Multisensory and Assistive Wearable Technology and Methods for Providing Sensory Relief Using the Same

Abstract

A system and method for providing sensory relief from distraction, inattention, anxiety, fatigue, and / or sensory issues to a user in need thereof, the user may be autistic / neurodiverse or neurotypical. The system may be configured to obtain user sensory sensitivity data indicative of the user's visual, sonic, or interoceptive sensitivity, use at least the user sensory sensitivity data to determine mediation data corresponding to a user-specific sensory threshold and a user-specific mediation, store the sensory threshold and mediation data, record sensory input stimuli to the user using one or more sensors, compare the sensory input stimuli to the sensory threshold, and determine a mediation to be provided to the user based at least on the mediation data in response to comparing the sensory input stimuli to the sensory threshold, the mediation configured to provide the user with relief from distraction, inattention, anxiety, fatigue, or sensory issues.

Description

[Background technology]

[0001] Where a document, act or item of knowledge is referenced or discussed in this specification, such reference or discussion is not an admission that that document, act or item of knowledge, or any combination thereof, was publicly available, was known to the public, was part of the common general knowledge, or constitutes prior art under any applicable legal provision, or is otherwise known to be relevant to any attempt to solve any problem to which this specification pertains.

[0002] A fairly high percentage of autistic adults (approximately 90%) report that sensory issues cause significant barriers at school and / or work (Non-Patent Document 1). Furthermore, 87% of autistic employees feel that environmental adjustments make a significant difference to their performance (Non-Patent Document 2). These figures make a compelling case for addressing sensory issues that affect autistic adults' ability to function well. Environmental factors are also known to cause persistent sensory and cognitive challenges (e.g., sensory overload) that lead to mental health challenges. Mental health is the most important autistic priority and major barrier to education / employment (Non-Patent Document 3) and contributes substantially to autism societal expenditure of over £27.5 billion annually in the UK - more than cancer, heart, stroke, and lung disease combined (Non-Patent Document 4). Summary of the Invention

[0003] This application addresses the aforementioned challenges by providing a wearable technology that (i) monitors the environment and tailors the user experience, (ii) reduces sensory load and enables greater participation, and (iii) offers groundbreaking opportunities for improving mental health through effective interventions. The wearable technology described herein enhances attentional focus, reduces sensory distraction, improves quality of life / reduces anxiety and fatigue.

[0004] It should be understood that the various individual aspects and features of the invention described herein may be combined in any number of any one or more of the individual aspects or features to form embodiments of the invention that are specifically contemplated and encompassed by the present invention.

[0005] One embodiment of the present application relates to a system comprising a wearable device comprising one or more sensors, one or more processors, and one or more non-transitory computer readable media having executable instructions stored thereon, which, when executed by the one or more processors, cause the system to: obtain (steps, operations) user sensory sensitivity data corresponding to a user input indicating whether a user of the wearable device is visually sensitive, ultrasonically sensitive, or interoceptively sensitive; and use at least the user sensory sensitivity data to determine one or more sensory thresholds specific to the user, and mediation data corresponding to one or more mediations specific to the user, wherein the one or more sensory thresholds are selected from auditory, visual, or physiological sensory thresholds. the sensory input stimuli to the user using one or more sensors; storing one or more sensory thresholds and mediation data; recording a sensory input stimulus to the user using one or more sensors; comparing the sensory input stimuli to one or more sensory thresholds specific to the user; determining a mediation to be provided to the user based at least on the mediation data in response to comparing the sensory input stimuli to the one or more sensory thresholds, the mediation being configured to provide the user with relief from distraction, inattention, anxiety, fatigue, or sensory issues; and providing the mediation to the user, the mediation comprising alert mediation, guidance mediation, or filter mediation.

[0006] In some embodiments, the operations further include storing a first identifier indicating whether the user is neurodiverse or neurotypical, and determining one or more sensory thresholds specific to the user and mediation data corresponding to one or more mediations specific to the user includes determining the one or more sensory thresholds and mediation data using at least the first identifier and the user sensory sensitivity data.

[0007] In some embodiments, the operations further include receiving user demographic data corresponding to user input indicating the user's age, education level, or gender, and determining one or more sensory thresholds specific to the user and mediation data corresponding to one or more mediations specific to the user includes determining the one or more sensory thresholds and mediation data using at least the first identifier, the user sensory sensitivity data, and the user demographic data.

[0008] In some embodiments, the first identifier indicates whether the user has autism.

[0009] In some embodiments, the first identifier indicates that the user has autism.

[0010] In some embodiments, the mediation is configured to provide relief to the user from fatigue, and the mediation comprises filter mediation, the filter mediation comprising filtering in real time an audio signal presented to the user or an optical signal presented to the user.

[0011] In some embodiments, the mediation is configured to provide the user with relief from the distraction by increasing the user's response time to the distraction.

[0012] In some embodiments, obtaining the user's sensory sensitivity data includes receiving, in the graphical user interface, one or more first responses by the user to one or more first prompts indicating whether the user is visually sensitive, ultrasonically sensitive, or interoceptively sensitive, and the operations further include deriving a first identifier indicating the user is autistic by receiving, in the graphical user interface, one or more second responses by the user to one or more second prompts indicating an anxiety level of the user, deriving one or more sensory sensitivity scores including a visual sensitivity score, an ultrasonic sensitivity score, or an interoceptive sensitivity score based on the sensory sensitivity data, deriving an anxiety score based on the one or more second responses, and predicting that the user is autistic based on at least the anxiety score and the one or more sensory sensitivity scores using a model that predicts a probability of autism based on at least the anxiety level and the one or more sensory sensitivity levels.

[0013] In some embodiments, obtaining the user's sensory sensitivity data includes receiving, in the graphical user interface, one or more first responses by the user to one or more first prompts indicating whether the user is visually sensitive, ultrasonically sensitive, or interoceptively sensitive, and the operations further include deriving a first identifier indicating the user is autistic by receiving, in the graphical user interface, one or more second responses by the user to one or more second prompts indicating the user's fatigue level, deriving one or more sensory sensitivity scores including a visual sensitivity score, an ultrasonic sensitivity score, or an interoceptive sensitivity score based on the sensory sensitivity data, deriving a fatigue score based on the one or more second responses, and predicting that the user is autistic based on at least the fatigue score and the one or more sensory sensitivity scores using a model that predicts a probability of autism based on at least the fatigue level and the one or more sensory sensitivity levels.

[0014] In some embodiments, obtaining the user sensory sensitivity data further includes recording a response by the user to a visual, acoustic, or physiological stimulus using at least one or more sensors.

[0015] In some embodiments, the mediation includes a combination mediation of at least two mediations selected from a warning mediation, a guidance mediation, and a filter mediation.

[0016] In some embodiments, the combination mediation includes alert mediation followed by filter mediation.

[0017] In some embodiments, the alert mediation includes alerting a user of a visual or audio distraction, and the filter mediation includes filtering in real time an audio or optical signal presented to the user, where the audio or optical signal is associated with the distraction.

[0018] In some embodiments, the system further comprises one or more fog nodes configured to locally store sensor data collected by the one or more sensors, the sensor data including first sensor data associated with the sensory input stimulus.

[0019] In some embodiments, storing the one or more sensory thresholds and the mediation data includes locally storing the one or more sensory thresholds and the mediation data using one or more fog nodes, and comparing the sensory input stimulus to the one or more sensory thresholds includes comparing the sensory input stimulus to the one or more sensory thresholds using the one or more fog nodes.

[0020] In some embodiments, the system further comprises one or more edge nodes configured to communicatively couple to the one or more fog nodes and to a cloud server located remotely from the wearable device.

[0021] In some embodiments, the one or more edge nodes are configured to encrypt first sensor data associated with the sensory input stimulus to obtain encrypted data, transmit the encrypted data to a cloud server, and receive a response from the cloud server.

[0022] In some embodiments, the one or more fog nodes and the one or more edge nodes are present on a local area network (LAN) that includes the wearable device, an ad-hoc network that includes the wearable device, a LAN of mobile devices directly coupled to the wearable device, or an ad-hoc network of mobile devices.

[0023] In some embodiments, the sensor data includes second sensor data that does not trigger mediation, and the system is configured such that the second sensor data that does not trigger mediation is not made available to any cloud server located remotely from the wearable device.

[0024] In some embodiments, the mediation includes filter mediation including filtering an optical signal presented to the user in real time, wherein the first sensor data associated with the sensory input stimulus includes first image data, and the one or more edge nodes or the one or more fog nodes are configured to determine whether the first image data is sufficiently similar to second image data stored in the cloud server, and determining the mediation to be provided to the user includes determining the filter mediation in response to determining that the first image data is sufficiently similar to the second image data.

[0025] In some embodiments, the mediation includes filter mediation including filtering an audio signal presented to the user in real time, wherein first sensor data associated with the sensory input stimulus includes first audio data, and the one or more edge nodes or the one or more fog nodes are configured to determine whether the first audio data is sufficiently similar to second audio data stored in a cloud server, and determining the mediation to be provided to the user includes determining the filter mediation in response to determining that the first audio data is sufficiently similar to the second audio data.

[0026] In some embodiments, the operations further include presenting to the user, in a graphical user interface, one or more access controls for controlling user data made available to one or more other users, where the user data includes sensor data collected by one or more sensors, one or more sensory thresholds, mediation data, or a record of mediation presented to the user, and receiving data corresponding to a user input selecting the one or more access controls. For example, the one or more access controls may be configured such that only sensor data that triggered the mediation is accessible to one or more other users (e.g., a general practitioner, a therapist, a family member, etc.). As another example, the one or more access controls may be configured such that certain types of sensor data (e.g., images or audio data of the environment) are not made available to other users. As a further example, the one or more access controls may be configured such that there are different hierarchies of data access, where some users have more access to certain types of data than other users.

[0027] In some embodiments, the operations further include presenting to the user in a graphical user interface one or more access controls that allow or deny access to one or more other users to affect the mediations presented to the user, and receiving data corresponding to a user input that activates the one or more access controls. For example, the wearer-user may give the therapist user access to modify user preferences to optimize the mediations presented to the user. In some embodiments, certain types of mediations may be disabled or enabled.

[0028] In some embodiments, the operations further include presenting to the user in a graphical user interface a graphical summary of the user's progress through use of the wearable device, the graphical summary including a running average of the time between mediations. As other examples, the graphical summary of the user's progress may show changes in sensory threshold and / or mediations over time, changes / running averages in the user's average response time to distracting stimuli, changes / running averages in the number of mediations required over a time frame (e.g., during the day) and / or for some event (e.g., while at work or in class), etc.

[0029] In some embodiments, the one or more sensors comprise a plurality of sensors of different types, the plurality of sensors comprising an auditory sensor, a galvanic skin sensor, a pupil sensor, a body temperature sensor, a head bob sensor, or an inertial motion unit, and recording the sensory input stimulus to the user includes obtaining first sensory data corresponding to the first sensory input stimulus from a first sensor of the plurality of sensors and second sensory data corresponding to the second sensory input stimulus from a second sensor of the plurality of sensors, and determining a mediation to be provided to the user includes inputting at least the first sensory data and the second sensory data into a fusion-based deep learning (FBDL) model that outputs an identification of the mediation to be provided to the user.

[0030] In some embodiments, determining the mediation to be provided to the user includes inputting at least the first sensory data, the second sensory data, and the mediation data into an FBDL model that outputs an identification of the mediation to be provided to the user.

[0031] Some embodiments of the present application include a step of obtaining, in a wearable device system, user sensory sensitivity data corresponding to a user input indicating whether a user of a wearable device of the wearable device system is visually sensitive, ultrasonically sensitive, or interoceptively sensitive; determining, in the wearable device system, one or more sensory thresholds selected from auditory, visual, or physiological sensory thresholds using at least the user sensory sensitivity data, one or more sensory thresholds specific to the user, and mediation data corresponding to one or more mediations specific to the user; storing the one or more sensory thresholds and the mediation data in a storage device of the wearable device system; and determining, in a storage device of the wearable device system, one or more sensory thresholds and the mediation data corresponding to one or more mediations specific to the user. The present invention relates to a method comprising the steps of: recording a sensory input stimulus to a user using one or more sensors; comparing, in a wearable device system, the sensory input stimulus to one or more sensory thresholds specific to the user; determining a mediation to be provided to the user based at least on the mediation data in response to comparing the sensory input stimulus to the one or more sensory thresholds, the mediation being configured to provide the user with relief from distraction, inattention, anxiety, fatigue, or sensory issues; and providing the mediation to the user using at least the wearable device, the mediation comprising alert mediation, guidance mediation, or filter mediation.

[0032] One embodiment of the present application relates to a system comprising a wearable device comprising one or more sensors, one or more processors, and one or more non-transitory computer readable media having executable instructions stored thereon, which when executed by the one or more processors, cause the system to perform operations including: connecting to a data store storing one or more sensory thresholds specific to a user of the wearable device, the one or more sensory thresholds being selected from auditory, visual, or physiological sensory thresholds; recording a sensory input stimulus to the user using the one or more sensors; comparing the sensory input stimulus to the one or more sensory thresholds specific to the user to determine an intervention to be provided to the user, the intervention being configured to provide the user with relief from distraction, inattention, anxiety, fatigue, or sensory issues; and providing the intervention to the user, the intervention comprising filtering an audio signal presented to the user or an optical signal presented to the user in real time. The physiological sensory threshold may be a physiological / psychophysiological sensory threshold.

[0033] In some embodiments, the operations further include communicatively coupling the system to an Internet of Things (IoT) device, where the sensory input stimulus is generated at least in part due to sound emitted by a speaker of the IoT device or light emitted by a lighting device of the IoT device, and providing the intervention to the user includes controlling the IoT device to filter the audio signal or the light signal in real time.

[0034] In some embodiments, the IoT device comprises a lighting device, and controlling the IoT device to filter the audio or light signal in real time includes controlling the IoT device to filter the light signal in real time, where filtering the light signal adjusts the brightness or color of the light output by the lighting device.

[0035] In some embodiments, the IoT device comprises a speaker, and controlling the IoT device to filter the audio or optical signal in real time includes controlling the IoT device to filter the audio signal in real time, where filtering the audio signal adjusts a frequency of the sound output by the speaker.

[0036] In some implementations, the wearable device further comprises a bone conduction transducer or hearing device, and providing intervention to the user includes filtering the audio signal in the frequency domain in real time in the wearable device, and after filtering the audio signal, presenting the audio signal to the user using the bone conduction transducer or hearing device by outputting vibrations or sound waves corresponding to the audio signal.

[0037] In some embodiments, the wearable device further comprises a head mounted display (HMD) that presents the light signals to the user, wherein the HMD is worn by the user, and providing intervention to the user further comprises filtering the light signals in real time by modifying a real time image of the real world environment presented to the user via the HMD.

[0038] In some embodiments, comparing the sensory input stimulus to one or more sensory thresholds specific to the user to determine an intervention to be provided to the user includes determining to filter the audio signal and to filter the light signal based on the same sensor data recorded by the one or more sensors.

[0039] In some embodiments, the wearable device further comprises an HMD that presents the optical signal to the user, the HMD being worn by the user, and providing intervention to the user includes filtering the optical signal in real time by modifying a real-time image of the real-world environment presented to the user via the HMD.

[0040] In some embodiments, modifying the real-time image includes inserting a virtual object into the real-time image or modifying the appearance of an object of the real-world environment in the real-time image.

[0041] In some embodiments, comparing the sensory input stimulus to one or more user-specific sensory thresholds to determine an intervention to be provided to the user includes inputting the sensory input stimulus and the one or more user-specific sensory thresholds into a trained model and automatically determining a visual intervention to be provided to the user based on an output of the trained model.

[0042] In some embodiments, the one or more sensors comprise a plurality of sensors of different types, the plurality of sensors comprising an auditory sensor, a galvanic skin sensor, a pupil sensor, a body temperature sensor, a head bob sensor, or an inertial motion unit, and recording the sensory input stimulus to the user includes recording a first sensor of the plurality of sensors and a second sensor of the plurality of sensors, and inputting the sensory input stimulus to the trained model includes inputting the first sensory input stimulus and the second sensory input stimulus to the trained model.

[0043] In some embodiments, the visual intervention includes presenting a warning to the user of a visually distracting object and filtering in real time the light signal presented to the user after it is determined that the user does not adequately respond to the warning within a certain period of time.

[0044] In some embodiments, the visual intervention includes filtering the light signal in real time to hide the visually distracting object without giving the user advance warning that the visually distracting object is present.

[0045] In some embodiments, the operations further include determining one or more user-specific sensory thresholds and one or more user-specific interventions by presenting a plurality of selectable templates to the user, each of the templates providing an indication of whether the user is visually sensitive, ultrasonically sensitive, or interoceptively sensitive, each of the templates being associated with a corresponding one or more sensory thresholds and one or more interventions, and receiving data corresponding to input by the user selecting one of the templates.

[0046] In some embodiments, determining the user-specific one or more sensory thresholds and the user-specific one or more interventions further includes receiving additional data corresponding to an additional user input selecting a preference including an audio preference, a visual preference, a physiological preference, an alert preference, a guidance preference, or an intervention preference, and modifying the one or more thresholds and the one or more interventions of the selected template in response to receiving the additional data to derive the user-specific one or more sensory thresholds and the one or more interventions. In some embodiments, the physiological preference is a psychophysiological preference.

[0047] In some embodiments, comparing the sensory input stimulus to one or more user-specific sensory thresholds to determine an intervention to be provided to the user includes inputting the sensory input stimulus and the one or more user-specific sensory thresholds into a trained model and automatically determining an intervention to be provided to the user based on an output of the trained model.

[0048] In some embodiments, the user is neurodiverse. In some embodiments, the user may be autistic.

[0049] In some embodiments, the intervention further comprises a warning intervention, wherein the warning intervention results in at least a 3% increase in response time and at least about a 26% increase in accuracy in the user from a baseline of error of commission, the error of commission being an indication of the user's failure to inhibit a response when prompted by the feedback device.

[0050] In some embodiments, the intervention further comprises guidance intervention, wherein the guidance intervention increases response time in the user by at least about 20% and accuracy by at least about 10% from a baseline of errors of commission, where the errors of commission are indicative of the user's failure to inhibit a response when prompted by the feedback device.

[0051] In some embodiments, the intervention further includes guidance intervention, where the guidance intervention increases the user's response time by at least about 2% and accuracy by at least about 30% from a baseline of errors of omission, where errors of omission are an indication of the user's failure to take appropriate action when a prompt is not received from the feedback device.

[0052] In some embodiments, the filtering intervention increases a user's response time by at least about 10% from a baseline of errors of omission, where errors of omission are indicative of a user's failure to take appropriate action when a prompt is not received from the feedback device.

[0053] In some embodiments, the filtering intervention results in a response time for the user that is at least about 15% faster than the response time for a neurotypical user using the system regarding errors of omission, where the error of omission is an indication of the user's failure to take appropriate action when a prompt is not received from the feedback device.

[0054] In some embodiments, the intervention further includes guidance intervention, where the guidance intervention results in a response time that is at least about 20% faster and about 8% more accurate in the user than the response time and accuracy in a standard user using the system for an error of commission, the error of commission being an indication of the user's failure to inhibit a response when prompted by the feedback device.

[0055] In some embodiments, the intervention further includes a warning intervention, where the warning intervention results in an accuracy in the user that is at least about 25% higher than the accuracy of a neurotypical user using the system with respect to an error of commission, the error of commission being an indication of the user's failure to inhibit a response when prompted by the feedback device.

[0056] One embodiment of the present application relates to a method including: connecting a wearable device system to a data store storing one or more sensory thresholds specific to a user of a wearable device of the wearable device system, the one or more sensory thresholds being selected from auditory, visual, or physiological sensory thresholds; recording a sensory input stimulus to the user using one or more sensors of the wearable device; comparing the sensory input stimulus to the one or more sensory thresholds specific to the user using the wearable device system to determine an intervention to be provided to the user, the intervention being configured to provide the user with relief from distraction, inattention, anxiety, fatigue, or sensory issues; and providing the intervention to the user using the wearable device system, the intervention comprising filtering an audio signal presented to the user or an optical signal presented to the user in real time. In some implementations, the physiological preference is a psychophysiological preference.

[0057] In some embodiments, the method further includes communicatively coupling the wearable device system to an IoT device, and providing the intervention to the user includes controlling the IoT device to filter audio or light signals in real time, and the sensory input stimulus is generated at least in part due to sound emitted by a speaker of the IoT device or light emitted by a light-emitting device of the IoT device.

[0058] In some embodiments, the IoT device comprises a lighting device, and controlling the IoT device to filter the audio or light signal in real time includes controlling the IoT device to filter the light signal in real time, where filtering the light signal adjusts the brightness or color of the light output by the lighting device.

[0059] In some embodiments, the IoT device comprises a speaker, and controlling the IoT device to filter the audio or optical signal in real time includes controlling the IoT device to filter the audio signal in real time, where filtering the audio signal adjusts a frequency of the sound output by the speaker.

[0060] In some implementations, the wearable device further comprises a bone conduction transducer or hearing device, and providing intervention to the user includes filtering the audio signal in the frequency domain in real time in the wearable device, and after filtering the audio signal, presenting the audio signal to the user using the bone conduction transducer or hearing device by outputting vibrations or sound waves corresponding to the audio signal.

[0061] An embodiment of this application relates to a system for providing sensory relief from distraction, inattention, anxiety, fatigue, sensory issues, or combinations thereof to a user in need thereof, the system comprising: (i) a wearable device; (ii) a database of one or more user-specific sensory thresholds selected from auditory, visual, and physiological sensory thresholds, one or more user-specific sensory resolutions selected from auditory, visual, and physiological sensory resolutions, or combinations thereof; (iii) an activation means for connecting the wearable device and the database; (iv) one or more sensors for recording sensory input stimuli to the user; (v) a comparison means for comparing the sensory input stimuli recorded by the one or more sensors with the database of one or more user-specific sensory thresholds to obtain a sensory resolution in the user; (vi) one or more feedback devices for transmitting the sensory resolution to the user; and (vii) a user-specific intervention means for providing the user with relief from distraction, inattention, anxiety, fatigue, sensory issues, or combinations thereof. The user-specific intervention means is selected from an alert intervention, a filter intervention, a guidance intervention, or combinations thereof, and the user may be a neurodiverse user or a neurotypical user. In a preferred embodiment, the neurodiverse user may be an autistic user. In some implementations, the physiological sensory threshold is a psychophysiological sensory threshold and the physiological sensory resolution is a psychophysiological sensory resolution.

[0062] In some implementations, the wearable device is a glasses frame comprising one or more sensors and one or more feedback devices.

[0063] In some embodiments, the one or more sensors are selected from one or more infrared sensors, one or more auditory sensors, one or more galvanic skin sensors, one or more inertial motion units, or a combination thereof.

[0064] In some embodiments, the one or more feedback devices are selected from one or more haptic drivers, one or more bone conduction transducers, or a combination thereof.

[0065] In some embodiments, the system further comprises a wireless or wired hearing device.

[0066] In some embodiments, the sensory input stimulus is selected from ecological auditory input, ecological visual input, egocentric physiological / psychophysiological input, or a combination thereof.

[0067] In some embodiments, the sensory input stimulus is measured by assessing one or more parameters selected from eye tracking, pupillometry, auditory cues, receptive sensations, physical movement, body temperature or ambient temperature fluctuations, pulse rate, respiration, or a combination thereof.

[0068] In some embodiments, the sensory resolution is provided by one or more alerts selected from visual alerts, auditory alerts, physiological / psychophysiological alerts, verbal alerts, or combinations thereof.

[0069] In some embodiments, the activation means is a power switch located on the wearable device.

[0070] In some implementations, the power switch is located on the left side of the wearable device.

[0071] In some implementations, the power switch is located on the right side of the wearable device.

[0072] In some embodiments, the power switch is a recessed power switch.

[0073] In some embodiments, the database is stored on a storage device.

[0074] In some embodiments, the storage device is selected from a fixed or mobile computer system, a portable wireless device, a smart phone, a tablet, or a combination thereof.

[0075] In some embodiments, warning intervention increases response time in autistic users by at least about 3% and accuracy by at least about 26% from a baseline in errors of commission, where errors of commission are indicative of a user's inability to inhibit a response when prompted by a feedback device.

[0076] In some embodiments, the warning intervention increases response time by at least about 18% and accuracy by at least about 2.0% in neurotypical users from a baseline of errors of commission, where errors of commission are indicative of a user's inability to inhibit a response when prompted by a feedback device.

[0077] In some embodiments, guidance intervention increases response time by at least about 20% and accuracy by at least about 10% in autistic users from a baseline of errors of commission, where errors of commission are indicative of a user's inability to inhibit a response when prompted by a feedback device.

[0078] In some embodiments, guidance intervention results in at least about a 2% increase in response time and at least about a 30% increase in accuracy in autistic users from a baseline of errors of omission, where errors of omission are indicative of a user's failure to perform appropriate action when no prompts are received from a feedback device.

[0079] In some embodiments, filter intervention increases response times in autistic users by at least about 10% from a baseline for errors of omission, where errors of omission are indicative of a user's failure to perform an appropriate action when no prompt is received from the feedback device.

[0080] In some embodiments, using filter intervention, response times in autistic users are at least about 15% faster than typical users in response to errors of omission, where errors of omission are indicative of a user's failure to take appropriate action when a prompt is not received from the feedback device.

[0081] In some embodiments, using guidance intervention, response times are at least about 20% faster and accuracy is at least about 8% higher in autistic users than in neurotypical users in errors of commission, where errors of commission are indicative of a user's inability to inhibit a response when prompted by a feedback device.

[0082] In some embodiments, with alert intervention, the accuracy of autistic users is at least about 25% higher than that of typical users in predicting errors of commission, where errors of commission are indicative of a user's inability to inhibit a response when prompted by a feedback device.

[0083] In one embodiment, a method for providing sensory relief from distraction, inattention, anxiety, fatigue, sensory issues, or combinations thereof to a user in need thereof includes creating a database of one or more user specific sensory thresholds selected from auditory, visual, and physiological / psychophysiological sensory thresholds, one or more user specific sensory resolutions selected from auditory, visual, and physiological / psychophysiological sensory resolutions, or combinations thereof; attaching a wearable device to the user, the wearable device comprising one or more sensors and one or more feedback devices; activating the wearable device to connect to the database; and detecting the sensory relief via the one or more sensors. the sensory input stimulus to the user based on the sensory threshold; comparing the sensory input stimulus to a database of one or more user-specific sensory thresholds; selecting an appropriate user-specific sensory resolution from the database; delivering the user-specific sensory resolution to the user via one or more feedback devices; and providing a user-specific intervention (a / k / digital mediation) to give the user relief from distraction, inattention, anxiety, fatigue, sensory issues, or combinations thereof, wherein the user-specific intervention is selected from an alert intervention, a filter intervention, a guidance intervention, or combinations thereof, and the user is an autistic user, a neuro-typical user, or a neurodiverse user.

[0084] In some embodiments, the one or more sensors are selected from one or more infrared sensors, one or more auditory sensors, one or more galvanic skin sensors, one or more inertial motion units, or a combination thereof.

[0085] In some embodiments, the one or more feedback devices are selected from one or more haptic drivers, one or more bone conduction transducers, or a combination thereof.

[0086] In some embodiments, the sensory input stimulus is selected from an auditory input, a visual input, a physiological / psychophysiological input, or a combination thereof.

[0087] In some embodiments, the sensory input stimulus is measured by assessing one or more parameters selected from eye tracking, pupillometry, auditory cues, receptive sensations, physical movement, body temperature or ambient temperature fluctuations, pulse rate, respiration, or a combination thereof.

[0088] In some embodiments, the sensory resolution is provided by one or more alerts selected from visual alerts, auditory alerts, physiological / psychophysiological alerts, verbal alerts, or combinations thereof.

[0089] In some embodiments, activation of the wearable device and connection to the database is via a power switch located on the wearable device.

[0090] In some embodiments, the power switch is located on the left side of the wearable device or on the right side of the wearable device.

[0091] In some embodiments, the power switch is a recessed power switch.

[0092] In some implementations, the wearable device is an eyeglass frame.

[0093] In some embodiments, the database is stored on a storage device.

[0094] In some embodiments, the storage device is selected from a fixed or mobile computer system, a portable wireless device, a smart phone, a tablet, or a combination thereof.

[0095] In some embodiments, warning intervention increases response time in autistic users by at least about 3% and accuracy by at least about 26% from a baseline in errors of commission, where errors of commission are indicative of a user's inability to inhibit a response when prompted by a feedback device.

[0096] In some embodiments, the warning intervention increases response time by at least about 18% and accuracy by at least about 2.0% in neurotypical users from a baseline of errors of commission, where errors of commission are indicative of a user's inability to inhibit a response when prompted by a feedback device.

[0097] In some embodiments, guidance intervention increases response time by at least about 20% and accuracy by at least about 10% in autistic users from a baseline of errors of commission, where errors of commission are indicative of a user's inability to inhibit a response when prompted by a feedback device.

[0098] In some embodiments, the guidance intervention increases response time by at least about 2% and accuracy by at least about 30% in autistic users from a baseline of errors of omission, where errors of omission are an indication of a user's failure to perform appropriate action when a prompt is not received from a feedback device.

[0099] In some embodiments, filter intervention increases response times in autistic users by at least about 10% from a baseline for errors of omission, where errors of omission are indicative of a user's failure to perform an appropriate action when no prompt is received from the feedback device.

[0100] In some embodiments, using filter intervention, response times in autistic users are at least about 15% faster than typical users in response to errors of omission, where errors of omission are indicative of a user's failure to take appropriate action when a prompt is not received from the feedback device.

[0101] In some embodiments, using guidance intervention, response times are at least about 20% faster and accuracy is at least about 8% higher in autistic users than in neurotypical users in errors of commission, where errors of commission are indicative of a user's inability to inhibit a response when prompted by a feedback device.

[0102] In some embodiments, with alert intervention, the accuracy of autistic users is at least about 25% higher than that of typical users in predicting errors of commission, where errors of commission are indicative of a user's inability to inhibit a response when prompted by a feedback device.

[0103] In one embodiment, the wearable device comprises one or more sensors and one or more feedback devices, the combination of the one or more sensors and one or more feedback devices providing sensory relief from distraction, inattention, anxiety, fatigue, sensory issues, or a combination thereof to a user / wearer in need thereof.

[0104] In some implementations, the wearable device is an eyeglass frame.

[0105] In some embodiments, the one or more sensors are connected to the eyeglass frame.

[0106] In some embodiments, the one or more feedback devices are connected to the eyeglass frame.

[0107] In some embodiments, the eyeglass frame comprises a rim, two earpieces, and a hinge connecting the earpieces to the rim.

[0108] In some embodiments, the one or more sensors are selected from the group consisting of one or more infrared sensors, one or more auditory transducers, one or more galvanic skin sensors, one or more inertial motion units, or combinations thereof.

[0109] In some implementations, the infrared sensor is surface mounted on the inside of the wearable device.

[0110] In some embodiments, the infrared sensor is positioned to strike the right eye, the left eye, or both eyes of the user.

[0111] In some embodiments, the acoustic transducer is a subminiature microphone.

[0112] In some implementations, a miniature microphone is surface mounted on the outside of the wearable device.

[0113] In some implementations, the wearable device comprises at least two acoustic transducers, a first acoustic transducer positioned at an angle of approximately 110° relative to the second acoustic transducer.

[0114] In some implementations, the galvanic skin sensor is surface mounted on the inside of the wearable device, such that the galvanic skin sensor is in direct contact with the user's skin.

[0115] In some implementations, the inertial movement unit is implemented internally inside the wearable device.

[0116] In some embodiments, the one or more feedback devices are selected from one or more haptic drivers, one or more bone conduction transducers, or a combination thereof.

[0117] In some implementations, the haptic drive is implemented internally inside the wearable device.

[0118] In some embodiments, the haptic drive is implemented internally behind the inertial motion unit inside the wearable device.

[0119] In some embodiments, the haptic drive provides vibration patterns in response to sensory input stimuli selected from eye tracking, pupil measurement, auditory cues, interoceptive recognition, body movement, body or ambient temperature fluctuations, pulse rate, respiration, or combinations thereof.

[0120] In some embodiments, the stereoscopic bone conduction transducer is surface mounted inside the wearable device, such that the stereoscopic bone conduction transducer is in direct contact with the user's skull.

[0121] In some embodiments, the stereophonic bone conduction transducer provides auditory tones, pre-recorded auditory guidance, real-time filtering, or a combination thereof, in response to sensory input stimuli selected from eye tracking, pupil measurement, auditory cues, receptive sensations, body movement, body or ambient temperature fluctuations, pulse rate, respiration, or a combination thereof.

[0122] In some embodiments, the wearable device further comprises an optional wireless or wired hearing device.

[0123] In some embodiments, the wearable device further comprises an intervention means for providing the user with relief from distraction, inattention, anxiety, fatigue, sensory issues, or combinations thereof, the intervention means being selected from an alert intervention, a filter intervention, a guidance intervention, or combinations thereof.

[0124] In some implementations, the wearable device further comprises a power switch. The power switch may be located on the left side of the wearable device or on the right side of the wearable device. The power switch may be a recessed power switch.

[0125] In one embodiment, a non-transitory computer readable medium has executable instructions stored thereon that, when executed by a processor, cause the wearable device to perform operations including connecting the wearable device to a data store storing one or more sensory thresholds and one or more sensory resolutions specific to a user, where the one or more sensory thresholds are selected from auditory, visual, or physiological / psychophysiological sensory thresholds and the one or more sensory resolutions are selected from auditory, visual, or physiological / psychophysiological sensory resolutions; recording sensory input stimuli to the user via one or more sensors; comparing the sensory input stimuli recorded by the one or more sensors with the one or more sensory thresholds to obtain the user's sensory resolution; and transmitting the sensory resolution to the user.

[0126] In some embodiments, the operations further include communicatively coupling to an IoT device that provides the sensory input stimulus to the user, and transmitting the sensory resolution to the user includes controlling the IoT device to transmit the sensory resolution after communicatively coupling to the IoT device.

[0127] In some embodiments, the IoT device comprises a networked lighting device, and controlling the IoT device to transmit the sensory resolution includes controlling the brightness or color output of the networked lighting device.

[0128] In some embodiments, the IoT device comprises a networked speaker, and controlling the IoT device to transmit the sensory resolution includes controlling a volume, equalization setting, or channel balance of the networked speaker.

[0129] In some embodiments, comparing the sensory input stimuli recorded by the one or more sensors with one or more user-specific sensory thresholds to obtain the sensory resolution of the user includes inputting the sensory input stimuli and the one or more user-specific sensory thresholds into a trained model to automatically determine the sensory resolution of the user.

[0130] In some embodiments, the operations further include determining one or more sensory thresholds and one or more sensory resolutions by presenting a plurality of selectable templates to a user, each of the templates providing an indication of whether the user is visually sensitive, ultrasonically sensitive, or interoceptively sensitive, each of the templates being associated with a corresponding one or more thresholds and one or more sensory resolutions, and receiving data corresponding to input by a user selecting one of the templates.

[0131] In some embodiments, determining the one or more user-specific sensory thresholds and the one or more user-specific sensory resolutions further includes receiving additional data corresponding to additional user input selecting preferences including audio preferences, visual preferences, physiological / psychophysiological preferences, alert preferences, guidance preferences, or intervention preferences, and modifying the one or more thresholds and the one or more sensory resolutions of the selected template in response to receiving the additional data to derive the one or more user-specific sensory thresholds and the one or more user-specific sensory resolutions.

[0132] Other features and aspects of the disclosed methods will become apparent from the following detailed description, taken in conjunction with the accompanying drawings, which illustrate, by way of example, features in accordance with embodiments of the present disclosure. The Summary is not intended to limit the scope of the claimed disclosure, which is defined solely by the claims appended hereto. [Brief description of the drawings]

[0133] The present disclosure, in accordance with one or more various embodiments, will now be described in detail with reference to the following figures, in which: The drawings are provided for purposes of illustration only and merely depict typical or exemplary embodiments of the present disclosure.

[0134] [Figure 1] FIG. 1 is a schematic diagram of a wearable device according to some implementations of the present disclosure.

[0135] [Diagram 2] FIG. 1 is a graphical representation of sensitivity across three modalities, visual, auditory, and anxiety, observed in the Pre-Test Battery (PTBE) described herein.

[0136] [Diagram 3] 1 is a graphical representation of interest in wearable devices among participants with autism spectrum condition (ASC) in PTBE.

[0137] [Figure 4] FIG. 1 is a flow chart of the standard testing protocol for the Sustained Attention to Response Task (SART) test.

[0138] [Diagram 5] 1 is a flow chart of the standard Wizard of Oz testing protocol.

[0139] [Figure 6] 1 is a flow chart of the SART / WoZ study protocol, according to some implementations of the present disclosure.

[0140] [Figure 7] 1 is a graphical representation of the recruitment scores of study participants for a wearable device study.

[0141] [Figure 8A] FIG. 1 is a graphical representation of errors of commission (EOC) across the cohort of participants in the SART / WoZ study described herein, showing EOC from baseline to baseline. [Figure 8B] FIG. 1 is a graphical representation of errors of commission (EOC) across the cohort of participants in the SART / WoZ study described herein, illustrating the EOC intervention effect. [Figure 8C] FIG. 1 is a graphical representation of errors of commission (EOC) across the cohort of participants in the SART / WoZ study described herein, showing the persistence effect of EOC.

[0142] [Figure 9A] FIG. 1 is a graphical representation of EOC with respect to response time (RT) across the cohort of participants in the SART / WoZ study described herein, showing EOC vs. RT from starting baseline to final baseline. [Figure 9B] FIG. 1 is a graphical representation of EOC in relation to response time (RT) across the cohort of participants in the SART / WoZ study described herein, showing the intervention effect of EOC versus RT. [Figure 9C] FIG. 1 is a graphical representation of EOC in relation to response time (RT) across the cohort of participants in the SART / WoZ study described herein, showing the sustained effect of EOC versus RT.

[0143] [Figure 10A] FIG. 1 is a graphical representation of EOC grouped by study participants. [Figure 10B] FIG. 1 is a graphical representation of EOC grouped by study participants. [Figure 10C] FIG. 1 is a graphical representation of EOC grouped by study participants.

[0144] [Figure 11A] FIG. 1 is a graphical representation of EOC versus RT grouped by study participants. [Figure 11B] FIG. 1 is a graphical representation of EOC versus RT grouped by study participants. [Figure 11C] FIG. 1 is a graphical representation of EOC versus RT grouped by study participants.

[0145] [Figure 12A] FIG. 1 is a graphical representation of errors of omission (EOO) across the cohort of participants in the SART / WoZ study described herein, showing EOO from starting baseline to final baseline. [Figure 12B]FIG. 1 is a graphical representation of errors of omission (EOO) across the cohort of participants in the SART / WoZ study described herein, showing the EOO intervention effect. [Figure 12C] FIG. 1 is a graphical representation of errors of omission (EOO) across the cohort of participants in the SART / WoZ study described herein, showing the sustained effect of EOO.

[0146] [Figure 13A] FIG. 1 is a graphical representation of EOO in relation to RT across the cohort of participants in the SART / WoZ study described herein, showing EOO vs. RT from starting baseline to final baseline. [Figure 13B] FIG. 1 is a graphical representation of EOO in relation to RT across the cohort of participants in the SART / WoZ study described herein, showing the EOO vs. RT intervention effect. [Figure 13C] FIG. 1 is a graphical representation of EOO in relation to RT across the cohort of participants in the SART / WoZ study described herein, showing the sustained effect of EOO versus RT.

[0147] [Figure 14A] FIG. 1 is a graphical representation of EOO grouped by study participants. [Figure 14B] FIG. 1 is a graphical representation of EOO grouped by study participants. [Figure 14C] FIG. 1 is a graphical representation of EOO grouped by study participants.

[0148] [Figure 15A] FIG. 1 is a graphical representation of EOO versus RT grouped by study participants. [Figure 15B] FIG. 1 is a graphical representation of EOO versus RT grouped by study participants. [Figure 15C] FIG. 1 is a graphical representation of EOO versus RT grouped by study participants.

[0149] [Figure 16]FIG. 1 is a block diagram of components of a wearable device according to some implementations of the disclosure.

[0150] [Figure 17] FIG. 13 is a block diagram of additional microprocessor details (ARM processor) of a wearable device according to some implementations of the present disclosure. [Figure 17-1] FIG. 13 is a block diagram (continued) of additional microprocessor details (ARM processor) of a wearable device according to some implementations of the present disclosure.

[0151] [Figure 18] 1 is a flow chart of various components of a test variable according to some implementations of the present disclosure.

[0152] [Figure 19] 1 illustrates a wearable device system including a wearable device in communication with a mobile device and a data store, according to some implementations of the present disclosure.

[0153] [Figure 20] FIG. 1 shows an operational flow diagram illustrating an example method for initializing and iteratively updating one or more sensory thresholds and one or more interventions associated with a particular user, according to some implementations of the present disclosure.

[0154] [Figure 21] 1 illustrates a wearable device system including a wearable device that communicates with a mobile device that controls an IoT device using a speaker, according to some implementations of the present disclosure.

[0155] [Figure 22] 1 illustrates a wearable device system including a wearable device that communicates with a mobile device that controls an IoT device having a light emitting device, according to some implementations of the present disclosure.

[0156] [Figure 23]1 illustrates an example wearable device that can be utilized to provide visual intervention, according to some implementations of the present disclosure.

[0157] [Figure 24A] 13 illustrates interventions that may be delivered using a real-time optical enhancement algorithm, including haptic alerts, tone alert guidance, and eraser effects, according to some implementations of the present disclosure.

[0158] [Figure 24B] 1 illustrates interventions that may be delivered using real-time optical enhancement algorithms, including text warnings, blurring effects, and obscuring effects, according to some implementations of the present disclosure.

[0159] [Figure 24C] 1 illustrates interventions that may be delivered using real-time optical enhancement algorithms according to some implementations of the present disclosure, including color balance, contrast effects, and enhancement effects.

[0160] [Diagram 25] 1 illustrates one particular example of a workflow using a real-time optical enhancement algorithm to provide real-time intervention in scenarios with distracting visual sources, according to some implementations of the present disclosure.

[0161] [Figure 26] 1 illustrates a susceptibility mental health distraction model, according to some embodiments of the present disclosure.

[0162] [Figure 27] 1 is a flow chart illustrating the design / methodology of the PPI study described herein.

[0163] [Figure 28] 1 shows a word cloud derived from alternative autistic speech representations during the PPI study described herein.

[0164] [Figure 29] 1 shows the mean distribution of anxiety and distractibility scores in diagnostic groups across demographic variables for the PPI study described herein.

[0165] [Diagram 30] 1 illustrates a non-autistic mediation model according to some embodiments of the present disclosure.

[0166] [Diagram 31] 1 illustrates an autism mediation model according to some embodiments of the present disclosure.

[0167] [Diagram 32] 1 illustrates an autism mediation model predicting auditory distraction due to fatigue, according to some embodiments of the present disclosure.

[0168] [Diagram 33] 1 illustrates an autism mediation model predicting distraction from fatigue physiology, according to some embodiments of the present disclosure.

[0169] [Figure 34A] 1 shows summary results of the PPI studies described herein.

[0170] [Figure 34B] 1 shows summary results of the PPI studies described herein.

[0171] [Figure 35A] 1 shows summary results of the SART / WOz clinical study described herein.

[0172] [Figure 35B] 1 shows summary results of the SART / WOz clinical study described herein.

[0173] [Figure 36A]FIG. 1 is an operational flow diagram illustrating an example of a method for initializing and iteratively updating one or more sensory thresholds and one or more mediations associated with a particular user, according to some embodiments of the present disclosure.

[0174] [Figure 36B] FIG. 1 is an operational flow diagram illustrating an example of a method for predicting whether a user is neurodiverse (e.g., autistic) or neurotypical, according to some embodiments of the present disclosure.

[0175] [Figure 36B] FIG. 1 is an operational flow diagram illustrating an example of a method for predicting whether a user is neurodiverse (e.g., autistic) or neurotypical, according to some embodiments of the present disclosure.

[0176] [Figure 37] 1 illustrates an example of a system architecture / topology for implementing fog data processing according to some embodiments of the present disclosure.

[0177] [Figure 38A] 1 illustrates a specific example of a wearable system architecture, including data flow, leveraging fog and edge computing in accordance with some embodiments of the present disclosure.

[0178] [Figure 38B] FIG. 38B is a flow diagram illustrating operations performed by the system of FIG. 38A according to some embodiments of the present disclosure.

[0179] [Figure 39] FIG. 1 is a high-level flow diagram of an Open Learner Model (OLM) framework according to some embodiments of the present disclosure. [Figure 39-1] FIG. 1 is a high-level flow diagram of the Open Learner Model (OLM) framework (continued), according to some embodiments of the present disclosure.

[0180] [Diagram 40] 1 shows a table illustrating the OLMs described herein that are available. [Figure 40-1] 1 shows a continuation of the table illustrating the OLMs described herein that are available.

[0181] [Diagram 41] 1 shows a flowchart of the OLM described herein that may be utilized.

[0182] [Diagram 42] 1 shows a table of the OLM described herein and illustrates how the model is presented to stakeholders. [Figure 42-1] 1 shows a table of the OLM described herein and a table (continued) explaining how the model is presented to stakeholders.

[0183] [Diagram 43] 1 shows a flow chart of the OLM described herein, illustrating how the model is presented to a stakeholder.

[0184] [Diagram 44] 1 shows a table of the OLM described herein that describes who controls access to others.

[0185] [Diagram 45] 1 shows a flow chart of the OLM described herein, describing who controls access to others.

[0186] [Diagram 46] 1 illustrates a system for implementing an augmented reality-based multimodal learning analytics framework according to some embodiments of the present disclosure.

[0187] [Figure 47] 1 illustrates an example of a fusion-based deep learning model according to some embodiments of the present disclosure.

[0188] [Figure 48] FIG. 1 is a high-level flow diagram conceptually illustrating the operation of a multisensory assistive wearable system according to some implementations of the present disclosure.

[0189] The figures are not exhaustive and are not intended to limit the disclosure to the precise forms disclosed. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0190] Further aspects, features and advantages of the present invention will become apparent from the following detailed description.

[0191] As used herein, the singular forms "a," "an," and "the" may be intended to include the plural forms as well, unless the context clearly indicates otherwise. Furthermore, the use of "or" is intended to include "and / or" unless the context clearly dictates otherwise.

[0192] As used herein, "about" is a term of approximation and is intended to include slight variations in the literally recited quantities as would be understood by one of ordinary skill in the art. Such variations include, for example, standard deviations associated with conventional measurement techniques or the particular measurement techniques described herein. All values ​​characterized by the above modifier "about" are also intended to include the exact numerical values ​​disclosed herein. Moreover, all ranges include upper and lower limits.

[0193] Any apparatus, device or article of manufacture described herein is intended to encompass apparatus, devices or articles of manufacture that consist of, consist essentially of, and include the various components / ingredients identified herein, unless expressly indicated to the contrary.

[0194] As used herein, the recitation of a numerical range for a variable is intended to convey that the variable can be equal to any value within that range, as well as any and all subranges encompassed within the broader range. Thus, the variable can be equal to any integer value or value within the numerical range, including the end points of the range. As an example, a variable described as having a value between 0 and 10 can be 0, 4, 2-6, 2.75, 3.19-4.47, etc.

[0195] In the specification and claims, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. As used herein, unless otherwise stated, the word "or" is used in the inclusive sense of "and / or" and not in the exclusive sense of "or."

[0196] Unless otherwise indicated, each individual feature or embodiment herein can be combined, without limitation, with any other individual feature or embodiment described herein, and such combinations are specifically contemplated as being within the scope of the present invention, regardless of whether they are explicitly described as combinations herein.

[0197] Unless otherwise defined, technical and scientific terms used herein have the meanings commonly understood by one of ordinary skill in the art to which this specification pertains. Reference is made herein to various methods and materials known to those skilled in the art.

[0198] As used herein, the term "alert intervention" can include the following: upon activation of ecological and / or physiological (e.g., psychophysiological) thresholds corresponding to the wearer's preferences, a signal is delivered to a haptic driver (i) that provides a gentle haptic vibration pattern to convey information to the wearer that focus, anxiety, fatigue or related characteristics require attention, and / or (ii) to a bone conduction transducer that delivers an auditory / sound wave message (e.g., pre-recorded text-speech, beeps, etc.) that augments the haptics with an auditory intervention and set of instructions.

[0199] As used herein, the term "filter intervention" can include: performing audio signal processing or optical signal processing in case of activation of echo and / or physiological (e.g., psychophysiological) thresholds that correspond to the wearer's preferences and require hearing or optical filtering. Digital audio signal processing can deliver real-time and low latency audio signals including corrected amplitude (compression, expansion), frequency (dynamic, shelving, low / high cut, and parametric equalization), spatial realignment (repositioning, stereo to mono) and / or phase correction (time delay, comb filtering, linear phase matching). In one embodiment, the filter invention can be delivered to a bone conduction transducer. In other embodiments, the filter invention can be delivered to any wireless or wired hearing device including, but not limited to, earbuds, earphones, headphones, etc.

[0200] As used herein, the term "guidance intervention" can include interventions similar to alert interventions that can provide guidance by step-by-step instructions for guidance on refocusing, head shaking, pupil activity, pulse, temperature, respiration, anxiety, and fatigue. These pre-recorded text-to-speech audio streams can be delivered to a bone conduction system that provides the step-by-step instructional interventions privately and unobtrusively.

[0201] As used herein, the term "combined intervention" can include interventions that can be a combination of the following: alert, filter and guidance interventions, and can be selected by the wearer, provided in response to a triggering mechanism. For example, only sonic disturbances can be addressed by filter intervention, while all other issues (attention-focus, anxiety, fatigue, etc.) can be intervened by haptics, text voice alerts, step-by-step guidance in the form of long sounds, etc.

[0202] As used herein, the terms "user" and "wearer" are used interchangeably.

[0203] As used herein, the terms "intervention" and "mediation" are used interchangeably.

[0204] As used herein, the term "error of commission" is an indication of a user's failure to inhibit a response when prompted by a feedback device.

[0205] As used herein, the term "error of omission" is an indication of a user's failure to take appropriate action when a prompt is not received from a feedback device.

[0206] As used herein, the term "response time" is intended to include the time it takes a participant to respond to a sensory cue and / or an alert, filter and / or guidance intervention. Response time is also referred to interchangeably as reaction time and is defined as the amount of time between when a participant perceives a sensory cue and when the participant responds to said sensory cue. Response Time or Reaction Time is the ability to detect, process and respond to a stimulus.

[0207] The term "real-time" when used herein to refer to processing, such as any processing that may include filtering of audio and / or optical signals presented to a user, is intended to refer to processing and / or filtering signals with minimal latency after the original audio and / or optical signals are generated. For example, the latency may be a non-zero value, such as about 500 milliseconds (ms) or less, about 250 ms or less, about 200 ms or less, about 150 ms or less, about 100 ms or less, about 90 ms or less, about 80 ms or less, about 70 ms or less, about 60 ms or less, about 50 ms or less, about 40 ms or less, about 30 ms or less, about 20 ms or less, or about 10 ms or less, ranges and / or combinations thereof. The minimum latency may be subject to system and hardware and software limitations, including communication protocol latency, digital signal processing latency, electrical signal processing latency, combinations thereof, and the like. In some cases, the real-time filtering of audio and / or optical signals may be perceived by a user to be immediate, instantaneous, or near-instantaneous and / or instantaneous.

[0208] Autism spectrum conditions (ASC) are a lifelong diagnosis with a subset of features that include hyperresponsiveness, exploratory and / or hyporesponsiveness to sensory input or unusual interests. These qualities are evident across environmental domains (e.g., responses to certain sounds, visual attraction with light or movement) and physiological / psychophysiological domains (e.g., anxiety, respiration or heat therapy). Scholars report that nine in ten (90%) adults with autism experience sensory problems, causing significant barriers at school / work (Leekam et al., 2007). Individuals with ASC often show persistent deficits in social communication and interaction across multiple contexts. Additional features include restricted and repetitive patterns of behavior and interests (RRBI). Importantly, RSBI includes hyperresponsiveness, exploratory and / or hyporesponsiveness to sensory input, along with attainable unusual interests in sensory aspects of the environment, and physiological / psychophysiological responses to sight, texture, smell, touch, and sound. As diagnoses of the ASC population grow exponentially over time, ever-widening social policy fissures proliferate, which often impair autistic individuals' smooth transition into the fabric of everyday life. Experts identify this as a gap resulting from one of the following: (i) stunted public / government support for neurodiverse individuals; ii) tensions between the autistic community and society; and iii) limited support for educational / occupational pathways in later life. The negative impact of policy-related factors results in social costs that can become even more significant and in some cases irreparable.

[0209] This application provides various interventions to modify, redirect, and / or attenuate disruptive stimuli. That is, systems, devices, and methods are described herein to determine whether distractions that may be exacerbated at school and work are present, and to provide interventions to compensate for such distractions, thereby reducing anxiety and providing sensory relief for neurotypical and neurodiverse individuals. This application aims to help individuals learn, adapt, and internalize how to best respond to intruding ecological stimuli and the resulting physiological / psychophysiological responses. As described herein, wearables can pave the way for a call and response process that can ultimately be transferred directly from the machine or system to the person through an iterative process observed and experienced by the user, thus embedding guidance for similarly repeated / future scenarios. For example, an autistic individual can observe, experience, and learn exactly how an artificial intelligence / cognitive enhancement system detects, filters, and guides him or her when faced with undesirable sensory stimuli.

[0210] One embodiment relates to a system for providing sensory relief from distraction, inattention, anxiety, fatigue, sensory issues, or combinations thereof to a user in need thereof, the system comprising: (i) a wearable device; (ii) a database of one or more user-specific sensory thresholds selected from auditory, visual, and physiological sensory thresholds, one or more user-specific sensory resolutions selected from auditory, visual, and physiological sensory resolutions, or combinations thereof; (iii) an activation means for connecting the wearable device and the database; (iv) one or more sensors for recording sensory input stimuli to the user; (v) a comparison means for comparing the sensory input stimuli recorded by the one or more sensors with the database of one or more user-specific sensory thresholds to obtain a sensory resolution in the user; (vi) one or more feedback devices for transmitting the sensory resolution to the user; and (vii) a user-specific intervention means for providing the user with relief from distraction, inattention, anxiety, fatigue, sensory issues, or combinations thereof. The user-specific intervention means can be selected from an alert intervention, a filter intervention, a guidance intervention, or combinations thereof. The user may be an autistic user, a neurodiverse user, or a neuro-typical user.

[0211] Another embodiment relates to a method for providing sensory relief from distraction, inattention, anxiety, fatigue, sensory issues, or a combination thereof to a user in need thereof, the method comprising the steps of: (i) creating a database of one or more user specific sensory thresholds selected from auditory, visual, and physiological / psychophysiological sensory thresholds, one or more user specific sensory resolutions selected from auditory, visual, and physiological / psychophysiological sensory resolutions, or a combination thereof; and (ii) attaching a wearable device to the user, the wearable device comprising one or more sensors and one or more feedback devices. (iii) activating the wearable device and connecting to a database; (iv) recording sensory input stimuli to the user via one or more sensors; (v) comparing the sensory input stimuli to a database of one or more user-specific sensory thresholds; (vi) selecting an appropriate user-specific sensory resolution from the database; (vi) delivering the user-specific sensory resolution to the user via one or more feedback devices; and (vii) providing a user-specific intervention to provide the user with relief from distraction, inattention, anxiety, fatigue, sensory issues, or combinations thereof.

[0212] Another embodiment relates to a wearable device comprising one or more sensors and one or more feedback devices. According to a further embodiment, the wearable device may be a glasses frame. One or more sensors and / or one or more feedback devices may be connected to the glasses frame. The glasses frame may comprise a rim, two earpieces, and a hinge connecting the earpieces to the rim. In an alternative embodiment, the wearable device may include jewelry, smart clothing, and accessories, including but not limited to rings, sensor textiles, wristbands, watches, pins, hearing aids, assistive devices, medical devices, virtual, augmented, and mixed reality (VR / AR / MR) headsets, and the like. The wearable device may have the ability to interface with mobile and / or network devices for alerts, filters, and guidance interventions, and may include various combinations of sensors and feedback devices.

[0213] According to further embodiments, the one or more sensors can be selected from one or more infrared sensors, one or more auditory sensors, one or more galvanic skin sensors, one or more inertial motion units, or combinations thereof. The infrared sensor can be surface mounted on the inside of the wearable device. The infrared sensor can be positioned to be incident on the right eye, the left eye, or both eyes of the user.

[0214] According to further embodiments, the one or more feedback devices may be selected from one or more haptic drivers, one or more bone conduction transducers, or a combination thereof.

[0215] According to further embodiments, the wearable device may further comprise a wireless or wired hearing device.

[0216] According to further embodiments, the sensory input stimuli can be selected from auditory input, visual input, physiological / psychophysiological input, or a combination thereof. The sensory input stimuli can be measured by evaluating one or more parameters selected from eye tracking, pupil measurement, auditory cues, interoceptive awareness, body movement, body or ambient temperature fluctuations, pulse rate, respiration, or a combination thereof. The sensory resolution can be provided by one or more alerts selected from visual alerts, auditory alerts, physiological / psychophysiological alerts, verbal alerts, or a combination thereof.

[0217] According to a further embodiment, the activation means can be a power switch located on the wearable device. The power switch can be located on the left side of the wearable device and / or on the right side of the wearable device. The power switch can be a recessed power switch. In another embodiment, when in standby mode, power can be provided from a user interface component including, but not limited to, a mobile phone, a laptop, a tablet, a desktop computer, etc., and any user interface known in the art can be used without limitation. In another embodiment, the activation means can include a power switch or power source that can be activated remotely (i.e., when not near the user).

[0218] In another embodiment, the activation means may be triggered by the wearable's accelerometer, pupil and head shaking sensors, etc. For example, if the accelerometer is selected as the activation means, the accelerometer detects when the wearer (and the wearable) is idle. In this case, the unit can be in a low power or power off mode, and when the wearable is engaged (e.g., the wearable is lifted off a surface, moved, or agitated), such engagement is recognized by the accelerometer, which switches the wearable to a power on mode. Similarly, if, for example, a pupil or head shaking sensor is used as the activation means, the power management system includes the functionality to put the unit into a battery conservation mode (e.g., a low power mode). For example, if the wearer closes their eyes while stationary on the wearable "in place", the sensor will react to new movement and immediately return the system to a power on state, etc., if / when the user eventually emerges from a period of inactivity.

[0219] In another embodiment, the wake-up means can include power-on activities programmed from a biopotential analog front-end (AFE), including galvanic skin sensor response applications including sweat, heart rate, blood pressure, temperature, etc., all of which can trigger wake-up of the wearable device.

[0220] In another embodiment, due to its IoT connectivity that allows the wearable device to communicate, activate, etc., the activation of the wearable device can be fully accessed by any type of network device / protocol.

[0221] According to further embodiments, the database can be stored on a storage device. The storage device can be selected from a fixed or mobile computer system, a portable wireless device, a smart phone, a tablet, or a combination thereof. In an alternative embodiment, the database can be stored locally on or within the wearable device. In another alternative embodiment, the database can be stored remotely, including but not limited to a cloud-based system, a secure data center behind a DMZ, etc., and the database can have encrypted and decrypted communication with the secure wearable device and its data.

[0222] Based on any of the exemplary embodiments described herein, the response time of an autistic user increases by at least about 0.5% to about 5%, about 1% to about 4.5%, about 1.5% to about 4%, about 2% to about 3.5%, preferably about 3% after the warning intervention, and accuracy increases by at least about 10% to about 50%, about 15% to about 40%, about 20% to about 30%, preferably about 26% from baseline for errors of commission, where errors of commission are indicative of a user being unable to inhibit a response when prompted by a feedback device. The numerical values ​​within these ranges can be equal to any integer value or values ​​within any of these ranges, including the endpoints of these ranges.

[0223] Based on the exemplary embodiments described herein, the response time of a neurotypical user increases by at least about 0.5% to about 50%, about 5% to about 40%, about 10% to about 30%, about 15% to about 20%, preferably about 18% after the alert intervention, and accuracy increases by at least about 0.01% to about 5%, about 0.05% to about 4%, about 1% to about 3%, preferably about 2.0% from baseline for errors of commission, where errors of commission are indicative of a user being unable to inhibit a response when prompted by a feedback device. The numerical values ​​within these ranges can be equal to any integer value or values ​​within any of these ranges, including the endpoints of these ranges.

[0224] Based on the exemplary embodiments described herein, the response time of an autistic user increases by at least about 0.5% to about 50%, about 1% to about 40%, about 10% to about 30%, preferably about 20% after guidance intervention, and accuracy increases by at least about 0.5% to about 30%, about 1.0% to about 20%, about 5% to about 15%, preferably about 10% from a baseline of error of commission, where error of commission is indicative of a user being unable to inhibit a response when prompted by a feedback device. The numerical values ​​within these ranges can be equal to any integer value or values ​​within any of these ranges, including the endpoints of these ranges.

[0225] Based on the exemplary embodiments described herein, the response time of an autistic user increases by at least about 0.01% to about 5%, about 0.05% to about 4%, about 1% to about 3%, preferably about 2% after the guidance intervention, and accuracy increases by at least about 10% to about 50%, about 15% to about 45%, about 20% to about 40%, preferably about 30% from baseline for errors of omission, where errors of omission are an indication of a user's failure to take appropriate action when a prompt is not received from a feedback device. The numerical values ​​within these ranges can be equal to any integer value or values ​​within any of these ranges, including the endpoints of these ranges.

[0226] Based on the exemplary embodiments described herein, the response time of an autistic user increases by at least about 0.5% to about 30%, about 1.0% to about 20%, about 5% to about 15%, preferably about 10% from baseline after a filter intervention for errors of omission, where errors of omission are indicative of a user's failure to take appropriate action when a prompt is not received from a feedback device. The numerical values ​​within these ranges can be equal to any integer value or value within any of these ranges, including the endpoints of these ranges.

[0227] Based on the exemplary embodiments described herein, the response time of an autistic user is at least about 0.5% to about 30%, about 1.0% to about 25%, about 5% to about 20%, and preferably about 15% faster than a neurotypical user after filter intervention for errors of omission, where the errors of omission are indicative of the user's failure to take appropriate action when no prompt is received from the feedback device. The numerical values ​​within these ranges can be equal to any integer value or values ​​within any of these ranges, including the endpoints of these ranges.

[0228] Based on the exemplary embodiments described herein, the response time of an autistic user is at least about 0.5% to about 50%, about 1% to about 40%, about 10% to about 30%, preferably about 20% faster after guidance intervention, and accuracy is at least about 0.5% to about 30%, about 1.0% to about 20%, about 5% to about 15%, preferably about 8% higher than a neurotypical user with respect to error of commission, where error of commission is indicative of a user being unable to inhibit a response when prompted by a feedback device. The numerical values ​​within these ranges can be equal to any integer value or values ​​within any of these ranges, including the endpoints of these ranges.

[0229] Based on the exemplary embodiments described herein, the accuracy of an autistic user is at least about 0.5% to about 50%, about 1% to about 4%, about 10% to about 30%, and preferably about 25% higher than a neurotypical user following a warning intervention for an error of commission, where the error of commission is indicative of a user being unable to inhibit a response when prompted by a feedback device. The numerical values ​​within these ranges can be equal to any integer value or values ​​within any of these ranges, including the endpoints of these ranges.

[0230] According to a further embodiment, the acoustic transducer may be a micro-microphone, which may preferably be surface mounted on the outside of the wearable device.

[0231] According to further embodiments, the wearable device may include at least two auditory transducers, and the placement of the first and second auditory transducers may be as known in the art, including, but not limited to, the first and second auditory transducers positioned at angles ranging from about 45° to about 135°, about 55° to about 130°, about 65° to about 125°, about 75° to about 120°, about 85° to about 120°, about 95° to about 115°, about 100°, about 110°, etc. The numerical value of any particular angle within these ranges may be equal to any integer value or value within any of these ranges, including the endpoints of these ranges.

[0232] According to further embodiments, the galvanic skin sensor can be surface mounted on an inside of the wearable device, where the galvanic skin sensor can be in direct contact with the skin of the user. The inside of the wearable device can be the skin-facing side or substantially the skin-facing side.

[0233] According to a further embodiment, the inertial movement unit may be mounted internally, preferably inside the wearable device.

[0234] According to further embodiments, the one or more feedback devices can be selected from one or more haptic drivers, one or more bone conduction transducers, or a combination thereof. The haptic drive can be mounted internally inside the wearable device. The haptic drive can be mounted internally inside the wearable device and behind the inertial motion unit. The haptic drive provides vibration patterns in response to sensory input stimuli selected from eye tracking, pupil measurement, auditory cues, nociceptive awareness, body movement, body or ambient temperature fluctuations, pulse rate, respiration, or combinations thereof. In another exemplary embodiment, the feedback device can also include a head-up visual component or other feedback device that provides pupil projection, distracting visual blurring, removal, squashing, recoloring, or combinations thereof.

[0235] According to further embodiments, the stereoscopic bone conduction transducer can be surface mounted inside the wearable device, and the stereoscopic bone conduction transducer can be in direct contact with the user's skull. The stereoscopic bone conduction transducer provides auditory tones, pre-recorded auditory guidance, real-time filtering, or combinations thereof, in response to sensory input stimuli selected from eye tracking, pupillometry, auditory cues, receptive sensations, body movements, body and ambient temperature variations, pulse rate, respiration, or combinations thereof.

[0236] According to further embodiments, the wearable device may further comprise an intervention means for providing the user with relief from distraction, inattention, anxiety, fatigue, sensory issues, or a combination thereof, the intervention means being selected from an alert intervention, a filter intervention, a guidance intervention, or a combination thereof. In an exemplary embodiment, the intervention means and the feedback means may be the same or different.

[0237] Various possible intervention measures available to the user and delivered by the wearable device are shown in the block diagram of Fig. 16. As shown in Fig. 16, according to the sensor(s) data stream delivery and the comparison of the real-time data with ecological / environmental and physiological / psychophysiological thresholds by the microprocessor 312, events deemed eligible for intervention processing can be delivered to one of two separate (or simultaneous) components: the haptic driver 313 or the bone conduction transducer 314. Depending on the wearer's previously defined preferences (stored in the microprocessor), one of four intervention strategies can be invoked: warning, filter, guidance, or combination.

[0238] Alerting Intervention: In case of activation of ecological and / or physiological / psychophysiological thresholds corresponding to the wearer's preferences, a signal is delivered to (i) a haptic driver that provides a gentle, haptic vibration pattern to communicate to the wearer that concentration, anxiety, fatigue or related characteristics require attention, and / or (ii) a bone conduction transducer that delivers an auditory / sonic message (e.g., pre-recorded voice synthesis, beeps, etc.) that augments the haptics with a set of auditory interventions and instructions.

[0239] Filter Intervention: In case of ecological and / or physiological / psychophysiological threshold activation that corresponds to the wearer's preferences and requires auditory filtering, digital audio signal processing delivers real-time and low latency audio signals with corrected amplitude (compression, expansion), frequency (dynamic, shelving, low / high cut, and parametric equalization), spatial realignment (relocation, stereo to mono) and / or phase correction (time delay, comb filtering, linear phase matching). Typically delivered to a bone conduction transducer, these can be delivered to any wireless or wired hearing device, including but not limited to earphones, earbuds, headphones, etc.

[0240] Guidance Intervention: Similar to warning interventions, guidance is provided to the wearer with step-by-step instructions for instruction on focus realignment, head shaking, pupil activity, anxiety, and fatigue. These pre-recorded text-to-speech audio streams are delivered to the bone conduction system, which provides the step-by-step instructional interventions privately and unobtrusively.

[0241] Combined Intervention: Selectable by the wearer, a combination of warning, filter and guidance interventions are provided depending on the triggering mechanism. For example, only sonic disturbances are addressed by filter intervention, while all other issues (attention-focus, anxiety, etc.) can be intervened by haptics, text voice warnings and long-form step-by-step guidance.

[0242] According to further embodiments, the wearable device may further include a power switch. The power switch may be located on the left side of the wearable device and / or on the right side of the wearable device. The power switch may be a recessed power switch.

[0243] In one embodiment, the wearable device may have a structure as shown in FIG. 1. As shown in FIG. 1, the wearable device 10 may be in the form of a eyeglass frame including a rim 109, left and right ear pieces having a temple portion 106 and a temple tip portion 108, respectively, and a screw 103 and a hinge 104 connecting the ear pieces to the rim 109. The frame may further include lenses 101, a nose pad 102, an end piece 107, and a bridge 105 connecting the left and right sides of the frame. The wearable device may have one or more sensors connected to the frame, including an infrared pupil measurement sensor 204, a galvanic skin sensor 205, an inertial motion unit 206, a wireless transceiver and A / D multiplexer 208, a microphone 201, etc. The wearable device 10 may also include one or more feedback devices connected to the frame, including a haptic driver 203, a bone conduction transducer 202, etc. The wearable device 10 may further include an optional wireless or wired hearing device 209, and a power switch (not shown) and / or a rechargeable power source 207.

[0244] Although the wearable device is shown in FIG. 1 as a glasses frame, it should be understood that the wearable device can be implemented using different types of head mounts, such as a visor or helmet. Other exemplary embodiments of the wearable device can include, but are not limited to, wrist-worn devices, bone conduction devices, etc. Any wearable device known in the art and adaptable to the methods described herein can be used, any of which can operate in conjunction with the user interface described herein. In some cases, the wearable device can be implemented as a combination of devices (e.g., wearable glasses, rings, wrist-worn, clothing / textiles, and watches).

[0245] In some implementations, the wearable device can be communicatively coupled to a mobile device (e.g., a smartphone and / or other smart device) that provides a user interface for controlling the operation of the wearable device, cooperating with the wearable device, and / or changing settings of the wearable device. For example, FIG. 19 illustrates a wearable device system including a wearable device 10 in communication with a mobile device 20 and a data store 30. In this exemplary system, the wearable device 10 communicates with the mobile device 20 via a wireless communication network. The wireless communication network can be any suitable network that allows communication between devices. The wireless communication network can be an ad-hoc network, such as a WiFi network, a Bluetooth network, and / or a network using some other communication protocol. In some implementations, the wearable device 10 can be tethered to the mobile device 20. In some implementations, the mobile device 20 processes sensor data collected by one or more sensors of the wearable device 10. For example, the mobile device 20 can determine one or more interventions to be applied using the wearable device 10 and / or some other device based on the processed sensor data. The determination may be based on one or more sensory thresholds 31 specific to the user wearing the wearable device 10. The applied intervention may be based on one or more user-specific sensory resolutions 32 specific to the user. Although the data store storing the thresholds 31 and resolutions 32 is shown in this example as separate from the wearable device 10 and the mobile device 20, in other implementations the data store 30 may be incorporated within the wearable device 10 and / or the mobile device 20.

[0246] Prior to use, the user can begin personalizing the wearable device by identifying individual sound, visual, and physiological / psychophysiological thresholds using software embedded in the wearable device. Personalization can identify unique sensory, attentional, and anxiety / fatigue-inducing cues that the user finds particularly distracting in educational, employment, social, and typical daily activities, and can be derived from the Participant Public Information (PPI) study described herein. The user-specific thresholds are used to customize subsequent alerts, filters, and guidance the user experiences while wearing the wearable device. Once the personalization process is complete, the thresholds are transmitted to the wearable device. The personalization thresholds may be updated over time (e.g., periodically or dynamically) as the user adapts to stimuli or is presented with new stimuli.

[0247] In some implementations, the device may be configured via a mobile application (app), web-based application, or other web-based interface (e.g., website). During the personalization process, the user may be presented with a graphical or other user interface via the wearable device or via a smartphone or other device communicatively coupled to the wearable device. For example, the wearable or other device may include a processor that executes instructions that cause the device to present (e.g., display) selectable controls or selections to the user in a discrete or combined format that are used to refine the set of thresholds, alerts, filters, and / or guidance. In some implementations, the personalization process may be performed by the wearer of the device, a healthcare provider, or a caregiver of the user. For example, the personalization process may be performed by running an application instance on the wearable or other device and receiving data corresponding to input from the wearer, healthcare provider, or caregiver making the selections (e.g., telemetry / biotelemetry). In some cases, different user interfaces and options may be presented depending on whether the personalization is performed by the wearer, healthcare provider, or caregiver.

[0248] In some implementations of the personalization process, a data store associated with the wearable device can pre-store initialization templates corresponding to particular sets of thresholds (e.g., sound, visual, and / or physiological / psychophysiological) and / or alerts, filters, and / or guidance. For example, templates corresponding to a wearer who is primarily ultrasound sensitive, a wearer who is primarily visually sensitive, a wearer who is primarily interoceptively sensitive, a combination wearer, etc. can be pre-configured and stored by the system. During initialization of the device, the wearer can select one of the templates (e.g., the user is primarily visually sensitive), and the configured parameters of the selected template (e.g., thresholds, alerts, filters, and / or guidance) can be further customized in response to additional user input. The additional user input can include responses to questions or selections of preferences, as further described below.

[0249] In some implementations, each of the templates can be associated with a trained model that generates one or more outputs (e.g., alerts, filters, guidance, sonic feedback, visual feedback, haptic feedback) experienced by the user given a set of inputs (e.g., sensor readings from the wearable device, user thresholds, etc.). The model can be trained and tested with anonymized historical data associated with the user to predict the appropriate output given the sensory inputs and thresholds. The model can be constructed using supervised, semi-supervised, or unsupervised learning. During the personalization process described further below, the parameters of the model (e.g., weights of input variables) can be adjusted according to the user's sensitivities and / or preferences.

[0250] In some implementations of the personalization process, the user may be presented with selectable preferences and / or answers to questions. For example, the personalization process may present the user with selectable options regarding demographics (e.g., gender, age, education level, handedness, etc.) and sensitivity (e.g., audio preferences, visual preferences, physiological / psychophysiological preferences, alert preferences, guidance preferences, intervention ranking preferences, etc.). Depending on the user's selection, a particular set of thresholds (e.g., audio, visual, and / or physiological / psychophysiological), alerts, filters, and / or guidance may be customized and stored for the user. For example, based on the user's selection of audio preferences, the system may be configured to perform digital signal processing of the audio signal before the audio is played to the user to adjust the energy of different frequency ranges (e.g., bass, midrange, treble, etc.) within the audible frequency band (e.g., 20 Hz to 20,000 Hz), the audio channel from which the sound originates, or other characteristics of the audio. During configuration, the user may specify preferences for filtering (e.g., enhancing, removing, or otherwise modifying) low, mid, high, bass, loud, reverberant, surround sounds, etc. As another example, the user may specify preferences for receiving alerts of sounds with particular sonic characteristics (e.g., alerting of loud, reverberant, and / or ambient sounds before the sound occurs). As a further example, the user may prefer that guided sounds have particular characteristics (e.g., soft speech, gentle sounds) when the user becomes anxious, distracted, or sensitive.

[0251] In some implementations, the user's selected preferences and / or answers during personalization can be used to build a model that generates one or more outputs (e.g., alerts, filters, guidance, sonic feedback, visual feedback, haptic feedback) experienced by the user given a set of inputs (e.g., sensor readings from a wearable device, user thresholds, etc.). For example, a website or mobile app component accessed via a user login can generate one or more tolerance scores based on the user's answers to questions related to visual, auditory, or physiological / psychophysiological stimuli. The one or more tolerance scores can be used to initialize the model. The model can also be initialized, modified, and / or monitored by an expert, healthcare provider, or caregiver.

[0252] During personalization, the user may rank types of interventions. For example, the user may rank and / or specify a preferred type of alert (e.g., beep, haptic, voice, or some combination thereof), a preferred audio filter (e.g., volume (compression, limiting), equalization (tone, formula), noise reduction, imaging (panning, phase), reverberation (echo), imaging (panning, phase), or some combination thereof), a preferred type of guidance (e.g., encouragement), etc.

[0253] In some implementations, one or more sensors of the wearable device can be calibrated during initialization of the device. The user can be presented with an interface to calibrate the sensors and / or adjust sensor parameters. For example, the user can specify whether all or only some sensors are active and / or collect data, adjust sensor sensitivity, or adjust sensor thresholds (e.g., brightness for optical sensors, volume for audio sensors) and the order in which they are desired to be implemented (e.g., first alerts, then guidance, then filters, etc.).

[0254] In some implementations, after an initial set of user-specific thresholds and warnings, filters, and guidance have been configured for the user, validation of the configuration may be performed by presenting external stimuli to the user and providing warnings, filters, and / or guidance according to the user-configured thresholds. Depending on the user's response, additional configuration may be performed. This validation process may also adjust sensor settings, such as sensor sensitivity.

[0255] In embodiments where trained models are used to provide alerts, filters, and / or guidance, the models can be retrained over time based on collected environmental and / or physiological / psychophysiological data. Retraining can be performed overnight and / or when the system is not in use to conserve computational resources and / or device battery life.

[0256] FIG. 20 illustrates an operational flow diagram illustrating an example method 400 for initializing and iteratively updating one or more sensory thresholds and one or more interventions associated with a particular user. In some implementations, the method 400 can be implemented by one or more processors of a wearable device system (e.g., one or more processors of the wearable device 10 and / or the mobile device 20) executing instructions stored on one or more computer-readable media (e.g., one or more computer-readable media of the wearable device 10 and / or the mobile device 20). Operation 401 includes presenting a plurality of selectable templates to a user, the plurality of templates corresponding to one or more sensory thresholds and one or more interventions. The plurality of selectable templates can be presented via a GUI (e.g., using the wearable device 10 and / or the mobile device 20). Operation 402 includes receiving data corresponding to an input by a user selecting one of the templates. After the user selects one of the templates, the one or more sensory thresholds and one or more interventions associated with the templates can be associated with the user. For example, the one or more sensory thresholds and the one or more interventions may be stored in data store 30 that includes an identification and / or user profile corresponding to a user. As discussed above, operations 401-402 may be performed during and / or after an initialization process.

[0257] Act 403 includes receiving data corresponding to a user input selecting a preference. The preferences may include audio preferences, visual preferences, physiological / psychophysiological preferences, alert preferences, guidance preferences, and / or intervention preferences. Act 404 includes modifying one or more sensory thresholds and one or more interventions associated with the user in response to receiving additional data corresponding to additional user input selecting a preference. For example, a data store thresholds and interventions may be updated. As shown, acts 403-404 may be repeated over time as the user desires to further define thresholds / interventions and / or as the user develops new preferences.

[0258] Act 405 includes collecting sensor data and environmental data while the user is wearing the wearable device. Act 406 includes modifying one or more sensory thresholds and one or more interventions associated with the user in response to collecting the sensor data and / or environmental data while the user is wearing the wearable device. As shown, acts 405-406 can be repeated over time as the user utilizes the wearable device system to provide sensory relief. The frequency at which the one or more sensory thresholds and one or more interventions are updated in response to newly collected data can be configurable, system-defined, and / or user-defined. For example, the updates can depend on the amount of data collected and / or the amount of time that has elapsed. In some implementations, acts 405-406 can be skipped. For example, the user can disable updating the thresholds and / or interventions based on actual use of the wearable device.

[0259] In some implementations, the wearable device can be configured to communicate with and / or control the IoT device that presents the stimuli. For example, based on thresholds configured for the user, the wearable device can control the operation of smart devices such as networked hubs, networked lighting devices, networked outlets, alert systems, networked thermostats, networked sound systems, networked display systems, networked appliances, and other networked devices associated with the user. For example, the audio output (e.g., volume and balance) of a networked sound system and / or the display output (e.g., brightness, contrast, and color balance) of a networked display system can be altered to meet individual sound or visual thresholds. To synchronize communication and operation between the wearable device and the IoT device, the device can be linked to a user's account, which can be configured via an application running on a smartphone (e.g., a native home control application) or other device (e.g., mobile device 20). In some cases, the behavior or one or more scenes of the IoT device can be pre-configured based on thresholds associated with the user. The behavior or scene can be activated when the wearable device detects that the IoT device is in the presence (e.g., in the same room) of the IoT device.

[0260] As an example, FIG. 21 shows a wearable device system including a wearable device 10 in communication with a mobile device 20 that controls an IoT device 40 having a speaker 41. As another example, FIG. 22 shows a wearable device system including a wearable device 10 in communication with a mobile device 20 that controls an IoT device 50 having a light emitting device 51. In operation, the wearable device 10 can use one or more sensors to collect sensory input stimuli. The sensory input stimuli can be transmitted to the mobile device 20, which compares the sensory input stimuli to one or more sensory thresholds specific to the user (e.g., threshold 31) to determine an intervention to be provided to the user to provide relief from distraction, inattention, anxiety, fatigue, and / or sensory issues.

[0261] 21, the sensory input stimulus may be generated at least in part due to a sound emitted by the speaker 41 of the IoT device 40. For example, a user may generate a physiological / psychophysiological response to music and / or other sounds being played at a particular frequency and / or frequency range by the speaker 41. In that scenario, the intervention may include the mobile device 20 controlling the IoT device 40 to filter the audio signal in the frequency domain such that the sounds output by the speaker 41 are played at frequencies that do not elicit the same physiological / psychophysiological response in the user.

[0262] In the example of FIG. 22, the sensory input stimulus may be generated at least in part due to light emitted by the light emitting device 51 of the IoT device 50. For example, if the output light is too bright or too cool in color temperature (e.g., >4000K), it may cause discomfort to the user. This discomfort may be measured using the sensory input stimulus collected by one or more sensors of the wearable device 10. In that scenario, the intervention may include the mobile device 20 controlling the IoT device 50 to filter the light signal of the lighting device 51 to reduce the brightness and / or color temperature of the light output by the lighting device 51.

[0263] Although the above examples show the mobile device 20 as communicating with and controlling the IoT devices 40, 50, it should be understood that these functions may be performed by the wearable device 10 instead.

[0264] The wearable device may include various user interface components, including but not limited to a mobile phone, laptop, tablet, desktop computer, etc., and may use any user interface known in the art. In some implementations, the wearable device may be synchronized with a smartphone. For example, the wearable device may be configured to accept calls, adjust call volume, provide notification sounds or vibrations, provide ringtones, etc. The wearable device may be granted access to user contacts, text messages or other instant messages, etc. In some cases, the intensity of sounds or vibrations, or patterns of sounds or vibrations provided via mobile integration may depend on user configured thresholds. Initial configuration and personalization of the wearable device may be performed via an application installed on the smartphone or other device.

[0265] The wearable device may include one or more network interfaces (e.g., WiFi, Bluetooth, cellular, etc.) for communicating with other network devices and / or connecting to the Internet. For example, a WiFi interface may enable the wearable device to select and communicatively couple to a local network, which may enable communication with IoT devices. Bluetooth may enable pairing between the wearable device and a smartphone or other device.

[0266] A wearable device may include or be communicatively coupled to one or more data stores (e.g., memory or other storage devices) that are accessed during its operation. The storage may be local, on a network, and / or on the cloud. The storage may maintain records of user preferences, user performance, trained models, and other data or instructions necessary to operate the device.

[0267] In an exemplary embodiment, the wearable device operates as described herein. The wearable device may remain in a passive, i.e., non-operational, mode before being worn by a user. This allows for optimized battery life.

[0268] Once in the active mode, i.e., when the wearable device is in the operating mode, the wearable device detects and responds to one or more sensory cues selected from a myriad of sensory cues received and detected by one or more sensors disposed on the wearable device. Such sensory cues may include environmental signals and physiological / psychophysiological signals, etc. The wearable device also responds to the sensory cues via warnings, filters, and guidance to provide the user with additional appropriate resolution whenever a personalized threshold for use is exceeded. Thresholds and interventions can be iteratively set, adjusted, muted, and canceled at any time during the user's use of the wearable device by returning to the computer / application.

[0269] Various types of sensory cues can be received and detected by the wearable device, including, but not limited to, visual, auditory, and physiological / psychophysiological cues. In an exemplary embodiment, visual distractions can be detected via eye tracking and pupil measurement monitored with an infrared sensor that can be surface mounted on the inside of the wearable device, such as at the intersection of the frame rim / right hinge temple, and pointed at the user's eyes, such as the right or left eye or both. In an exemplary embodiment, auditory distractions and hearing thresholds can be monitored by a miniature wired electret microphone that can be surface mounted on the outside of the end piece, near the intersection of the frame front and temple. In an exemplary embodiment, physiological / psychophysiological distractions, interoceptive thresholds, and user's head shaking can be monitored by a galvanic skin sensor that is surface mounted on the inside of the left earpiece and in direct contact with the skin just above the user's neckline, and / or an inertial motion unit that can be internally mounted on the inside of the wearable device and positioned behind the earpiece. The various detection components described herein are merely exemplary, and any suitable components may be used.

[0270] Various types of resolutions (interventions or digital mediations) can be provided to the user in response to sensory cues received by the wearable device. Resolutions can include, but are not limited to, visual, auditory, and physiological / psychophysiological resolutions. In an exemplary embodiment, visual resolutions can be delivered via haptic drivers that can be mounted internally inside the earpiece, behind the inertial movement unit intersection of the frame rim / right hinge temple. If pupil or inertial / head shaking thresholds are detected, visual resolutions can be provided via inherent vibrations associated with optical distraction. In another exemplary embodiment, visual alerts can be delivered by stereoscopic bone conduction that can be surface mounted inside the wearable device, for example, on both temples between the hinge and temple tips, making direct contact with the user's left and right skull in front of each ear, and providing either a beep and / or pre-recorded audio guidance if pupil or inertial / head shaking thresholds are detected. In another exemplary embodiment, auditory resolution can be delivered to the user through a single haptic driver by providing a uniquely coded vibration alert when an acoustic threshold is detected, and / or through a bone conduction transducer that provides both beeps, pre-recorded audio guidance, and / or real-time filtering using digital signal processing (DSP) for distracting environmental auditory events (e.g., compression, equalization, noise reduction, spatial panning, limiting, phasing, and gating) when an acoustic threshold is detected and can be processed according to the user's personalization settings. For example, in an exemplary embodiment, a real-time digital audio stream recorded by a microphone connected to the wearable device provides audio data to a microprocessor that undergoes system manipulation to achieve a predetermined goal. As previously mentioned, the DSP generates feedback in the form of modified audio signals (filtered interventions) that improve volume (amplitude, compression, noise reduction), tone (equalization), directionality (spatial, etc.). As another example, the guidance can include one or more tonal alerts retrieved from a data store.

[0271] In an exemplary embodiment, the device can be configured to boost certain audio frequencies depending on the user's age or hearing ability. For example, the device can boost low, medium, and / or high frequencies depending on the user's age and / or hearing profile. In some cases, the device can execute instructions to provide a hearing test to generate a hearing profile. Controls can be provided to enable or disable sound boosting.

[0272] In an exemplary embodiment, physiological / psychophysiological resolution can be provided to the user via the haptic drivers described above, and by providing uniquely coded vibration alerts when physiological / psychophysiological, anxiety, fatigue or other sensory reception thresholds are detected, and / or via bone conduction transducers, which provide both beeps and / or pre-recorded audio guidance for similar threshold alerts and guidance. The various resolution components described herein are merely exemplary and any suitable components may be used.

[0273] In an exemplary embodiment, the wearable device may also include an internally mounted central processing unit that may further include a miniature printed circuit board combined with a built-in connected rechargeable power source, with a wireless transceiver and analog / digital multiplexer present in both earpieces to provide an evenly weighted distribution to the wearer.

[0274] As described herein, the comparison means compares the sensory input stimuli recorded by the one or more sensors to a database of one or more user-specific sensory thresholds to obtain the user's sensory resolution. The comparison means performs the aforementioned functions as follows:

[0275] The user-specific thresholds can be obtained by having the user complete a decision tree style survey (similar in scope to the survey described in the Sustained Attention Paired Response Test (SART) protocol described herein) and then the microprocessor measures the user-specific thresholds against the ecological and physiological / psychophysiological data streams to provide appropriate intervention assistance. In an exemplary embodiment, the user-specific thresholds can be dynamically altered using machine learning capabilities.

[0276] An exemplary embodiment of how input stimuli are compared to stored data to generate user-specific interventions is illustrated via a block diagram in FIG. 16, although the application is not limited thereto. In the exemplary embodiment illustrated in FIG. 16, six components make up the input section of the wearable device, including an optical module 301, an inertial measurement unit (IMU) 304, an audio sensor 305, a galvanic module 306, a temperature sensor 309, and a biopotential analog front-end (AFE) 310. In combination, these components provide both ecological (environmental) and physiological / psychophysiological data to a sensor hub 311 (multiplexer), where the data is processed by the system's microprocessor (e.g., ARM Cortex) for rapid analysis and comparison with existing thresholds, characteristics, and user preferences (typically via wireless bidirectional communication, but can also be directly connected). The microprocessor 312 (e.g., ARM microprocessor) then delivers appropriate commands for intervention activities that are processed by the associated system components described herein. The six components are further described below.

[0277] The optical module 301 includes (i) a pair of inward-facing infrared sensors 302 that monitor pupillary responses, indicative of a user's lapses in focus and attention, and (ii) a single outward-facing sensor for determining ecological / environmental cues of a visual nature. The tuned optical AFE 303 provides an appropriate pupil data stream for processing and simultaneously provides an environmental data stream for image recognition, allowing the microprocessor 312 to determine the visual environmental cues to which the user is responding. In either case, image recognition (pupil responses, saccades, computer screens, books, roadways, office / academic environments, etc.) relies on computer vision techniques that allow the microprocessor 312 to interpret and classify what is seen in the visual data stream. This type of image classification (or labeling) is the core task and foundational component in comparing real-time visual input to a library / catalog of pre-labeled images that are interpreted and, if they exceed (or do not exceed) the user's threshold, are the basis for intervention.

[0278] The IMU 304 measures and reports specific forces of the body (in this case the user's head / face). The IMU 304 also provides angular velocity and orientation using a combination of accelerometers, gyroscopes, and magnetometers to deliver a data stream regarding the user's head movement and attention when compared and contrasted to the optical AFE 303 and similarly processed against pre-labeled and classified data. In some implementations, the IMU can include a 3-axis gyroscope / accelerometer.

[0279] Similar to the optical module 301, the audio sensor 305 provides an environmental data stream of sonic nature that can be compared to known auditory signatures that are labeled and available for computer microprocessing. Auditory signatures that reach user-selected thresholds of frequency, amplitude, spatial, time delay / phase, and the like can then be delivered for intervention processing.

[0280] Both the galvanic module 306 and temperature sensor 309 provide physiological / psychophysiological and ambient / physiological data streams measuring the wearer's electrodermal activity (EDA), galvanic skin response (GSR), body and ambient temperature. These are utilized in conjunction with the biopotential AFE 310 to provide real-time, continuous monitoring of the wearer's electrical skin properties, heart rate, respiratory rate, and blood pressure detection. As with the sensors previously described, all are time-stamped / synchronized for microprocessor processing, analysis, labeling / comparison, and interventional activation.

[0281] The biopotential AFE 310 provides electrocardiogram (ECG) waveforms, heart rate, and respiration which are then sent to the microprocessor 312 to assist with the user's physiological / psychophysiological state, processing, and attention / anxiety / fatigue interventions.

[0282] An additional block diagram providing additional microprocessor details (ARM processor) is shown in FIG.

[0283] The catalog of user-specific cues and resolutions can be stored in a database stored in and communicating with software executed from the wearable device and the control program / app, and is available for machine learning purposes to provide applications and hardware with an increasingly greater understanding of the user's environment and physiological cues, alerts, filters, and guidance. Artificial intelligence (AI) algorithms continuously process the user's personalization, input cues, and custom-created resolutions to more narrowly and accurately predict and respond to physiological / psychophysiological inputs and responses. This machine learning and AI algorithms increase user training, promoting greater autonomy, comfort, attention, focus, and mental well-being. The catalog is available for analysis by the user and experts, and data streams and progress reports are available for clinical research, physician / telemedicine, evaluation, and further review.

[0284] In some implementations, the preferences of the wearable device may be modified by the user to optimize the battery life of the device. For example, the device may be configured to operate in a power saving mode that conserves battery life by reducing the sensitivity of the sensors, limits power for unused operations, or operates to maximize battery life. Alternatively, the user may have the option to select an extended processing mode that emphasizes processing (e.g., making the sensors more sensitive) but uses more battery per unit time.

[0285] In some implementations, the wearable device can be associated with an application that provides diagnostic data regarding the user, the system, or the caregiver / healthcare professional. For example, user diagnostic data can include user preferences, user responsiveness, and generated problems and warnings. System diagnostic data can include environmental and device responsiveness, and problems and warnings. Caregiver / healthcare professional diagnostic data can include user efficacy performance (e.g., sonic, visual, or interoception) and any areas of interest such as wearer guidance or device guidance.

[0286] As alluded to above, real-time filtering of audio signals can be performed in response to collecting sensory data from one or more sensors of the wearable device. As opposed to adjusting the time or overall amplitude of the signal experienced by the listener, this filtering can be performed in the frequency domain, affecting at least the center frequency (Hz), cut or boost (dB), and / or width (Q). For example, any low frequency hum associated with real-time detection of machinery and / or light ballasts in the environment can be eliminated and / or reduced, minimized and / or mitigated. This can be performed by adjusting the problematic and nearby frequencies in real-time, either with a band-pass or low-cut high-pass filter, with specific adjustments fine-tuned to the user's personalized profile. In some implementations, audio filtering can also be applied in additional domains including time, amplitude, and spatial positioning (e.g., to filter distracting sounds modulating from a given direction).

[0287] Although some implementations have been described primarily in the context of modifying and / or filtering distracting sounds (i.e., audio interventions), the techniques described herein can implement a similar set of interventions related to visual stimuli, either separately or in combination with other types of stimuli. For example, interventions such as warnings, guidance, and / or combinations can be implemented without filtering mediation. As described further below, visual interventions can be based on pupil response, accelerometer, IMU, GSR detection, and / or video of the wearer's environment. In some implementations, identified visual interventions can work in concert with audio modifications.

[0288] FIG. 23 illustrates an exemplary wearable device 500 that can be utilized to provide a visual intervention, according to some embodiments of the present disclosure. As illustrated, the wearable device 500 can include sensors and / or transducers of the wearable device 10. To support certain visual interventions, the wearable device 500 also includes a camera 550 and a display 551. Thus, the wearable device 500 is implemented as a wearable HMD. Although an eyeglass form factor is shown, the HMD can be implemented in a variety of other form factors, such as, for example, a headset, goggles, visor, combinations thereof, etc. Although shown as a binocular HMD, in some embodiments, the wearable device can be implemented as a monocular HMD.

[0289] Display 551 may be implemented as an optical see-through display, such as a transparent LED and / or OLED screen that uses wave guides to display virtual objects overlaid on the real-world environment. Alternatively, display 551 may be implemented as a video see-through display that supplements video of the user's real-world environment with the overlaid virtual objects. For example, virtual objects may be overlaid on video captured by a camera 550 aligned with the field of view of the HMD.

[0290] The integrated camera 550 can capture video of the environment from the perspective of the wearer / user of the wearable device 500. Thus, as described further below, the camera's live video / image feed can be used as one input for detecting visual objects to which the user is potentially visually sensitive and triggering visual interventions.

[0291] In some implementations, real-time overlay interventions can be implemented in which visual objects and / or light interruptions are suppressed, squashed, minimized, mitigated, and / or otherwise removed from the wearer's field of vision. In some implementations, the transducer components of the system (e.g., microphones and / or outward-facing optics) can be used in concert with on-board biological sensors and / or projection technology to train / detect, analyze / match / predict, and / or modify optical cues and / or visible items that correlate to the wearer's visual sensitivity, attention, fatigue, and / or anxiety thresholds. In some implementations, visual, optical, and / or related scene clutter can be filtered in real time, and the wearer is not aware that it is distracting.

[0292] In certain embodiments, one or two types of Real-Time Optical Enhancement (REOPEN) algorithms can be implemented to detect, predict, and / or modify visual inputs that reduce distraction / mental health issues and increase attention, calm, and / or focus. The algorithms can provide (i) editing of visual scenes, images, and / or objects, and advanced notification of distracting optics that match the user's visual profile, and / or (ii) live correction of visual distractions without prior notification in real time. Interventions that can be delivered in real time using the REOPEN algorithm are illustrated by Figures 24A-C, which are further described below.

[0293] In one embodiment, a real-time optical enhancement and visual a priori intervention algorithm (REOPEN-VAIL) may be implemented to train / detect, analyze / match / predict, and / or modify visual items that the user considers distracting (e.g., based on the above and / or created personalized preference profile) and compare these to previous and / or current physiological / psychophysiological responses to the environment. Upon detecting a threshold crossing and / or match between interoceptive reactivity (egocentric) and detection of visual cues (egocentric video), REOPEN-VAIL may provide iterative analysis, training, augmentation, context modification, and / or advanced warning of optical distractions prior to the wearer's ability to sense these visual and / or associated physiological / psychophysiological cues.

[0294] In one embodiment, a visual post-intervention algorithm (VASILI) can be implemented using multimodal learning methods to train / detect, analyze / match, and / or correct visual items that the user previously deemed distracting (e.g., based on previously described or created preferences, previous and / or current physiological / psychophysiological responses, etc.). In contrast to REOPEN-VAIL, in the case of VASILI, the optical system provides a situation change without advanced warning of the distraction in the form of an intervention delivered in real-time after the user is exposed to a visual distraction and after identification of either the ecological and / or wearer's physiological / psychophysiological cues systems as part of an iterative process that can serve as the basis for future training, sensing, and / or a priori algorithms.

[0295] Using the REOPEN algorithm, various interventions may be possible in real time. For example, as shown in FIG. 24A, in response to detection of a particular visual object (e.g., a distracting visual subject and / or a visual anomaly), a haptic alert, a tone alert, a guidance alert, a combination thereof, etc. may be delivered to inform the wearer. The guidance alert may provide user-selectable verbal instructions of an anticipated visual distraction and / or guidance to intervene with continued focus, calm, and / or attention. The aforementioned guidance alerts may be implemented as visual and / or text-based guidance visible to the wearer via a displayed (e.g., using display 551) user-selectable system visible to one and / or both eyes and / or gaze to intervene with continued focus, calm, attention, a combination thereof, etc. (e.g., FIG. 24B).

[0296] In some implementations, visual distractions (e.g., certain objects, faces, etc.) can be rendered to appear opaque and / or unclear. This blurring effect can blur identified, distracting, and / or otherwise annoying images as a user-selectable and / or predefined intervention that affects what the wearer sees (e.g., FIG. 24B).

[0297] In some implementations, the visual scene can be rendered with a modified background to eliminate visual distractions and / or identified images. The system can interpolate nearby images to distracting objects and / or replicate the background by overlaying a series of "stitched" images that naturally hide and / or suppress the sensory impact of the problematic optics, all of which are user-selectable (e.g., FIG. 24A).

[0298] In some implementations, predetermined user-selectable pictograms and / or placeholder images can be rendered that camouflage distracting optics and / or visual clutter (eg, FIG. 24B).

[0299] In some implementations, visual distractions may be rendered with a modified color palette and / or associated pigmentation may be modified to a user's preferences to reduce effects such as distraction, sensitivity, concentration, anxiety, and / or fatigue, and / or combinations thereof (e.g., FIG. 24C).

[0300] In some implementations, visual distractions can be rendered as muted and / or altered visuals (e.g., FIG. 24C) with edited brightness and / or sharpening of user-selectable images.

[0301] In some implementations, visual distractions can be rendered at an edited size such that the image is expanded and / or altered to make it more prominent, larger, and / or highly visible (e.g., FIG. 24C).

[0302] The principles of the present invention, including certain exemplary features and embodiments, and advantages thereof, are illustrated with reference to the following non-limiting examples.

[0303] FIG. 25 illustrates one particular example of a workflow using the REOPEN algorithm to provide real-time intervention in a scenario where there is a singular distracting visual source (e.g., a bird flying across the sky, distracting an individual from their work). In this example, the algorithm is implemented using a convolutional neural network (CNN). As shown, the distracting stimulus can be visualized by an individual wearing a wearable device (e.g., wearable device 500) that captures the cues and then processes and trains the data (e.g., environmental and / or psychophysiological). The convolutional layers of the CNN are formed and / or iteratively examined, for example, visual data that causes distraction or physiology such as pupil movement including edges, shape, and / or directional movement. The connection layer digitizes the traditional convolutional layer. This can be repeated for multi-modal data types such that a learned state of focused activity can be recorded when the layers correlate with a predefined "eyes on target". Conversely, if pupil movement is not coordinated with the target and / or activity, a separate learned event can be stored and / or tagged as non-focused activity, resulting in digital mediation being delivered until a "focused" state is observed.

[0304] For example, in the case of the REOPEN-VAIL algorithm, the a priori intervention may be one that has already been trained with the flow shown in FIG. 25 and then sensed by the system following similar external cues and / or early reflexes of pupil concentration lapses in anticipation. This may generate a warning before the actual prolonged individual state in an attempt to intervene before the distraction occurs. In the event of a failure of REOPEN-VAIL (e.g., the individual remains distracted for a period of time, e.g., more than about 5 seconds or as dictated by the personalized paradigm), the a posteriori flow may be repeated, this time providing a mediation consisting of a warning, guidance, and potential filtering to mute, eliminate, mitigate, minimize, and / or otherwise correct the offending distraction.

[0305] PPI research

[0306] A PPI study was conducted to identify dependency / thematic variables and dependency / demographic factors associated with wearable device usage. PPI participants included language-capable autistic and neurotypical adolescents and adults aged 15–84 years. All participants had intelligence above the normal or average range and lived largely independent lives, i.e., study participants did not fall within the general learning disability range. Participants provided health / medical conditions and disability information relevant to their opinions on distraction, concentration and anxiety, both at school and at work. Prior to testing, participants provided informed consent along with a verifiable ASC diagnosis, where applicable. All participants were invited to participate in a focus group, a user research group, or both.

[0307] The focus group included 15 participants (aged 17-43 years) who were studied for distraction and attention focus. The main task of the focus group was to comment on sensory issues and provide input into the design of the user study to ensure autism relevance and adherence to an autism-friendly format.

[0308] The user research group included 187 participants aged 18-49 years, who provided first-person perspectives on distraction and focus while gathering views and opinions on which aspects of technological aids / assistances were most welcome and had the greatest impact on sensory, attention, and quality of life issues.

[0309] Incorporated into the PPI study was the Lived Experience Attention Anxiety Sensory (LEA) questionnaire, which was developed for participants' expression and used as a warm-up point for discussing concentration, distractibility, anxiety, sensory and attention difficulties, and needs. 2 Se)). LEA 2 Participants were prompted to indicate their interests, attitudes, and opinions regarding receiving technology assistance. The study was administered online.

[0310] Pre-prototype Battery Experiment (PTBE)

[0311] One hundred and ninety-six autistic participants were recruited by opportunity sampling. After exclusions, 188 participants were left (109 males, 79 females), of which 12.2% were aged 18–20 years, 21.2% were aged 21–29 years, 60.8% were aged 30–39 years, and 5.8% were aged 40–49 years. For the purposes of PTBE, variables were designed that tapped different aspects of the experiences of individuals with autism. After identifying variables of interest, questions addressing these variables were assigned numerical variables. Some questions were not assigned to any variable and therefore were not analyzed in this study. Table 1 shows examples of the resulting variables, the number of questions that fell under each variable, and the type of questions used to investigate that variable. Each participant received a score for each of these variables, which was calculated by averaging the responses to the questions that fell under that variable. The variables were mutually exclusive (i.e., no question was included in the calculation of more than one variable). To assess the validity of the variables, the inter-item correlations of each variable were examined, and all of them had Cronbach's alpha greater than 80, indicating high internal validity.

[0312] [Table 1]

[0313] The various pilot outcomes using benign data are shown in Figures 2 and 3. These graphics report sensitivity across three modalities (visual, auditory and anxiety) along with the wearable attention of ASC participants to visually distracting stimuli. Kruskal-Wallis H tests on the test population showed the following:

[0314] Statistically significant differences in age-based sensitivity effects, anxiety tendency, distractibility index, technological tolerance, visual difficulties and physiological / psychophysiological difficulties, but no statistically significant differences in sound difficulties;

[0315] Statistically significant differences in sensitivity effects and sound problems based on gender, but no statistically significant differences in anxiety tendency, distractibility index, technology tolerance, visual problems, physiological / psychophysiological problems;

[0316] Statistically significant differences in the effects of sensitivity, anxiety tendency, distractibility index, technology tolerance, sound problems, visual problems and physiological / psychophysiological problems based on education level; and

[0317] There are statistically significant differences in sensitivity impact, anxiety tendency, distractibility index, technology tolerance, sound problems, visual problems, but no statistically significant differences in physiological / psychophysiological problems based on different employment levels.

[0318] Continued focus on response tests (SART)

[0319] The PPI study and PTBE were followed by the SART study, which included an online test designed to test for sensory issues affecting participants diagnosed or identified with ASC. Specifically, the study tested a subset of components within a wearable prototype to answer two questions: (i) is it possible to classify and predict autistic responsiveness / reactivity to auditory (ecological) disturbances and physiological / psychophysiological distractors when autistic individuals are assisted through alerts, filters, and guidance, and (ii) can the exploration of multimodal learning analytics (MMLA) combined with supervised artificial intelligence / machine learning contribute to understanding the heterogeneity of autism with high accuracy, thereby enhancing attentional focus while reducing distractibility and anxiety.

[0320] The study is based on theories of attentional schema, zone of proximal development, multimodal lecture analysis and multimodal learning analytics and utilizes both evaluator-involved and user-involved methodologies including iterative development and evaluation, early user integration, phenomena of interest, and persistent collaborative methodologies.

[0321] In an exemplary testing protocol, baseline tests and associated scores are procedurally derived both before and after the subtest, and a putative cognitive conflict occurs during the subtest that may result in a hypothesized and measurable increase in both distractibility and anxiety. At the same time, this surge is likely to pool with a decline in focus and cone-of-attention performance. Finally, during the latter subtest, the "Associate" (a human wizard) presents a collection of handcrafted warnings, filters, and guidance. These mimic the behavior of the wearable intervention by offsetting and counterbalancing distracting auditory stimuli. To mitigate fatigue effects, these interventions are counterbalanced and randomized, possibly across multiple sessions. Alternatively, a combination of warnings, filters, and guidance is provided to shorten the overall length of the experiment, as shown in Table 2 below and illustrated in Figure 4.

[0322] Delivery of isolated and displaced supports can yield broad indices of test responses. Mixed methods (qualitative and quantitative assessments) combined with overt participant behavior obtained via audio and video recordings can provide encoding and analysis with sample-accurate synchronization to the system software. Depending on sample size and time constraints, the design allows for post-hoc video analysis (e.g., participant walk-throughs) in either structured or free-form. These examinations can help facilitate recall, accuracy, and provide further understanding of anxiety and other episodic test moments.

[0323] [Table 2]

[0324] This study tests a subsystem mockup using multimodal artificial intelligence-driven (MM / AI) sensors designed to provide individualized warnings, filters, and guidance to help reduce distraction and anxiety while enhancing focus and attention by increasing the cognitive load associated with unexpected ecological and physiological / psychophysiological stimuli. The study uses a series of online experiments in which the wearable's behavior is simulated by a collaborative human operator. The study proposes a within-subject, two-condition SART using multimodal sensors, where the user's performance is measured (Robertson, IH, Manly, T., Andrade, J., Baddeley, BT, & Yiend, J. (1997). "Oh no!": Daily attention deficit performance correlates in traumatic brain injury and normal subjects. Neuropsychology, 35(6), 747-758). Importantly, the task was performed and data collected with and without the influence of distracting sonic stimuli (single modality) with various combinations of advanced warnings, audio filtering, and return-to-task guidance. This phase of the study included 40 participants, including 19 autistic and 21 non-autistic participants.

[0325] The classic SART paradigm is considered an example of both high reliability and validity, requiring participants to refrain from pressing a computer key during the on-screen appearance of a target image. This study modifies the SART by flipping the keystroke sequence; that is, rather than holding the "down" key through the majority of the test, the assigned key was pressed only when the target appeared. This resulted in greater reliability and reproducibility when testing at a distance and online (Anwyl-Irvine, AL, Massonie J., Flitton, A., Kirkham, NZ, Evershed, JK (2019). Gorilla in our midist: An online behavioral experiment builder. Behavioral research methods). Performance on the SART correlates significantly with performance on tests of sustained attention. However, studies have shown that the SART does not correlate well with other types of attention measures, "supporting the view that [the SART] is in fact an index of sustained attention" (Robertson et al., 1997, 747, 756). This study uses the SART specifically because studies such as Robertson's study support that the methodology is fundamentally unaffected by age, estimated intelligence scores, or other intellectual measures.

[0326] Furthermore, within this study, the SART task is performed and data are collected with and without the influence of distracting sonic stimuli. This modality acts as both singular and irrelevant foils when accompanied by a combination of various subtests of the Advanced Alerting, Audio Filtering, and Return to Task Guidance Model. These combinations act as interventions. The subtests of the study utilize visual search of targets against competing irrelevant foils (e.g., alphanumeric characters). Complementing these text targets with additional competing modalities (e.g., sonic foils and interventions) makes this SART study novel compared to previously conducted studies. The SART requires participants to "actively inhibit the selective activation of competing distractor and target representations. Memory factors are minimal in these tasks because the targets are simple and prominently displayed to the subject over the course of the test." (Robertson, IH, Ward, T., Ridgeway, V., & Nimmo-Smith, I. (1996). The structure of normal human attention: testing everyday attention. Journal of the International Neuropsychological Society, 2(6), 525, 526) Although reaction times and other temporal indices are taken into account in creating participant scores, this study excludes the possibility that the subtests only measure sampling speed of processing, as qualitative mental health indices are also integrated.

[0327] The initial PPI study and PTBE facilitated a deeper understanding of the everyday experiences of autistic individuals, as well as their focus, distractibility and anxiety concerns with particular regard to later life, educational and workplace experiences. The PPI study and PTBE also provided information on the potential decline of both anxiety and sensitivity as autistic individuals age, with these trends differing within a particular modality. Stability is achieved across various ages for sonic variables, but varies for both visual and physiological / psychophysiological variables. Furthermore, anxiety and sensitivity may not be gender-related. There is also a downward age-related trend for both technological tolerance and distractibility, but with variation from ages 30 to 39, possibly due to the large size of this particular sample.

[0328] The study design was based on the SART / WoZ design, involving prior manual intervention to support participant testing and online experiments in which system operation was simulated by a human operator equipped with scripts (Bernsen, N.O., Dybkjaer, H., & Dybkjaer, L. (1994). Wizard of oz prototyping: How and when. Proc. CCI Working Papers Cognit. Sci. / HCI, Roskilde, Denmark). The Wizard of Oz (WoZ) study design offers economical and rapid implementation and evaluation and has gained academic acceptance and popularity for decades. (Bernsen et al., 1994);(Robertson et al., 1997);(Fiedler, A., Gabsdil, M., & Horacek, H. (August 2004). Tools to aid in the incremental refinement of oz experiments in natural language. International Conference on Intelligent Tutoring Systems (p. 325-335));(Maulsby, D., Greenberg, S., & Mander, R. (May 1993). Prototyping intelligent agents via the Wizard of Oz. Proceedings of the INTERACT'93 and CHI'93 Conferences on Human Factors in Computing Systems (pp. 277-284)). These aids can reduce distraction / anxiety while increasing attention by increasing the cognitive load associated with unexpected stimuli. WoZ proposes an in-subject two-state SART that uses multimodal sensors from which the user's errors of commission, errors of omission, reaction time, state anxiety, and fatigue level are calculated. (Burchi, E., & Hollander, E. (2019). Anxiety in Autism Spectrum Disorder); (Ruttenberg, D. (2020). The SensorAble project: Multisensory assistive technologies to filter distractions and improve focus in individuals diagnosed with an autism spectrum condition. MPhil / PhD upgrading report University College London).

[0329] Memory factors are minimized in the SART test because the visual tasks are unobtrusive and are tried by the participants prior to testing. Scoring of participants' performance includes interval measures, but there are other important metrics that can be derived from both qualitative and quantitative scoring. As stated in Robertson (1996), these do not create cognitive burdens of similar dynamics and characteristics, and therefore they do not constitute myopic or simplified sampling rates of processing measures. Furthermore, this study separates the visual sustaining task from the auditory distractors, which avoids cross-modality and interference concerns.

[0330] In this study, a t-test / correlation-point bimaterial model was utilized (Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G*Power3: A flexible statistical power analysis program for social, behavioral, and biomedical research. Behavioral Research Methods, 39, 175-191). Calculations were based on a given alpha, power, and effect size. In general, a T-test is a calculation that informs the significance of differences between groups. In this case, it answers whether the data between autistic scores and non-autistic scores (measured on the average) could have occurred by chance or not. Alpha (α) is the threshold used to determine whether the test statistic is statistically significant and was chosen by the inventors (typically 0.05). A statistically significant test result (p=probability and p≦0.05) indicates that the test hypothesis is incorrect or should be rejected, whereas a p-value above 0.05 means that no effect was observed. The statistical power of a significance test (t-test) depends on (i) the sample size (N), where as N increases, the power increases; (ii) the significance level (α), where as α increases, the power increases; (iii) the effect size, where as effect size increases, the power increases.

[0331] Half of the sample included neurotypical participants and half were identified as or had an ASC diagnosis. All participants were available before and after the WoZ manipulation. Baseline test and related scores were procedurally derived both before and after the subtests. Cognitive conflict was created during the subtests that could result in a hypothesized, measurable increase in both distractibility and anxiety. At the same time, this increase was likely pooled with a decline in concentration and cone-of-attention performance.

[0332] Finally, during the latter subtest, a "confederate" (a human wizard) presented a collection of hand-crafted warnings, filters, and guidance. These mimicked the behavior of the wearable intervention by offsetting and counterbalancing distracting auditory stimuli. To reduce fatigue effects, these interventions were present in counterbalanced, randomized, multiple sessions. Alternatively, combinations of warnings, filters, and guidance were prepared to shorten the total length of the experiment, as shown in Figure 3.

[0333] Delivery of isolated and displaced supports yielded extensive indices of test responses. Mixed methods (qualitative and quantitative assessments) are combined with participants' overt behavior captured by audio and video recordings providing coding and analysis with synchronization of sample accuracy to the system software. Depending on sample size and time constraints, the design allows for post-hoc video analysis (e.g., participant walk-throughs) in either structured or free-form. These examinations may help facilitate recall, accuracy, and provide further understanding of anxiety and other episodic test moments.

[0334] Reliability was tested by administering the procedure once to subgroups of autistic and non-autistic subjects over seven separate test sessions. The IH Robertson protocol was used because of its tradition and wide acceptance in the scientific community (Robertson et al., 1997).

[0335] The SART procedure consisted of visually presenting 100 single letters (e.g., A–Z) for up to 5 min. Each letter was presented for 250 ms, followed by a 900 ms mask. Subjects responded with a key press for each letter, except for 10 occasions when the letter "X" appeared, in which case subjects had to withhold (inhibit) their response. Subjects used their preferred hand. Target letters were distributed in a non-fixed randomized manner across the 100 trials. The period between successive letter onsets was 1150 ms. Subjects were asked to prioritize accuracy first, followed by speed, in performing the task.

[0336] The letters were presented equally in the same font (Arial) and size (36 point) corresponding to a height of 12.700008 mm. The mask following each digit consisted of a white square with no border or fill coloring. The total area of ​​the mask depended on the user's screen size (e.g., the entire screen was considered as the maskable area). As a comparison, a 10" diagonal screen would generate a 25cm diagonal mask for a laptop and a 40cm diagonal mask for a tablet. Similarly, a 15" and a 20" diagonal screen would generate a 38.10cm and a 50.80cm mask for both laptops and tablets, respectively.

[0337] Before each session, there was a practice period consisting of the presentation of 15 letters, two of which were targeted. In addition, participants self-rated the State-Trait Anxiety and State-Trait Fatigue Inventory (STAFI) before and after each of the seven SART / WoZ tests (Spielberger, C.D. (1972). Conceptual and methodological issues in research on anxiety. Anxiety: current trends in theory and research on anxiety). Trait and state anxiety and fatigue were measured by combining the fatigue portion of the more commonly used anxiety-only inventory. The STAFI has historically been used in clinical settings to diagnose anxiety disorders and differentiate them from depressive symptom groups. The STAFI is appropriate for those with at least a reading level of Grade 6 or higher (American Psychological Association. (2011). State-Trait Anxiety Inventory (STAI). American Psychological Association).

[0338] Participants chose from five state anxiety items containing illustrations and text describing their feelings at the moment of inquiry, including "1-extremely anxious," "2-slightly anxious," "3-neither anxious nor calm," "4-slightly calm," and "5-extremely calm;" "worried;" "calmed;" and "reassured." Lower scores indicated greater anxiety.

[0339] Similarly, the five fatigue items included illustrations and text to indicate feelings at the moment of inquiry, including "1-extremely tired," "2-slightly tired," "3-not tired when awake," "4-slightly awake," or "5-extremely awake." Lower scores indicated greater fatigue.

[0340] Performance of the SART clearly requires the ability to inhibit or withhold a response, which becomes more difficult when distractors are introduced into the testing paradigm. Specifically, handmade sonics of varying amplitude, frequency, time / length, distortion, localization, and phase were introduced to mimic sounds that may occur in office, work, educational, and school environments.

[0341] A total of 28 sounds were played over a five-minute period, including office industry, fire alarms, telephone ringing, busy signals and dial tones, classroom lectures, copier and fax machine operations, footsteps, sneezing, coughing, and pencil scribbles.

[0342] Prior to testing, this study accomplished similar testing with pre-programmed sensing and associated interventions. Scripted sonic stimuli were run along with created participant alerts, filters, and guidance to provide customized intervention-assisted sensations and responses. These smart system components are predefined, and sensor cause and effect are evaluative to stabilize system operation and facilitate autonomous testing and synchronous data recording. As in Forbes-Riley, K. and Litman, D. 2011, Designing and evaluating a wizarded uncertainty-adaptive spoken dialogue tutoring system, Computer Speech & Language 25, 105-126, this study leverages WoZ in place of multiple system components, the combination of which presents a fully intelligent and integrated system. The human wizard is primarily a conductor / evaluator whose function and oversight of the program material are not identified to the participant. The user makes selections via a "dumb" control panel to provision customized alerts, filters, and guidance. Importantly, the mechanisms increase autonomy by providing specific features for participant evaluation while ostensibly eliminating the influence of the evaluator. When selecting these components, the following questions are considered: What requirements should the evaluator meet before conducting the study? How will the evaluator adhere to the plan and which measurements will reflect the flow of tests and subtests? How should the control panel component be designed and how will this affect its operation? How will the evaluator's personal behavior affect system operation?

[0343] All studies were managed and hosted in the Gorilla Integrated Development Environment (IDE) and available via most common web browsers and devices (Anwyl-Irving et al., 2019). All audio, video, and associated ecological / interoceptive data were presented and collected in real-time via the IDE and raters.

[0344] Study variables:

[0345] The overarching study was divided into four components, including: (i) the PPI study and the previously mentioned PTBE; (ii) the assessor (including tasks, self-reports and controls); (iii) the system prototype (non-wearable subsystem); and (iv) the participant (recorded). Test variables are listed in Table 3A and shown in Figure 18:

[0346] [Table 3A]

[0347] Oz Protocol Design SART / Wizard:

[0348] Figure 6 is a flow chart showing the SART / WoZ protocol used in this study, containing four superordinate classes including research objectives, variables, assessments, and outcome measures. The research questions, independent and dependent variables, potential assessments / activities, and expected outcomes are also shown. Based on this SART / WoZ protocol design, the corresponding class descriptions are listed in Table 3B.

[0349] [Table 3B]

[0350] Test procedure:

[0351] Each participant participated in a single experimental session after initially completing consent and demographic forms. The session began with a short (1-2 min) tutorial to ensure participants were familiar with the proper operation of the testing software and to introduce participants to the importance of remaining within range of the webcam and pointing device to properly monitor their environment and their physiology. After the tutorial, participants were advised that an evaluator would be available throughout the session to help monitor the system and answer any questions between tests. Participants were not informed of the evaluator's contribution to the test (WoZ), that any warning, filtering or guidance programming was predefined prior to the experiment, or that the system's preference control was placebo in nature.

[0352] The WoZ test (baseline to multiple interventions, then return to baseline) included three phases. Phase I started with baseline I cognitive testing; i.e., there were no distracting cues or interventions. Phase II was introduced concomitantly with randomized sonic distraction and filters, warnings and guidance applied in concert with the test. Phase III reintroduced a return to baseline to ensure that participants' recovery and responses were not memorized and the randomization effect was properly sustained.

[0353] Warnings, filters and guidance structures:

[0354] The warnings and guidance in this test protocol utilize Amazon Polly™, a neural text-to-speech (TTS) cloud service designed to increase engagement and accessibility across multiple platforms (Neels, B. (2008). Retrieved December 17, 2020 from https: / / aws.amazon.com / polly / ). As listed in Table 4, Polly's output is cached within the test system to depict a safe, non-controversial news anchor personification that speaks in a style tailored to the specific use case.

[0355] [Table 4]

[0356] During this test, a single modality of various sonic distractors was scheduled for testing. For fidelity and deeper understanding, both sonic and visual cues can be easily programmed, but the experiment was conducted with audio cues only. The stimulation events and cues are listed in Table 5 along with their associated filter names and descriptions. According to an exemplary embodiment, the success and effectiveness of the prototype wearable device can be evaluated based on participant data (both quantitative and qualitative) collected during the test administration of these stimulation events and the participant's performance.

[0357] [Table 5]

[0358] Protocol Test Procedures:

[0359] Participants were instructed to remain near their computer's webcam and in direct contact with at least one of their pointing devices (e.g., mouse, trackpad, keyboard) at all times during the experiment. Participants were also informed that measures of engagement, concentration, comfort, productivity, and autonomy would be tested; that environmental and physiological / psychophysiological monitoring (e.g., ecology and endoreception) would occur during testing; and that participants' head bobbing, pupil response, GSR, environmental sounds, and vision would be collected.

[0360] intervention:

[0361] As the study progressed, participants received a combination of pre-distraction and / or pre-filtered audio cues with warning supports (e.g., muted distractors, spatially focused distractors, etc.). Optionally, participants also received post-stimulus guidance to help them return to the task / activity / test.

[0362] Data collection methods:

[0363] This study utilized three data capture methods: direct computer entry / scoring, video analysis, and self-report. The first data capture method was integrated into the Gorilla application, the second data capture method was intended to record and enable observation of subjects' system interaction, and the third data capture method could reflect participants' and evaluators' operating experience (Goldman, N., Lin, I.-F., Weinstein, M. and Lin, Y.-H. 2003. Evaluating the quality of self-reports of hypertension and diabetes. Journal of Clinical Epidemiology 56, 148-154).

[0364] Participants:

[0365] PPI studies of verbally impaired autistic (ASC) participants included: (i) a focus group exploring distraction / attention; (ii) the Lived Experience Attention Anxiety Sensory Survey (LEA), which provides a first-person perspective on sensory, attentional, and mental health indicators; 2 Se). LEA 2 Se was developed, customized, and further modified by fifteen (15) participants who voiced autism-related questions and problems related to sensory, attention, and anxiety. 2 The SE was presented to both autistic and non-autistic participants, who agreed to complete an online questionnaire consisting of 103 items for autistic participants and 48 items for non-autistic participants (Autistic: N = 187, Female = 75; Non-binary = 5; Non-autistic: N = 174, Female = 85; Non-binary = 3). 2Se was used as the basis for the WoZ Proof of Concept / Testing (POC / T, N=5, 2=ASC, 3=NT, 4 / 1=F / M) and final testing / experimentation. The POC / T verified proper system operation and translation of the user interface to data collection devices and downstream to analytical applications.

[0366] After the POC / T and before the SART / WoZ test, the PTBE was administered (N=131; 71=ASC, 60=NT; 59=M, 72=F). Each participant received four separate tests, including the Matrices Reasoning Item Bank (MaRs-IB): a novel open access abstract reasoning item for adolescents and adults; the Autism Spectrum Quotient (AQ): a 50-item self-report questionnaire to measure the degree to which adults with normal intelligence have characteristics related to the autism spectrum; and the Adult ADHD Self-Report Scale (ASRS A and ASRS B) Symptom Checklist: self-report questionnaires used to aid in the diagnosis of adult Attention Deficit Hyperactivity Disorder (ADHD), specifically everyday problems related to cognitive, academic, occupational, social and economic situations.

[0367] Based on the PTBE results, a well-matched cohort of SART / WoZ participants was selected for demographic and test battery outcomes, which provided a near 50 / 50 balance in neural differences between experimental and control groups (N=40; 19=ASC / 21=NT; 15=M / 24=F / 1=non-binary) and allowed for seven randomized controlled tests before and after the sensory manipulation.

[0368] Data analysis examined the use of variables derived from the aforementioned PPI study and PTBE to understand the lived experiences of autistic individuals regarding distraction, attention, and anxiety. These variables and supporting data were used to predict how participants of different ages and genders might perform tasks involving distracting visual, audio, and physiological / psychophysiological cues. These variables included: sensitivity affect; anxiety proneness; distractibility index; visual difficulties index; audio difficulties index; physiological / psychophysiological (interoceptive) difficulties index; and correlations. Of these, six variables, three of which are contextually related to different modalities, were tested in the present study. Descriptive statistics and correlations for these variables are listed in Table 6.

[0369] [Table 6]

[0370] Stepwise regression:

[0371] Dummy variables were created for both age and sex (i.e., the only demographic factors that were not correlated) and combined with the sensitivity, anxiety and distractibility variables (SI, AP and DQ) embedded within stepwise regression analyses to predict scores within speech, vision and physiological / psychophysiological / sensory receptors. The model(s) with the highest R2 / significance are reported in Table 7.

[0372] [Table 7]

[0373] Standard regression

[0374] One categorical variable was regressed onto the three modalities (e.g., sound, visual and physiological / psychophysiological) with a continuous variable (either Sensitivity Affect [SI], Distraction Quotient [DQ] or Anxiety Proneness [AP], with which the correlation was highest accordingly) (i.e., either age or gender, depending on which was significant in the previously performed ANOVA), all of which served as dependent variables in the present study. Standard regression values ​​are shown in Table 8:

[0375] [Table 8]

[0376] PTBE Results / Data Analysis:

[0377] For the initial run of SART / WoZ participants (N=37, mean age 25.70, SD=7.442), their mean PTBE scores were ranked as follows: MaRs-IB=62.20% and 18.64; AQ=24.51 and 12.66; ASRS-1=3.19 and 1.66; ASRS-2=5.51 and 3.30). Independent samples tests for all PTBE outcomes yielded MaRs-IB of (F = .166, t = -.295, df = 35, and Sig. 2-tailed = 0.769), AQ of (F = .046, t = 4.494, df = 35, and Sig. 2-tailed = 0.000), ASRS-1 of (F = .281, t = 2.757, df = 35, and Sig. 2-tailed = 0.009), and ASRS-2 of (F = .596, t = 2.749, df = 35, and Sig. 2-tailed = 0.009). Demographic independent samples tests were not significant across age, sex, handedness, education, employment, income, status, children, home, and location.

[0378] For PTBE participants (N = 131; autistic = 71, non-autistic = 60, and those tapped for SART / WoZ), an ANOVA comparing autistic and non-autistic participants using Levene's test showed that the variance was significant for all scores, such as MaRs-IB scores (F(1,129) = 4.143, p = .044), AQ scores (F(1,129) = 81.090, p < .001), ASRS-1 scores (F(1,129) = 4.832, p = .030), and ASRS-2 scores (F(1,129) = 8.075, p = .005).

[0379] Similarly, for the same samples, ANOVAs comparing the gender of participants using Levene's test showed that the variance was not significant for MaRs-IB scores (F(1,129) = .143, p = .705), AQ scores (F(1,129) = .008, p < .930), ASRS-1 scores (F(1,129) = .973, p = .326), and ASRS-2 scores (F(1,129) = .018, p = .893). The means of cohort and group scores are listed in Table 9 and shown in Figure 7:

[0380] [Table 9]

[0381] SART / WoZ Results / Data Analysis:

[0382] Errors of Commission (EOC) Performance

[0383] Performance of a cohort (N=40) consisting of autistic / ASC participants (N=19) and neurotypical / non-autistic / NT participants (N=21) across the SART study and from baseline to baseline retest showed improved performance. Namely, there was a mean reduction in Sustained Attention Response Task (SART) errors equal to 7.46% for all inhibition measures across the cohort (Figure 8A). The same cohort averaged an improvement of 14.50% (again, in error reduction) for the different intervals; i.e., from the onset of the distractor cue to the alert intervention. Finally, a similar improvement of 10.27% occurred from the distractor cue to the different interventions (e.g., alert, filter, and guidance; Figure 8B) for the combined aids. Improvements were significantly prevalent across the cohorts, regardless of the intervention. Notably, even after cessation of the intervention, the long-term improvement of a 17.52% reduction in errors persisted across the cohorts when the four technological aids were discontinued (Figure 8C). This resulted in a specific average improvement of 1.45 errors per participant, regardless of participant diagnosis (group membership). For each measure, improvement trend lines were well correlated (baseline to baseline, intervention only, and intervention removed).

[0384] Error Response Time of Commission (EOC-RT)

[0385] In general, response times increased across cohorts when participants experienced exposure to the intervention aid. Despite counterbalanced testing and their internal randomization, improvements in accuracy across cohorts occurred due to increased / slower RTs (e.g., 21.74% increase from baseline to the alert intervention). Note that the slowing of RTs is actually a desired effect of the intervention, as explained in detail below. It is worth mentioning that, unlike EOC, the net effect of RTs was small (10.69% faster responses were obtained once the intervention was discontinued).

[0386] In comparison, autistic response times were shorter (faster) than neurotypical controls. This may be due to a variety of factors that distinguish neurodiverse responsiveness, including, but not limited to, greater neural processing, differences in genetic makeup that affect sensory responsiveness, and superior activity in the visual cortex (Schallmo, M.-P., & Murray, S. (2016) People with Autism May Be Faster Movements. 19). For errors of commission, autistic participants experienced an RT increase of 19.39% (i.e., the desired slowing from the onset of the distraction to the guidance intervention), while their neurotypical counterparts experienced an undesired decrease (speeding up) in RT of nearly 1 percent (-0.74%) over the same period. The effect of response timing on accuracy resulted in an improvement of 8.67% of ASC participants and an increase of 1.27% of neurotypical (NT) participants. These results are shown in Figure 9A-C, which are graphical representations of EOC in relation to response times (RT) across the cohort of participants in the SART / WoZ study described herein. FIG. 9A shows EOC vs. RT from starting baseline to final baseline, FIG. 9B shows the EOC vs. RT intervention effect, and FIG. 9C shows the sustained effect of EOC vs. RT.

[0387] It is noted that slower reaction times tend to indicate greater mindfulness. Mindfulness can be defined as the participant's awareness of their internal feelings and their ability to maintain awareness without subsequent evaluation or judgment (e.g., defined as an outcome). Therapeutically speaking, the wearable device described herein fosters mindfulness through custom interventions (assistive technology). This helps to change and shape the participant's wandering mind and awareness. In essence, participants in this study become more aware, productive, and comfortable through alerts, filters, and guidance when exposed to sensory interruptions during a sustained attention to response task (SART). Over time, participants become more alert, less sensitive, less anxious, and less fatigued.

[0388] Recognizing that slowing RTs is a desirable outcome rather than an undesirable outcome, the data suggest that NT participants who previously experienced a decrease in RT (speeding up resulting in a smaller performance gain) may be further improved by utilizing warnings rather than guidance. This causes NTs to experience a desirable increase in RT (slowing down). Specifically, during the distraction onset period, both ASCs and NTs slow their RTs. As a result, EOC improved by 26.01% and 17.59% among both autistic and non-autistic participants, respectively. Over the same period, ASCs and NTs improved their RTs by 2.94% and 1.9%, respectively.

[0389] Although the neural normed increase in accuracy appears small (i.e., 1.27% to 1.9%), this represents an improvement of 50% (49.60%). Thus, the delayed response time resulting from the custom intervention aid produces better performance outcomes. Autistic performance also improved by 200% (e.g., 8.67% to 26.01% accuracy). These results are shown in Tables 10A and 10B:

[0390] [Table 10A]

[0391] [Table 10B]

[0392] The dissociation between speeded and slowed RTs (and its effect on experimental and control groups) is not accidental. Evidence of an inverse RT effect on accuracy; i.e., faster RTs lead to higher accuracy, supported after repeated intervention aids were removed and then measured. Long-lasting effects of fewer errors (e.g., 17.92% and 17.09% reductions for ASC and NT, respectively) occurred even when response times were shortened (e.g., 21.48% and 2.688% faster RTs for ASC and NT, respectively). These were small but significant reductions, amounting to an average increase of 19.095 ms for autistic participants and 2.913 ms for non-autistic participants. Nevertheless, continued enhancement of performance occurred despite reduced response times.

[0393] The trend or tendency of response times provides interesting considerations. Specifically, across the seven trials, autistic and non-autistic RTs diverge. Autistic RTs increased from 88.89 ms to 69.80 ms from baseline to retest to baseline (a speedup of 29.78 ms over the period). In comparison, neurotypical RTs decreased from 82.59 ms to 105.46 ms, or a slowing of 22.86 ms. This results in a 52.64 ms gap between the experimental and control groups that is adjusted once the intervention is applied.

[0394] Explicitly, RTs increase (slow) by 17.72 ms for both ASC (i.e., from onset of distraction to guidance intervention) and NT (i.e., 19.06 ms for onset of distraction to warn intervention). These represent the maximum increase in RT for both groups and are not contrasted (i.e., again, both slow down). Equally important is the RT sustainment effect; i.e., once the intervention is removed, neither autistic nor non-autistic participants benefit from a slowing of RT. Both the ASC and NT groups accelerate their responses by 19.10 ms and 2.91 ms, respectively (even though there is a positive sustained performance due to fewer errors). These results are shown in Figure 10A-C.

[0395] Furthermore, as shown in Figures 11A-C, autistic reaction times are typically faster than neurotypical participants for the same task and intervention. Similarly, while error reduction (improved performance) occurs in both groups, autistic participants have greater variability in their improvement, while neurotypical participants make fewer errors overall. The only exception we see is with the combined intervention (e.g., alerts, filters, and guidance) where both NT and ASC are equivalent in their reduction to 7.4 errors each.

[0396] In summary, as a cohort and within subjects / groups, improved performance (e.g., fewer errors) resulted from the interventional assistance applied and measured with the assistive technology from the onset of sensory distraction. A 14.5% improvement across the cohort resulted with the alert intervention. Adjusting the intervention (i.e., applying filters and guidance to the alert) similarly resulted in variable improvements. The cohort improved in performance by 10.27% from the combination of these interventions. Autistic participants revealed higher performance (26.01% fewer errors) with the alert intervention, whereas non-autistic participants improved by 5.7% with the filter intervention. The sustained effect on performance improvement across the cohort (17.52%) and among individual groups (ASC 17.92% and NT 17.09%) continued well after discontinuing the intervention.

[0397] Reaction times increase when participants receive the assistive technology. By slowing down, participants allow and experience greater mindfulness that improves performance. From baseline to the warning intervention, the cohort showed an average slowing in RT of 21.74%, and from the onset of the distraction to the warning, the slowing was 11.35%. When any type of intervention was removed from baseline to retest to baseline, RT accelerated (decreased) by 10.69%. Comparison of the experimental and control groups shows that unlike autistic and non-autistic participants, RT decreases in ASC participants and increases in NT subjects. Nevertheless, both groups benefit under the intervention measures with improved performance, but neither group benefits from a sustained effect on RT once the assistive technology is removed.

[0398] Errors of Omission (EOO) Performance:

[0399] For the entire SART study, as well as for the study of cohort performance (N=40, ASC=19, NT=21) on errors of commission (e.g., not inhibiting a response when instructed to do so), we also analyzed errors of omission. EOO refers to not responding appropriately to any stimulus when inhibition is neither justified nor instructed. Over the same testing period and from baseline to retest, EOO increased by 49.16%. This means that there was a mean increase in Sustained Attention to Response Task (SART), measuring an average of 2.2 errors per participant (Figure 12A).

[0400] Although not a desirable outcome, when intervention was present, the cohort averaged improvement. Specifically, there was a 47.60% reduction in EOO from distractor cue onset to alert intervention. As can be seen in Figure 12B, a similar improvement occurred from distractor cue to combined intervention (albeit this time with a smaller improvement of 23.12%).

[0401] Notably, when the four technical aids were suspended, the improvement persisted over time with a 10.10% reduction in EOO across the cohort (Figure 11C). This was calculated by measuring the percentage error increase from baseline to baseline retest and then subtracting the error improvement measured from distraction by baseline retest. For each participant, this amounted to an average of 1.86 fewer errors for each participant, regardless of diagnosis / group membership. For each measure, the improvement trend lines correlated well (baseline vs baseline, intervention only, and sustained effect).

[0402] Error Response Time to Omission (EOO-RT)

[0403] In general, response times across the cohort increased when participants experienced exposure to the intervention aid, but did not increase from baseline to retest to baseline (remaining relatively flat at -0.20%; Figures 13A-C). Despite counterbalanced testing and their internal randomization, the initial EOO accuracy declines and final EOO accuracy improvements in the cohort occurred due to increased / slowered RTs (e.g., 5.16% increase from baseline to filter intervention). Again, as previously explained in the EOC section, this slowing of RTs is a desired effect of the intervention. It is worth mentioning that, unlike errors of commission, the sustained effect of RTs was small (3.48% faster responses were obtained once the intervention was discontinued).

[0404] The EOO response times of the autistic participants were similar to the EOC; in that both were faster than neuro-normal controls, in part due to the aforementioned neuronal processing and responsiveness. Thus, the autistic participants experienced an RT increase of 9.28% (i.e., the desired slowing from distraction onset to the filter intervention), whereas their neuro-normal counterparts produced an undesired decrease (speeding up) in RT of nearly 1 percent (4.44%) to the same intervention.

[0405] Compared to EOCRT, neurally normalized results were slightly faster (inferior), e.g., EOC vs. EOO yielded 126.94 ms to 122.05 ms. ASC participants performed significantly better (e.g., EOC vs. EOO yielded 91.02 ms to 114.19 ms). In contrast, the impact of RT on sustained performance was observed as a slight degradation (1.2%) in NT participants, but a slower (faster) performance in ASC (7.25%) and a relative flattening. These results are shown in Figure 14A-C.

[0406] RT also results in errors of omission when comparing autistic and non-autistic groups. There is a decrease in EOO among neurodiverse participants (but these still result in inaccuracies) (-15.12%). Similarly, an increase in accuracy (decreased EOO) is shown among neurotypical participants. Not surprisingly, faster RTs (4.44% due to distraction to filter for NT participants) actually resulted in more errors (15.09%). Also as expected, slower RTs (9.28%) among autistic participants resulted in less EOO (15.12%). Interestingly, both groups responded with equal and opposite magnitudes in accuracy with the opposite to similar interventions (by RT) and with NT (but not ASC participants) experiencing larger errors.

[0407] This seems unlikely. However, for the entire study participants' data (i.e., baseline to retest to baseline), the RT and EOO curves are indeed inversely proportional. Longer RTs result in less EOO, as expected. However, neurotypical RTs are less correlated. Higher (desirable) NT RTs produce fewer errors under filter intervention (desirable). However, larger RTs with guidance generate more EOO (undesirable). Thus, depending on the group, longer RTs result in reduced accuracy returns. If errors of commission correlate better with response time variance, errors of omission correlate poorly with RT.

[0408] Even though there was improvement among autistic participants with higher accuracy, this occurs due to an abnormal decrease in EOO response time. Higher accuracy (29.07%) and improvement from distraction onset to guidance intervention occurs with less RT slowing. Further slowing from distraction to filtering (9.28%) results in more imprecision (15.12%). Accuracy for non-autistic participants behaves as expected; i.e., an increase (e.g., slower RT) from -4.4% to -1.39% results in an 8.5% increase in accuracy (from 15.09% to 23.59%). These results are shown in Tables 11A and 11B:

[0409] [Table 11A]

[0410] [Table 11B]

[0411] As presented, the correlation between faster and slower RTs (and its effect on accuracy in experimental and control groups) is not coincidental for EOC. In contrast, there is a difference in EOO scores. Evidence of an inverse RT effect on accuracy emerges; i.e., faster RTs do not always result in greater inaccuracies.

[0412] In the previous table, an autistic reaction time of 9.28% increased resulted in a negative increase in accuracy (e.g., inaccuracy). However, speeding up (or reducing response time to 1.24%) resulted in higher accuracy (29.07%). This unexpected autistic divergence is not exhibited in the neurostandard EOO. An increase in speed (-4.44%) resulted in a lower accuracy of 15.09% errors, while a slowing down to 1.39% (increased speed) results in the expected higher accuracy (23.59%). These results are shown in Figures 15A-C.

[0413] While there is no long-lasting effect of fewer errors (e.g., both ASC and NT show EOC increases of 51.16% and 29.25%, respectively), RT accelerated by 7.58 ms for autistic participants and 1.53 ms for non-autistic participants. With faster RT, one would expect an increase in errors, but this is the opposite of the EOC-lasting effect. Simply put, the lasting effect shown in EOC performance / accuracy does not hold for EOO, just as RT errors for omissions correlate well with EOC.

[0414] Similar to EOC RTs, EOO responses in autistic participants remain narrowly and consistently faster than their more variable neurotypical counterparts. Also, while a reduction in omission errors (improved performance) occurs across autistic participants, there is less variability in this improvement, whereas neurotypical participants do not necessarily make fewer errors. This is in stark contrast to the EOC data. These results are shown in Figure 15A-C.

[0415] In summary, unlike identical cohorts, within subjects / groups who experienced improved performance (e.g., fewer errors of commission), errors of omission did not decrease equally from sensory distraction onset to assistive technology: a 15.16% decrease in EOO occurred in autistic participants (from distraction onset to filter intervention), whereas no decrease occurred with intervention application in neurotypical participants.

[0416] The combined effect on the entire cohort also proved to be unremarkable in terms of EOO improvement. Here again, only the autistic (experimental) group experienced benefits. It is worth mentioning that adjusting the intervention to other forms (e.g., modification, guidance, and combination) did not result in any discernible improvement for the neurodiverse participants. Only the filter intervention proved to be supportive. A sustained effect of improved performance was avoided for the entire cohort, as there were more than 10.10% errors of omission. The same remained consistent in the experimental and control groups when the intervention was discontinued (e.g., 51.16% and 29.25% increase in EOO for ASC and NT).

[0417] When assistive technology was invoked, reaction times increased across the cohort. By slowing down, participants allowed and experienced greater mindfulness, which improved performance. From baseline to filter intervention, ASC participants showed an average slowing of 9.28% in RT, and from distraction onset to warning, the slowing was 1.535%. Neurotypical participants undesirably accelerated 4.44% under the same filter intervention, but slowed down for both the guidance intervention (2.61%) and the combined intervention (4.58%).

[0418] When the entire cohort intervention of any kind was removed from baseline to baseline retest, RT accelerated (decreased) by 3.48%. Thus, there was no significant sustained effect on EOO RT.

[0419] Similarly, for both experimental and control groups, autistic and non-autistic participants experienced a decrease in RT, with no significant sustained effect on EOO. ASC RTs increased by 7.25%, while NT RTs increased by 1.20%.

[0420] As alluded to above, the PPI study investigated issues and associations between three variables: sensory (sensitivity), mental health (anxiety and fatigue), and distraction (attention). Using the PPI study, we developed the sensitivity mental health distraction model shown in FIG. 26, which shows how anxiety and fatigue can mediate sensory sensitivity and distraction in both autistic and non-autistic diagnostic groups. The model in FIG. 26 links sensory cues (e.g., labeled #1, including the individual's hyper, hypo, and sensor-seeking characteristics) to mental health mediators (e.g., labeled #2, describing the individual's anxiety and / or fatigue) and distraction (e.g., labeled #3, describing the individual's ability to focus / maintain attention). In terms of ordering, the model extends sensory cues through mental health characteristics that can further modulate the individual's attentional reactivity, versus a straight line going from cues to distraction alone. While sensitivity was previously hypothesized to confound top-down and bottom-up attention, the model in Figure 26 embodies new lateral relationships that more fully illustrate sensitivity and attentional processing in autistic individuals.

[0421] Figure 27 is a flow chart showing the design / methodology of the aforementioned PPI study. The PPI study was conducted in two parts. The first phase consisted of five autistic-only (14 adults, ages 18-54) online focus groups to better understand everyday experiences regarding sensory sensitivity (Figure 27, item #2). The focus groups explored how sensory sensitivity affected both attention and mental health across three themes, namely: (i) experiences in adult contexts of higher education, employment, and social places; (ii) technological tolerance, and digital mediations that can help autistic individuals in adverse sensory environments; and (iii) a questionnaire for phase 2 with a larger group of participants in easy-to-understand, autistic-friendly language.

[0422] As shown in items #6 and #7 of Figure 27, the second phase included an online questionnaire for both autistic (N=187) and non-autistic participants (N=174). Both groups answered questions about their sensory sensitivity (visual, auditory, and physiological), concentration (distraction), and mental health (anxiety and fatigue), and in the case of the autism-only group, questions about interest, personal desire, and tolerance for assistive technology. The questionnaire design provided diversity of life experiences by allowing for unlimited response opportunities in addition to multiple choice questions.

[0423] Table 12 shows the demographics of participants in the first wave of the PPI study.

[0424] [Table 12]

[0425] Participants were recruited through opportunity sampling using university databases in both the UK and the US combined with social media across the globe. All individuals identified themselves as autistic and / or indicated they had a form diagnosis. Participants (N=14, female=7; non-binary=1) aged 17-38 years attended one of five scheduled focus groups with a minimum of two and a maximum of four people each. Each hour-long online focus group utilized a screen presentation to guide the discussion, helping to delve into participants' everyday experiences with sensory sensitivity, technology, attention, and mental health. A collection of over 100 questions was summarized and combined into 17 items and arranged across six overarching topics (Figure 27, item #1) shown in Table 13.

[0426] [Table 13]

[0427] These questions corresponded to the screen presentation and were read aloud by the researcher to prompt the participant's response. As the participant responded, they were prompted to express their thoughts and feelings about how technology could be used to accommodate and assist in their daily activities (Figure 27, item #3). For example, Q.6.: How do you "deal with interruptions, staying focused, or working in a cluttered environment?" could allow the researcher to probe the participant's response in more depth by asking, "What type of assistive device could help you resist or combat the interruptions just described in your work; or how does this technology succeed?"

[0428] Participants were encouraged to state any concerns they had regarding the wording of the questions (e.g., ambiguity, terms that were difficult to understand, and relevance of the topic under consideration) and to freely provide alternative wordings using autism-friendly language (Figure 27, item #4). For example, participants identified Q.3 as unclear because it was too broad. They suggested rephrasing it as a yes / no question: "Soft sounds make me tense and tired." Upon completing the final meeting, audio / video recordings of all focus groups were transcribed and imported into Qualitative Data Analysis (QDA) computer software for further analysis. These transcripts were coded first by placing the responses into the six previous topics, then by the participants' valence, and finally by any suggested alternative wordings. The QDA software was programmed to output a list of response words that were mentioned more than 10 times, as shown by Table 14 below.

[0429] [Table 14]

[0430] The 22 response words, together with their counts, were arranged into the six original topics allowing for the examination of presence, salience, and the relationship of the words to sensory, attentional, skill, and mental health issues.

[0431] As shown in Table 14, salient words were divided into one of four mean valence categories, which indicates how participants expressed concerns about these issues within the context of the subject matter and questions they were asked. For example, participants may have expressed a "positive" interest in using technology to help overcome anxiety, or expressed neither a "positive" nor a "negative" interest in light sensitivity (i.e., "neutral"). Higher valences (a total of 10 or more) were judged to be more important than lower numbers. These counts were used to create the Phase 2 PPI questions, discussed further below.

[0432] Suggested modifications to the question wording were studied and QDA computer software was used to output a list of alternative wordings, shown in Table 15.

[0433] [Table 15]

[0434] The 46 items ranked by word count and weighting (i.e., individual word count divided by the total number of all options) were sorted. Table 15 was used to rephrase 79 of the original 103 questions by replacing troublesome words with alternative words suggested by participants. The modification of the questions was also based on the visualization report (Figure 27, item #5; and Figure 28).

[0435] At the end of the final focus group, a word group representing autistic voiced color expressions was created using the data from Table 15. The word cloud is shown in Figure 28. Participants submitted the word cloud to rate their satisfaction with the alternatives and received unanimous approval (both in written email responses and / or in subsequent internet chats). This agreement was used to help select which questions would be re-voiced to reflect the autistic alternative expressions. For example, participants stated: o The centrally located, larger, and darker word cloud terms are appropriate for structuring Phase 2 questions using alternatives such as "things," "people," and "distractions." o Items just outside the midpoint of the cloud are appropriate when questioning sensory cues and reactivity, including "needing help," "feeling anxious," and "worried." o Words near the edge of the cloud were suitable for framing questions about technology, including “wearable,” “coaching,” and “assistance.” These were supportive in asking outcomes about “comfort,” “participation,” and “performance.”

[0436] Table 16 shows the demographics, including gender, age, and education level, of participants in the second phase of the PPI study.

[0437] [Table 16]

[0438] Participants were recruited similarly to individuals in the focus group using university websites and social media. Non-autistic individuals were recruited and well-matched to ensure compatibility with their neurodiverse counterparts. Eligibility criteria for both groups included age 18-49 years and English proficiency. Non-autistic participants were specifically requested not to self-identify or possess an autism diagnosis. After removals, a total sample of N=361 individuals was analyzed (i.e., 187 autistic participants and 174 non-autistic participants).

[0439] The online survey was designed to expand the questions of the original focus group beyond word use and their changes to collect data describing sensory patterns, technology experiences, and accommodation requests, all described using authentic, clear, and relatable language by a larger sample of respondents who are likely to share the experiences identified by the smaller group of original participants.

[0440] The majority of the questions in the second phase were drawn from a variety of sources, including those from the UCL Student Mental Health Survey (SENSE; McCloud et al., 2019), the Stanislaus State Concentration Questionnaire (CQ, 2019), the CogniFit Online Cognitive Assessment Battery for Concentration (CAB-AT, 2018), the Cognifit Cognitive Assessment for ADHD Research (CAB-ADHD, 2018), and the semi-structured interviews developed by Ashburner and colleagues (2013). These questions were supplemented and / or refined by the researchers.

[0441] Using alternative wording, 79 of the original 103 questions deployed online for autistic-only participants (Figure 27, item #6) were modified. A second questionnaire was deployed using the same wording but for non-autistic participants (Figure 27, item #7). 55 of the original 103 questions that focused on technology use and tolerance were removed from the non-autistic version, including, for example: "How often do you use a smartwatch?"; "I would be looking forward to owning a wearable device that knows my preferences regarding distractions"; "I would be interested in knowing how a wearable device could help me in environmentally noisy situations" because such questions were considered less useful when comparing sensitivity, attention, anxiety, and fatigue characteristics between diagnostic groups. Their removal resulted in a 48-item non-autistic questionnaire.

[0442] Online demographic responses were imported using QDA software and examined as follows: Demographic questions. Education was used as a proxy to measure intelligence level, to match groups on abilities and to avoid collecting Intelligence Quotient (IQ) data, respecting autistic heterogeneity (Motron, 2004). Sensory sensitivity variables. A wide range of modalities were queried using three input variables: visual, auditory, and physiological. A 5-point Likert scale was used with responses ranging from strongly disagree (1 point) to strongly agree (5 points) unless otherwise stated. A mean score was calculated and ranged from 1 to 5, with higher scores indicating greater sensitivity. Sample queries included, "Have you ever felt sensitive, distracted, or anxious at work or school for visual reasons (e.g., lighting, movement, colors, etc.)?" and "Sensitive to vibrating wearables (e.g., cell phone, FitBit, iWatch, or another notification device)?" Visual Sensitivity Variable (VSV). Nine questions taken from the Stanislaus State Concentration Questionnaire (CQ, 2019) and the Cognitive Assessment Battery for Concentration (CAB; Cognifit, 2018) defined movement, light, and other cues that negatively affect activity. Sample questions included, “Easily distracted or sensitive to certain environmental sights / views” and “Avoid visually stimulating environments.” Auditory Sensitivity Variable (ASV) Seven questions taken from the Stanislaus State Concentration Questionnaire (CQ, 2019), the Cognitive Assessment Battery for ADHD Research (CAB-ADHD; Cognifit, 2018), and semi-structured interviews (Ashburner et al., 2013) described loud, startling, and other cues that have an undesirable effect on the individual. Example questions included, “I am easily distracted or sensitive to certain environmental sounds (e.g., noises, volume, pitch, conversations)” and “I describe the humming of lights or refrigerators, fans, heaters, or the ticking of clocks as distracting.” Four questions were reverse scored, reflecting higher scores indicating greater sensitivity. Physiological Sensitivity Variable (PSV). Six questions derived from the Stanislaus State Concentration Questionnaire (CQ, 2019) and the Cognitive Assessment Battery for ADHD Research (CAB-ADHD; Cognifit, 2018) related to egocentric effects of attention and anxiety and self-centered sensations. Sample questions included, “I do not like to be touched” and “I am easily distracted or sensitive to certain physiological thought-feelings (e.g., anxiety, racing thoughts, tinnitus).” Anxiety variable (AV). Sixteen questions derived from the UCL Student Mental Health Survey (Sense; McCloud et al., 2019), interviews (Ashburner et al., 2013), and the Cognitive Assessment Battery for Concentration (CAB; Cognifit 2018) described how sounds, sights, and sensations affect anxiety. Questions included: “Certain sounds, sights, or stimuli make me tense, anxious, or on edge” and “Nothing makes me feel dazed or anxious.” Three questions were reverse scored, reflecting higher scores indicating greater anxiety. Distraction variable (DV). Ten questions taken from the UCL Student Mental Health Survey (Sense; McCloud et al., 2019), the Stanislaus State Concentration Questionnaire (CQ, 2019), the Cognitive Assessment Battery for ADHD Research (CAB-ADHD; Cognifit, 2018), and interviews (Ashburner et al., 2013) described participants' sensitivity to distractions in daily life. Inquiries included "I often start new tasks and leave them unfinished" and "If interrupted, I can return to what I was doing very quickly." Two questions were reverse scored to reflect higher scores indicating sensitivity to distractions. · Autism Spectrum Quotient (AQ-10). The 10 questions were derived from the Autism Spectrum Quotient (Allison et al., 2012) to test only non-autistic individuals who may be undiagnosed and / or have milder levels of autistic symptomatology (Baron-Cohen et al., 2001). These measures were used to ensure within-group participant concordance and to determine whether subclinical traits are present in undiagnosed individuals. The AQ-10 is a subset of a larger 50-item survey and uses a 4-point Likert scale ranging from definitely agree to definitely disagree across the 10 questions. A single point maximum is scored for each response, with a total score of 6 or greater indicating a concentration of autistic traits.

[0443] Demographics were examined by descriptive statistics, chi-square, and Mann-Whitney U tests. Significant differences between matched diagnostic groups and individuals were obtained by performing independent sample t-tests. Sequential one-way ANOVA tests revealed significance in differences in anxiety and distractibility as related to demographic characteristics. Tukey post-hoc tests confirmed significance between diagnostic groups. Pearson correlations established robustness between variables, and linear regressions identified predictive anxiety and distractibility traits.

[0444] Diagnostic mediation analyses tested the hypothesis linking anxiety as a mediator between sensation and distraction (see Figure 26). Analyses were performed to assess measurement error while adjusting for bootstrapping techniques (e.g., random sampling with permutation bias correction and accelerated 95% confidence intervals; 1000 resamples) using a statistical package.

[0445] A chi-square test of independence was conducted to examine the relationship of gender between diagnostic groups. The relationship between these variables was not significant at p<.05 (x2=2.946, p=.229). A Mann-Whitney U test was conducted to determine whether there were differences in age (U=16213, p=.948) and education (U=15461.5, p=.356) between diagnostic groups. Neither result was significant at p<.05.

[0446] The mean distribution of anxiety and distractibility scores for diagnostic groups across demographic variables is shown in Figure 29, which shows graphs of mean anxiety and distractibility / attention scores. A series of one-way between-subject ANOVA tests were conducted to compare differences in anxiety and distractibility scores across the various demographic characteristics. The autism group revealed a significant effect of age on both anxiety, F(3,183)=6.733, and distractibility scores, F(3,183)=10.856, both at the p<.001* level.

[0447] Post hoc comparisons using Tukey HSD tests indicated that on anxiety, 30-39 year olds scored significantly higher than 18-21 year olds (mean difference = .377, p = .008), and 40-49 year olds scored significantly higher than 18-21 year olds (mean difference = .830, p < .001), 22-29 year olds (mean difference = .525, p = .022), and 30-39 year olds (mean difference = .453, p = .042). For distraction, post hoc comparisons using Tukey HSD tests indicated that 30-39 year olds scored significantly higher than both 18-21 year olds (mean difference = .468, p < .001) and 22-29 year olds (mean difference = .261, p = .005), and 40-49 year olds scored significantly higher than both 18-21 year olds (mean difference = .614, p = .001) and 22-29 year olds (mean difference = .407, p = .040).

[0448] There were no significant differences in non-autistic participants across any age group, including anxiety scores, F(3,170) = 1.462, p = .227, or distractibility scores, F(3,170) = 2.430, p = .067. However, there were significant differences in anxiety scores and gender for non-autistic participants, F(2,171) = 4.289, p = .015. Post-hoc comparisons using Tukey HSD tests showed that females scored significantly higher than males (mean difference = 2.911, p = .024). However, there were no significant differences in gender for non-autistic participants in distractibility scores, F(2,171) = .514, p = .599.

[0449] Tukey post-hoc tests revealed that there was a significant difference in anxiety scores between genders for autistic participants, F(2,184) = 16.944, p < .001, with non-binary participants scoring significantly higher than male participants (mean difference = 1.301, p < .001) and female participants (mean difference = 1.156, p < .001). For autistic individuals, there was a significant difference in distractibility between genders, F(2,184) = 6.551, p = .002. Post-hoc distractibility comparisons using Tukey HSD tests showed that non-binary autistic participants scored significantly higher than male autistic participants (mean difference = .702, p = .002) and female autistic participants (mean difference = .588, p = .015).

[0450] There was a significant effect of education for autistic individuals on anxiety scores, F(2,184) = 3.915, p = .022; however, it was not significant for non-autistic participants, F(2,171) = .148, p = .863. Post hoc comparisons using Tukey HSD tests indicated that autistic participants who had attended graduate school scored significantly higher on anxiety than individuals with a bachelor's degree (mean difference = 3.418, p = .016). There was no significant effect of education on distractibility for autistic participants, F(2,184) = 2.108, p = .124; however, there was a significant effect from education on distractibility for non-autistic participants, F(2,171) = 4.360, p = .014. Post hoc comparisons using Tukey HSD tests revealed that non-autistic individuals with no college degree scored significantly higher on distractibility than those with some (mean difference = .364, p = .016).

[0451] Table 17 shows the sensory sensitivity variables and mental health scores and statistics for the diagnostic groups, including mean scores and differences between groups.

[0452] [Table 17]

[0453] The 187 autistic participants showed significantly higher levels of visual, t(359) = -8.090, p < .001, and physiological sensitivity, t(359) = -6.681, p < .001, compared to the 174 non-autistic individuals. Autistic members showed significantly higher levels of anxiety, t(359) = -1.993, p = .047; however, the study did not show any significant effects for either auditory sensitivity, t(359) = .585, p = .559, or distractibility scores, t(359) = 1.443, p = .157.

[0454] Tables 18 and 19 show the correlations between sensitivity and outcome variables for non-autistic and autistic participants.

[0455] [Table 18]

[0456] [Table 19]

[0457] Pearson correlation coefficients (Bonferroni adjusted) were calculated to assess linear relationships between non-autistic and autistic adults, as shown in Tables 18-19. There were positive correlations for non-autistic participants, showing significance between each of the sensitivity and outcome variables (mental health) and all sensory variables. There were similar significant positive correlations between each sensitivity variable and outcome variable (mental health) for autistic participants; however, the relationship between visual and hearing scores did not reach significance (r=.172, p=.180), while the association between hearing scores and physiology was nominally correlated (r=.255, p<.01). Both diagnostic groups showed correlations with outcome variables (i.e., anxiety and fatigue).

[0458] Unique predictors were modeled with anxiety and attention scores (Tables 20-21). Both autistic and non-autistic visual, auditory, and physiological variables were examined as predictors, and demographic variables were added as covariates. For both diagnoses, education and age were not significant predictors and were removed from further analysis. Non-binary individuals were excluded due to small sample sizes. Gender was dummy coded (i.e., 0-male; 1-female), and positive beta scores indicated significance between non-autistic females and outcome variables. All sensitivity variables and gender uniquely predicted non-autistic anxiety; the model accounted for 64.7% of the variance in anxiety scores (F=81.46, p<.001). Similarly, for autistic individuals, all sensory variables uniquely predicted anxiety, with the model accounting for 76.4% of the variance in anxiety scores (F=112.54, p<.001). Note that gender did not contribute to the anxiety scores of the autism group.

[0459] [Table 20]

[0460] [Table 21]

[0461] All input variables were entered as predictors of distractibility for both diagnostic groups, in addition to the demographic variables. For non-autistic individuals, age and gender did not significantly predict distractibility and were removed from the assessment. Visual cues, physiological cues, and education significantly predicted attention. Interestingly, sound sensitivity did not (p=.239) (see Table 22). The model explained 40.2% of the non-autistic variance in distractibility scores (F=28.407, p<.001); by extending the model with anxiety scores, this amount increased by 2%, and all variables (except sound) remained significant predictors (see Table 23).

[0462] [Table 22]

[0463] [Table 23]

[0464] Visual, sound, and physiological sensitivity were significant predictors of distractibility for an autism diagnosis; however, demographic variables were not significant (see Table 24). The non-autistic model explained 59.2% of the variance in distractibility scores (F=88.52, p<.001), which increased by 9% when anxiety scores were included (F=98.645, p<.001). Note, however, that once anxiety was included, visual and auditory scores were not significant distractibility (see Table 25).

[0465] [Table 24]

[0466] [Table 25]

[0467] To test the hypothesis that there is an indirect effect of anxiety mediating the relationship between sensory sensitivity (cues) and attention (distractions), hierarchical multiple regression analyses were performed (see Figure 30 for the non-autistic mediation model and Figure 31 for the autistic mediation model). There were bias and acceleration corrections with bootstrap confidence intervals (1000 resamples) that did not overlap with 0, and these all showed a significant relationship.

[0468] For non-autistic participants, the relationship between visual sensitivity and distractibility was significantly directly mediated by anxiety scale scores (b = .247, 95% BCaCI (.136, .352)). We also confirmed that anxiety was a significant and direct mediator between physiological sensitivity and distractibility (b = .2361, 95% BCaCI (.1261, .3680)). Education was added as a covariate in both cases. Auditory sensitivity was not tested in the mediation model because acoustic sensitivity was not predictive in the regression model.

[0469] For autistic individuals, there was a significant indirect effect of anxiety on vision (b=.3121, 95% BCaCI(.2242, .4423)) and hearing (b=.2689, 95% BCaCI(.1869, .3503)); however, the direct effects for neither were significant; thus, anxiety does not fully mediate the relationship between both visual and auditory sensitivity and distractibility in autistic individuals. However, anxiety significantly mediated the indirect relationship between physiological sensitivity and distractibility (b=.2835, 95% BCaCI(.1990, .3691)), and the direct effect was significant, indicating partial mediation.

[0470] Relationships between AQ scores and study variables for non-autistic individuals in the PPI study were examined to ensure matching of both in-group participants and to exclude individuals exhibiting subclinical traits that would affect out-group comparisons (see Table 16).

[0471] [Table 26]

[0472] The PPI study found significant positive correlations between non-autistic AQ scores and all study variables except sound, suggesting that individuals scoring higher on the autism spectrum (individuals with fewer non-autistic traits) were more likely to experience visual and physiological sensitivity accompanied by higher levels of anxiety and distractibility, consistent with the between-group results.

[0473] To examine the predictive outcomes of anxiety and inattention, the regression models were supplemented with the non-autistic AQ, but there was no significant effect (p=.198). However, the amount of model variance increased by 8% with the addition of the AQ score as a predictor of inattention, and it was a significant predictor of inattention scores when included in the regression model (p<.001) (see Table 27).

[0474] [Table 27]

[0475] The means and standard deviations of participants' self-rated anxiety and fatigue levels following the distraction SART and intervention SART, as obtained from the state anxiety and state fatigue surveys, are shown in Table 28. Given that the data were not normally distributed, Wilcoxon signed-rank tests were conducted to investigate whether the digital mediator was tolerated by participants based on their self-reported anxiety and fatigue.

[0476] [Table 28]

[0477] In the non-autistic group, self-rated anxiety was significantly lower after the filter-intervention SART (Mdn=3.0) compared to the distraction SART (Mdn=4.0; T=10.0, Z=-2.36, p=0.02). Similarly, self-reported anxiety and fatigue were significantly lower after the combined-intervention SART (anxiety: Mdn=3.0; fatigue: 2.0) compared to the distraction SART (anxiety: T=610.50, Z=-2.35, p=0.02; fatigue: Mdn=3.0; T=10.00, Z=-1.94, p=0.05). Lower anxiety scores are associated with participants feeling calmer, whereas lower fatigue scores are associated with participants feeling more alert.

[0478] In contrast, there were no significant differences in anxiety or fatigue between the distraction and each intervention SART for autistic participants. Given that autism is a very heterogeneous disorder and the aim of this intervention is to tailor the mediation to the various preferences of individuals, the analysis was subsequently performed on participants' scores from their best performing intervention (and the unaveraged mediation). Indeed, when using the best intervention, autistic participants' self-reported anxiety was lower in the best intervention SART (Mdn=2.0) than in the distraction SART (Mdn=3.50; T=5.00, Z=-2.24, p=0.02). Similarly, autistic participants' self-reported fatigue was lower in the best intervention SART (Mdn=2.00) compared to the distraction SART (Mdn=2.50; T=0.00, Z=-2.33, p=0.02).

[0479] Applying the same logic to the non-autistic group, where best-performing intervention scores were analyzed, similar results were found: Self-reported anxiety was lower in the best-intervention SART (Mdn=3.00) compared to the distraction SART (T=13.00, Z=-2.78, p=0.01). Self-reported fatigue was also lower in the best-intervention SART (Mdn=2.00) compared to the distraction SART (T=12.00, Z=-2.32, p=0.02).

[0480] The Mann-Whitney test showed that both diagnostic groups responded well to the Distraction SART and the Best Intervention SART (U 不安 = 174, p = 0.66; U 疲労 The results did not differ significantly in terms of the difference in self-reported anxiety or fatigue between the two groups (=182, p=0.83), indicating that they found the best intervention to be similarly useful.

[0481] To examine whether the intervention improved performance on the SART despite the presence of distractors, repeated measures ANOVAs were conducted (checked for independence, sphericity, and normal distribution) to compare performance between conditions (baseline, distractors, intervention) and diagnostic groups. Although variables were not normally distributed (with the exception of EoC), the sample size was large enough that the repeated measures ANOVAs were robust to violations of this assumption. Furthermore, as in previous analyses, scores from participants' best performing interventions were used. Means and standard deviations of variables are shown in Table 29.

[0482] [Table 29]

[0483] RTEoO was significantly different across the three conditions, F(1.37,50.81)=11.74, p<0.01. Post hoc pairwise comparisons using Bonferroni corrections indicated that RTEoO was slower in intervention (M = 142.50, SE = 12.28) compared to baseline (M = 111.79, SE = 8.34, p < 0.01) and distraction (M = 115.75, SE = 10.05, p < 0.01), but there was no main effect of group, F(1,37) = 1.40, p = 0.25, and no interaction between condition and diagnostic group, F(2,74) = 0.56, p = 0.51.

[0484] Similarly, RTEoC was significantly different across the three conditions, F(1.52,56.36)=16.92, p<0.01. Post hoc pairwise comparisons using Bonferroni corrections indicated that RTEoC increased significantly between baseline (M = 76.90, SE = 9.00) and intervention (M = 154.98, SE = 17.10), as well as between distraction (M = 85.57, SE = 13.33) and intervention (p < 0.01). However, there was no main effect of group, F(1,37) = 3.00, p = 0.09, or an interaction between condition and diagnostic group, F(2,74) = 1.12, p = 0.32.

[0485] Both assumptions of normality and sphericity were met, and a repeated measures ANOVA determined that EoC was significantly different across the three conditions, F(2,74)=12.5, p<0.01. Post hoc pairwise comparisons using Bonferroni corrections indicated fewer EoCs in intervention (M = 4.97, SE = 0.44) compared with both baseline (M = 6.27, SE = 0.51, p = 0.02) and distraction (M = 7.29, SE = 0.57, p < 0.01). However, there was no main effect of group, F(1,37) = 0.00, p = 0.95, or an interaction between condition and diagnostic group, F(2,74) = 0.39, p = 0.68.

[0486] Finally, surprisingly, EoO did not differ significantly across the three conditions, F(1.38, 50.95) = 2.54, p = 0.11, there was no main effect of group, F(1, 37) = 0.19, p = 0.67, nor was there an interaction between condition or diagnostic group, F(2, 74) = 1.89, p = 0.17.

[0487] To determine whether performance and self-reported mental health data correlated with each other, Spearman's rank correlation (if data were non-normal) was performed. Table 30 below shows the results.

[0488] [Table 30]

[0489] In the autism group, the best fatigue intervention score (the intervention that produced the lowest fatigue ratings) was correlated with the best anxiety intervention score (the intervention that produced the lowest anxiety ratings) (r s = 0.71, p < 0.01, N = 18), and as a result, lower anxiety was associated with lower fatigue in the intervention SART. Similarly, best EoO was negatively correlated with best EoC (r s=-0.51, p=0.03, N=18), such that more errors of omission were associated with fewer errors of commission. None of the performance variables appeared to correlate with any of the self-reported mental health variables in the autism group.

[0490] In contrast, correlations in the non-autistic group revealed a different pattern. The best EoC was significantly higher than the RTEoO (r s = 0.-0.46, p = 0.04, N = 21) and RTEoC (r s =-0.45, p=0.04, N=21), such that more errors of commission were associated with faster reaction times for errors of omission and, unsurprisingly, faster reaction times for errors of commission. Best RTEoO was also highly associated with best RTEoC (r s = 0.79, p < 0.01, N = 21). Interestingly, the best EoC was also associated with the best fatigue intervention score (r s =-0.56, p<0.01, N=21), suggesting that when participants felt less fatigued, they made fewer commission errors.

[0491] Multiple linear regression with Enter's method was used to predict EOC performance of the intervention SART from participants' anxiety and fatigue levels, education level, diagnostic group, and best intervention type. Preliminary analyses suggested that the assumption of normal distribution was met (Shapiro-Wilk test; p = 0.17). Furthermore, the assumption of multicollinearity was met as none of the predictor variables correlated with each other more than 0.7. However, none of the predictor variables correlated with the dependent variable more than 0.3. Cook's distance test revealed that there was one influential data point, which was subsequently removed from the analysis to ensure that Cook's distance remained less than 1 and the standardized residuals remained between -3 and 3. The model did not explain a statistically significant amount of variance in EOC performance, F(5,32) = 0.83, p = 0.54, R2 = 0.12, R 2 調整 =-0.02. The variables are shown in Table 31.

[0492] [Table 31]

[0493] Multinomial logistic regression was performed to model the relationship between diagnostic group, age, education, and sex and the best performing intervention. This regression was repeated such that the reference category varied across intervention types (alerts, filters, guidance, combined), as shown in Table 32.

[0494] [Table 32]

[0495] Regression yielded the same results. Adding predictors to the model significantly improved the fit between the model and the data, χ 2 (28, N=39)=43.40, Nagelkerke R 2 =0.70, p=0.03. A significant unique contribution was found for diagnostic group (χ 2 =16.95, p=0.02) and education (χ 2 = 18.31, p = 0.01). The goodness of fit was examined using the Pearson chi-square statistic, which was not significant, indicating a good fit, and 2 =141.61, p=1.00.

[0496] Post hoc chi-square tests of independence were then performed to identify diagnostic group and educational characteristics associated with best intervention type. There did not appear to be a significant relationship between diagnostic group (autistic vs. non-autistic) and best intervention type, with χ 2 (7, N = 39) = 3.87, p = 0.80), there also appeared to be no significant relationship between education (university vs. non-university) and best intervention type, χ 2(7, N = 38) = 11.43, p = 0.12).

[0497] Binary logistic regression was performed to determine the effect of diagnostic group, age, sex, or education on the likelihood that participants performed best using the alarms. Preliminary analyses suggested that the assumption of multicollinearity was met (tolerance = diagnostic group: 0.88; age: 0.67; sex: 0.93; education: 0.62). Inspection of standardized residual values ​​revealed that there were two outliers (standardized residuals -2.24; 2.56) that were retained in the data set. The model was not statistically significant, with χ 2 (5,N=39)=4.82, p=0.44, suggesting that the model was unable to distinguish between those who performed best with the alarm and those who did not. The model explained between 11.6% (Cox&Snell R square) and 16.4% (Nagelkerke R square) of the variance in the dependent variable and correctly classified 69.2% of the cases. As shown in Table 33, none of the predictors contributed significantly to the model.

[0498] [Table 33]

[0499] Binary logistic regression was performed to determine the effect of diagnostic group, age, sex, or education on the likelihood that participants performed best using the filter. Preliminary analyses suggested that the assumption of multicollinearity was met (tolerance = diagnostic group: 0.88; age: 0.67; sex: 0.93; education: 0.62). Inspection of standardized residual values ​​revealed no outliers. The model was not statistically significant, with χ 2(5,N=39)=5.93, p=0.31, suggesting that the filter was unable to distinguish between those who performed best and those who did not. The model explained between 14.1% (Cox&Snell R square) and 19.6% (Nagelkerke R square) of the variance in the dependent variable and correctly classified 66.7% of the cases. As shown in Table 34, none of the predictors contributed significantly to the model.

[0500] [Table 34]

[0501] Binary logistic regression was performed to determine the effect of diagnostic group, age, sex, or education on the likelihood that participants performed best using guidance. Preliminary analyses suggested that the assumption of multicollinearity was met (tolerance = diagnostic group: 0.88; age: 0.67; sex: 0.93; education: 0.62). Inspection of standardized residual values ​​revealed no outliers. The model was not statistically significant, with χ 2 (5,N=39)=2.64, p=0.76, suggesting that it was not possible to distinguish between those who performed best with the guidelines and those who did not. The model explained between 6.5% (Cox&Snell R square) and 9.4% (Nagelkerke R square) of the variance in the dependent variable and correctly classified 76.9% of the cases. As shown in Table 35, none of the predictors contributed significantly to the model.

[0502] [Table 35]

[0503] Binary logistic regression was performed to ascertain the effect of diagnostic group, age, sex, or education on the likelihood that participants performed best using the intervention combination. Preliminary analyses suggested that the assumption of multicollinearity was met (tolerance = diagnostic group: 0.88; age: 0.67; sex: 0.93; education: 0.62). Inspection of standardized residual values ​​revealed no outliers. The model was not statistically significant, with χ 2 (5,N=39)=0.39, p=1.00, suggesting that the combination did not distinguish between those who performed best and those who did not. The model explained between 1.0% (Cox&Snell R square) and 1.4% (Nagelkerke R square) of the variance in the dependent variable and correctly classified 66.7% of the cases. As shown in Table 36, none of the predictors contributed significantly to the model.

[0504] [Table 36]

[0505] A series of one-way ANOVA tests were conducted to determine whether there were significant differences in fatigue and distractibility scores based on various demographic categories. Meeting the assumption of homogeneity of variance for the non-autistic group, there was no significant effect of age on fatigue, F(3,158) = 0.43, p = 0.73, or distractibility, F(3,158) = 2.23, p = 0.09. In contrast, for the autism group, the assumption of homogeneity of variance was violated, so analyses were performed using Welch's test, which showed a significant effect of age on fatigue in the autism group, t(30.78) = 12.53, p < 0.01, as well as a significant effect of age on distractibility, t(32.99) = 7.83, p < 0.01. Post hoc comparisons using Games-Howell tests revealed that for fatigue, 30-39 year olds scored significantly lower than 18-21 year olds (mean difference = 1.01, p < 0.01) and significantly lower than 22-29 year olds (mean difference = 0.49, p = 0.01). Furthermore, 41-49 year olds scored significantly lower than 18-21 year olds (mean difference = 1.26, p < 0.01) and significantly lower than 22-29 year olds (mean difference = 0.74, p = 0.04). For distractibility, post hoc Games-Howell tests revealed that 41-49 year olds scored significantly lower on distractibility compared to 18-22 year olds (mean difference = 0.61, p < 0.01) and 22-29 year olds (mean difference = 0.41, p = 0.02). Furthermore, scores of 31-39 year olds were significantly higher than those of 18-22 year olds (mean difference = 0.47, p = 0.01) and 22-29 year olds (mean difference = 0.26, p = 0.02).

[0506] Given that the assumption of homogeneity of variance based on the median was met for both diagnostic groups, a one-way ANOVA revealed no significant differences between genders with respect to fatigue scores (Autistic: F(2,184) = 2.83, p = 0.06; Non-Autistic: F(2,159) = 0.85, p = 0.43). Similarly, in the non-autistic group, gender did not differ significantly with respect to distractibility scores, F(2,159) = 0.46, p = 0.64. However, in the autism group, gender differed significantly on distractibility scores, F(2,184) = 6.55, p < .01. Post hoc Tukey tests revealed that autistic participants who had attended graduate school rated distractibility significantly more than participants without the condition (mean difference = -0.70, p < 0.01) and participants with the condition (mean difference = -0.59, p = 0.02). Finally, for the non-autistic group, there were no significant differences between education groups for fatigue scores, F(2,159) = 0.56, p = 0.57. However, the education groups differed significantly in distractibility scores in the non-autistic group, F(2,159) = 4.47, p = .01. Post hoc Tukey tests revealed that those without a degree had significantly higher distractibility scores than those who had attended graduate school (mean difference = 0.40, p = 0.01).For the autism group, the education group was significantly different in fatigue scores, F(2,184) = 5.03, p = 0.01. A post-hoc Tukey test revealed that participants who had attended graduate school had significantly lower scores on fatigue than those who had not (mean difference = 0.64, p = 0.01).For distractibility, the Autism Education group was not significantly different using a Welch test, t(41.93) = 1.85, p = 0.17.

[0507] One-way ANOVA revealed that fatigue scores did not differ significantly between genders (autistic: F(2,184) = 2.83, p = 0.06; non-autistic: F(2,159) = 0.85, p = 0.43). Similarly, fatigue scores were significantly different for non-y, F(2,184) = 5.03, p = 0.01. A post-hoc Tukey test revealed that participants who had attended graduate school had significantly lower scores on fatigue than those who had not (mean difference = 0.64, p = 0.01).For distractibility, the Autism Education group was not significantly different using a Welch test, t(41.93) = 1.85, p = 0.17.

[0508] Table 37 shows the mean scores for the three sensory sensitivity variables, as well as the mean anxiety, fatigue, and attention scores across the two diagnostic groups. Independent samples t-tests were performed to compare the scores for each variable across the diagnostic groups. Autistic adults reported significantly higher levels of visual sensitivity, t(313.31)=-7.83, p<0.01, and physiological sensitivity, t(292.46)=-6.09, p<0.01, compared to the non-autistic group. Autistic individuals also reported lower levels of fatigue (meaning they tend to be more alert) compared to the non-autistic population, t(347)=5.13, p<0.01. There were no significant differences in mean scores for other variables between the groups.

[0509] [Table 37]

[0510] Table 38 shows the Bonferroni adjusted Spearman rank correlations between the variables in both the non-autistic and autistic groups. Application of the Kolmogorov-Smirnov test revealed that most variables were non-normal, except for vision and anxiety in the non-autistic group. In the non-autistic group, significant associations were found between each sensory sensitivity variable and the outcome variables (anxiety, fatigue, and distractibility), as well as between all sensory sensitivity variables. A similar pattern of associations was found for the autistic group, except that the relationship between vision and hearing was not significant (r=0.11, p=0.13).

[0511] [Table 38]

[0512] Given the significant associations between the variables in Table 38, a series of linear regressions were performed to analyze the unique contribution of each variable in predicting fatigue and distractibility, first checking for multicollinearity. Visual, audio, and physiological sensitivity variables, as well as demographic variables (age, sex, education) were added as predictors of fatigue scores. For the non-autistic group, Table 39 shows that none of the demographic variables predicted fatigue, but all sensory variables uniquely predicted fatigue, with the model accounting for 47% of the variance in fatigue (F(6,155)=23.07, p<0.01). For the autistic group, Table 40 shows that all sensory variables as well as age uniquely predicted fatigue, but education and sex did not (F(6,180)=64.39, p<0.01). The model accounted for 68% of the variance in fatigue.

[0513] [Table 39]

[0514] [Table 40]

[0515] Table 41 shows the regression predicting distractibility from sensory and demographic variables for the non-autistic group. The model predicted 41.9% of the variance in distractibility, F(6,155)=18.65, p<0.01. Both visual and physiological sensory variables were significant unique predictors of fatigue. In contrast, as shown by Table 42, the model predicting distractibility for the autism group showed a different pattern, F(6,180)=47.35, p<0.01, predicting 61.2% of the variance in distractibility. All variables except education were significant unique predictors of distractibility.

[0516] [Table 41]

[0517] [Table 42]

[0518] The next regression again predicted distractibility, but included fatigue. For the non-autistic group, the model predicted 43.2% of the variance in distractibility, F(7,154) = 16.72, p < 0.01. Visual, physiological, and educational were significant unique predictors of distractibility. However, other variables, including fatigue, were not significant predictors of distractibility. In contrast, in the autistic group, a regression predicting distractibility including fatigue showed that not only did the model predict 70.5% of the variance in distractibility (F(7,179)=61.19, p<0.01), but also that fatigue, auditory, and physiological sensory variables were unique predictors of distractibility. The results for the non-autistic group are shown in Table 43. Table 44 shows the results for the autistic group.

[0519] [Table 43]

[0520] [Table 44]

[0521] Mediation analyses were performed to investigate the indirect effect of fatigue on the relationship between distractibility and the auditory as well as physiological sensitivity variables (when both variables were unique predictors of distractibility). Given that fatigue did not appear to be a significant predictor of distractibility in the non-autistic group, analyses were performed only for the autistic group. First, the direct effects of hearing and then physiology separately on fatigue were calculated by bivariate regression. There were significant results for both (hearing: B = -0.35, SEB = 0.08, p < 0.01; physiology: B = -0.77, SEB = 0.06, p < 0.01). A multiple regression including both fatigue and hearing (Fatigue: B=-0.46, SEB=0.03, p<0.01; Hearing: B=0.14, SEB=0.03, p<0.01), as well as separate regressions including both fatigue and physiology (Fatigue: B=-0.39, SEB=0.04, p<0.01; Physiology: B=0.17, SEB=0.04, p<0.01) were run to predict distraction. Sobel tests were then performed to test the indirect effects for statistical significance in each model. For the model including hearing, as shown in Figure 32, the indirect effect was z=4.13, p<0.01, concluding that partial mediation occurred between hearing sensation on distraction due to fatigue. For the model including physiology, as shown in Figure 33, the indirect effect was z = 8.12, p < 0.01, suggesting that partial mediation occurred between physiological sensations on fatigue-induced distraction.

[0522] Given that physiological sensory sensitivity, fatigue and AQ scores were non-normally distributed, Spearman's rank correlation was performed to determine the relationship between AQ scores and study variables (Table 45).

[0523] [Table 45]

[0524] AQ scores were only collected from 140 non-autistic individuals in the study, so only these participants were included in the analysis. Distractibility was the only variable significantly associated with AQ scores (r=0.38, p<0.01), suggesting that non-autistic individuals with higher autistic symptoms were more likely to score higher on distractibility.

[0525] AQ scores were then included in a regression model predicting distractibility in non-autistic participants. The results are shown in Table 46. This model accounted for 53% of the variance in distractibility, F(8,131) = 18.07, p < .01. Vision, physiology, and education remained significant predictors of distractibility. AQ scores also appeared to be significant predictors of distractibility. Interestingly, however, fatigue now appeared to be a significant predictor of distractibility when AQ was included in the model compared to the model without AQ (Table 43). Such a change implies that there may be a confounding effect of AQ.

[0526] [Table 46]

[0527] 34A-34B show summary results of the PPI study described herein, which provided a basis for conducting and refining the SART / WOz clinical study described herein. As shown in FIG. 34A, different associations between individual demographics (e.g., age, sex, education) and anxiety, fatigue, and / or concentration were found in the autistic group versus the neurotypical group. Also, the autistic group was found to be more sensitive to visual and physiological stimuli and tend to be less fatigued. As shown in FIG. 34B, in some cases, sensory sensitivity was found to be mediated by anxiety and fatigue in different ways in the autistic and non-autistic groups. For example, fatigue was found to significantly mediate the indirect and direct relationship between hearing sensitivity and distractibility in autistic individuals, but no such relationship was found in non-autistic individuals. Some of the regression analysis findings in the PPI study included the following: For distractibility alone and for non-autistic participants, both visual, physiological, and education were predictors. For autistic participants, all variables were predictors except education. For predicting fatigue only and for non-autistic participants, all sensory variables were predictors (demographics were not). For autistic participants, all variables were predictors except for gender and education. For predicting anxiety only and for non-autistic participants, all sensory variables were predictors and gender. For autistic participants, all variables were predictors but demographic data were not. For fatigue and anxiety prediction and non-autistic participants, anxiety and gender were predictors. For autistic participants, all visual, physiological, anxiety and age were predictors. - Vision, physiology, and education (but not demographics) were predictors of anxiety and fatigue in non-autistic participants. For autistic participants, hearing, physiology, and fatigue were predictors. For non-autistic participants, physiological, education and anxiety were predictors of anxiety and for autistic participants, physiological and anxiety were predictors. For distractibility, predictors of fatigue and anxiety in non-autistic participants were vision, physiology, and anxiety. For autistic participants, predictors were hearing, physiology, anxiety, and fatigue. For the prediction of inattention by the AQ-10 test only, and for non-autistic participants only, physiological, educational, anxiety, and AQ were predictors. Autistic participants were excluded as they did not take the AQ-10.

[0528] Figures 35A-B show summary results of the SART / WOz clinical study described herein. As mentioned above, the study included multiple trials that tested different mediations / interventions and different sensory cues. As shown by Figure 35A, all mediations / interventions (e.g., alerts, filters, guidance, or combinations) were found to improve anxiety in both autistic and non-autistic individuals. In contrast, only some mediations (e.g., filters, or combinations) were found to improve fatigue, while other mediations (e.g., guidance) were found to be potentially harmful. In each case, mediations customized to the specific individuals in each group were found to be most effective. Although it was initially expected that only individuals in the autism diagnosis group (i.e., the experimental group) would benefit from the digital mediation, surprisingly, there was also an improvement in performance in the control group of non-autistic participants. Thus, benefits of reduced errors, improved timing, and / or improved performance due to reduced anxiety and fatigue were found in both groups.

[0529] As shown in FIG. 35B, the overall performance of all users, as measured by reduced EoC, reduced EoO, better reaction time of EoC (slower), and response time of EoO (slower), is improved with personalized mediation. In this case, the percentage change in error change or increase in reaction / response time is measured. As shown, “Baseline to Distraction” shows the change from baseline (no distraction cues present) to added distraction. Regardless of whether the mediation was added to the baseline or after the distraction, all users’ performance improved when the mediation was personalized. As mentioned before, one of the mechanisms to achieve performance improvement is expected to be that personalized mediation slows users down (e.g., as measured by increased reaction time) and allows participants to stay focused longer, resulting in greater mindfulness. This was found to be especially true for autistic participants. In practice, this increase in response / reaction time can be on the order of tens of ms (e.g., 50 ms) depending on the given task the user is performing.

[0530] As the aforementioned PPI study and SART / WOz clinical study show, optimal mediation for a given user can be predicted taking into account user-specific data and correlations between the user's performance and the user's characteristics, such as demographics, sensory sensitivity, and autistic or non-autistic status. Thus, the multi-assistive wearable technology described herein can better initialize and / or optimize one or more sensory thresholds and mediations associated with a user. For example, FIG. 36A is an operational flow diagram illustrating an exemplary method 3600 for initializing and iteratively updating one or more sensory thresholds and one or more mediations associated with a particular user. In some implementations, the method 3600 can be implemented by one or more processors of the wearable device system (e.g., one or more processors of the wearable device 10 and / or mobile device 20) executing instructions stored in one or more computer-readable media (e.g., one or more computer-readable media of the wearable device 10 and / or mobile device 20).

[0531] Act 3601 includes obtaining demographic data of a user of the wearable device, which may include receiving user input at a user interface indicating the user's age, education level, gender, or other demographic data.

[0532] Operation 3602 includes obtaining user sensory sensitivity data indicative of whether the user is visually sensitive, ultrasonically sensitive, or interoceptively sensitive. This may include receiving user input at a user interface indicative of whether the user of the wearable device is visually sensitive, ultrasonically sensitive, or interoceptively sensitive. In some cases, the user input may include input at a GUI including one or more responses by the user to one or more prompts indicative of whether the user is visually sensitive, ultrasonically sensitive, and / or interoceptively sensitive. These responses may indicate the user's preferences for particular sensory inputs, such as stimuli that the user likes, stimuli that cause discomfort to the user, the user's perception and / or measured sensitivity to different stimuli, etc. In some implementations, the responses may include responses to questions such as those described with reference to the studies above.

[0533] Operation 3603 includes obtaining neurodiversity data indicating whether the user is neurodiverse or neurotypical. For example, the neurodiversity data can indicate whether the user is autistic or non-autistic. In some implementations, the system can store a first identifier indicating whether the user is neurodiverse or neurotypical. In some implementations, the neurodiversity data can be obtained by a user input at a user interface indicating whether the user has been diagnosed as neurotypical. In other implementations, described further below, the wearable device system can be configured to perform a method for providing a diagnostic prediction of whether the user is neurodiverse or neurotypical.

[0534] Operation 3604 includes initializing and storing one or more sensory thresholds and one or more mediations associated with the user. The thresholds and mediations associated with the user can be based on the user's sensory sensitivity data, demographic data, and / or neurodiversity data. In some cases, the demographic data can be ignored.

[0535] Act 3605 includes collecting sensor data and environmental data while the user is wearing the wearable device.

[0536] Operation 3606 includes modifying one or more sensory thresholds and one or more mediations associated with the user in response to collecting sensor data and / or environmental data while the user is wearing the wearable device. As shown, operations 3605-3606 can be repeated over time as the user utilizes the wearable device system to provide sensory relief. The frequency with which the one or more sensory thresholds and one or more mediations are updated in response to newly collected data can be configurable, system-defined, and / or user-defined. For example, the updates can depend on the amount of data collected and / or the amount of time that has elapsed. In some implementations, operations 3605-3606 can be skipped. For example, a user can disable updating the thresholds and / or mediations based on actual use of the wearable device.

[0537] As the aforementioned PPI study and SART / WOz clinical study also show, anxiety and fatigue can mediate sensory sensitivity differently in autistic and non-autistic users. In some implementations, features found to correlate with autistic vs. non-autistic users can provide a basis for training a model that outputs a prediction (e.g., as a likelihood / probability) that a given feature corresponding to a user (e.g., sensory sensitivity feature, anxiety feature, fatigue feature, demographic feature, etc.) is autistic or non-autistic. For example, Figures 36B and 36C are operational flow diagrams illustrating exemplary methods 3610, 3620 for predicting whether a user is neurodiverse (e.g., autistic) or neurotypical. In some implementations, method 3610 or method 3620 may be performed by one or more processors of the wearable device system (e.g., one or more processors of wearable device 10 and / or mobile device 20) executing instructions stored on one or more computer-readable media (e.g., one or more computer-readable media of wearable device 10 and / or mobile device 20).

[0538] In method 3610, operation 3602 may be performed as described above. Operation 3611 includes deriving one or more sensory sensitivity scores, including a visual sensitivity score, a sonic sensitivity score, and / or an interoceptive sensitivity score, based on the sensory sensitivity data. For example, the one or more scores (e.g., normalized on a scale such as 0-100) may be derived based on the user's responses to the prompts. Operation 3612 includes obtaining anxiety data measuring a general anxiety level of the user. For example, this may include receiving, in a GUI, one or more responses by the user to one or more prompts indicating the user's anxiety level in different situations. Operation 3613 includes deriving an anxiety score based on the anxiety data. For example, the score (e.g., normalized on a scale such as 0-100) may be derived based on the user's responses to the prompts. Operation 3614 includes using the trained model to predict the likelihood that the user is neurodiverse based on the one or more sensory sensitivity scores and the anxiety score. The model can be configured / trained to predict the probability of autism based at least on features including a given user's anxiety level / score and one or more sensory sensitivity levels / scores. Each feature can be weighted differently. It should be noted that the model can also be trained to consider other features (e.g., demographic data) when making the prediction. Operation 3615 includes making a determination that the user is neurodiverse or neurotypical and storing the associated identifier. For example, if the prediction output by the model meets a threshold (e.g., >80% probability), a prediction can be made that the user is autistic. In some implementations, the system can validate the prediction by measuring the user's performance in response to a particular task in the presence and absence of mediation. This performance can be measured using a wearable device and / or a mobile device by administering the SART as described above. Given a particular mediation, the level of improvement in the user's performance can further validate whether the predicted diagnosis is correct or incorrect.

[0539] In method 3620, operations 3602, 3611, and 3615 may be performed as described above. Operation 3621 includes obtaining fatigue data measuring a general fatigue level of a user 3621. For example, this may include receiving, in the GUI, one or more responses by the user to one or more prompts indicative of the user's fatigue level in different situations. Operation 3622 includes deriving a fatigue score based on the fatigue data. For example, a score (e.g., normalized on a scale such as 0-100) may be derived based on the user's responses to the prompts. Operation 3623 includes using the trained model to predict 3623 a likelihood that the user is neurodiverse based on one or more sensory sensitivity scores and the fatigue score. The model can be configured / trained to predict the probability of autism based at least on features including a given user's fatigue level / score and one or more sensory sensitivity levels / scores. Each feature can be weighted differently. It should be noted that the model can also be trained to consider other features (e.g., demographic data) when making the prediction. In some implementations, both anxiety and fatigue features can be considered in the trained model.

[0540] In some implementations, the multisensory assistive wearable technology described herein can be implemented using a network topology that ensures privacy of user data and facilitates ethical relationships between device layers, systems, and parties, such as the user / wearer, the user's family, the user's therapist, and / or the user's general practitioner. To this end, edge and fog computing can be implemented using devices located at the periphery of the system to facilitate and secure cloud connectivity. These devices can be independent and can connect to both sensors and applications while acting as components, software, and data transceivers between the cloud only when necessary. This can provide desirable reliability constraints for data computation, especially since sensitive data is substantial, often unregulated, and can be subject to exploitation. Due to cloud limitations on exposure, fog computing can provide an additional layer of efficiency and security.

[0541] 37 illustrates an exemplary system architecture / topology for performing fog data processing according to some implementations of the present disclosure. The system architecture includes IoT sensors 3710, an edge layer 3720 including edge nodes 3721, a fog layer 3730 including fog nodes 3731, and a cloud layer 3740 including one or more cloud computing devices 3741. Although FIG. 37 is described primarily in the context of a system architecture that applies to a single user / wearer, it should be understood that this system architecture can be extended to multiple independent users.

[0542] The IoT sensor 3710 may be a sensor implemented as part of a wearable device (e.g., wearable device 10 or wearable device 500). For example, the sensors may include pupil measurement sensor 204, galvanic skin sensor 205, inertial motion unit 206, temperature sensor 309, audio sensor 309, image sensor (e.g., as part of camera 550), etc. IoT sensors 3710 may also include sensors implemented in the same environment as the wearer but on a different device. For example, sensors may include sensors implemented on mobile device 20 (e.g., GPS or motion sensors), ambient temperature sensors, image sensors on an external IoT device, audio sensors on an external IoT device, etc.

[0543] The edge node 3721 and fog node 3731 may be implemented in hardware including, but not limited to, a client-side wearable device (e.g., wearable device 10 or 500), a mobile device 20, and / or a local (i.e., pre-cloud) operating server or database device that may be provided by a wearable device system provider.

[0544] As shown, the fog layer 3730 exists between the edge layer 3720 and the cloud layer 3740. In some implementations, edge nodes can be between cloud nodes and fog nodes. In some implementations, some edge nodes are between cloud nodes and fog nodes, and some fog nodes are between edge nodes and cloud nodes. When fog node 3731 receives data from edge node 3721, it can filter the data by deterministically passing only appropriate data to cloud computing device 3741 for processing, storage, networking, etc. For example, edge and fog computing can be implemented where ecological parameters (e.g., temperature) or physiological parameters (e.g., heart rate) are periodically sensed and collected as data (e.g., every second of movement) to align the user's fatigue and anxiety with other ecological / physiological measures. In the absence of fog layer 3730, any sensor measurements can potentially be sent to a cloud application to support downstream monitoring for the user / wearer, as well as a therapist, general practitioner, and / or family member. The rules-based fog layer 3730 can prevent this excessive data transfer from congesting the network and / or compromising user privacy / security. For example, fog nodes 3731 can be configured to pass on only critical data as it occurs (e.g., excessive temperature spikes) or only data collected by specific sensors (e.g., image or audio data is not made available to the cloud).

[0545] The fog node 3731 or edge node 3721 may also encrypt any data before making it available to the cloud computing device 3741 so that the information can remain pseudonymized, thereby protecting the user's privacy. In operation, all encryption, decryption, and purging of data can be done locally at the user level without the use of cloud software or hardware.

[0546] The edge node 3721 may be responsible for maintaining the location of the middleware that manages data flow, encryption / decryption, and ultimately deleting the data when it is no longer needed. In some implementations, the same device may function as both a fog node and an edge node.

[0547] By locating data management and processing in the "fog", various benefits can be realized by stakeholders (e.g., users) and / or wearable device systems. First, data processing close to the end user can improve data access and allows buffering storage from expensive, inefficient, and unsecure / unethical activities. For example, this can prevent data collection for marketing or sales. "Fogging" by centralizing data processing within the wearable device or local network of wearable devices can also bypass slow connections, i.e., expensive and slower cloud transmission rates are not required for grounded applications. Furthermore, by localizing data management / processing / storage, data protection and privacy rules can be controlled and managed by the user, allowing configuration of what can and cannot be collected, transmitted, and / or stored. For example, location can be performed within a LAN and / or ad-hoc network of wearable devices (e.g., 10) and / or mobile devices (e.g., 20) coupled to the wearable devices.

[0548] In some implementations, the cloud tier 3740 can include a data lake (DL) repository that stores non-personalized machine learning (ML) data, including images, audio, and / or video. Some or most of this data can be in the public domain. In operation, the fog tier 3730 can compare private and nuisance conditions (e.g., as determined from data collection by wearable devices) to data stored in the cloud tier 3740. In this configuration, the edge layer 3720 can regulate data flow to the cloud layer 3740, allowing only the most limited flow to the cloud, while the fog layer 3740 can be used to detect user distractions based at least in part on a repository of data stored in the cloud layer 3740. Thus, the system can operate without the cloud layer personally identifying users.

[0549] In some implementations, all personalized user data can be stored at a local level, including thresholds, sensory resolution, mediation, demographic data, diagnostic data, etc. Distraction data of deep learning and machine learning data (e.g., auditory, visual, etc.) can be stored globally encrypted, while real-time comparative responsiveness to ecological and physiological data can be stored locally and temporarily.

[0550] 38A illustrates a specific example of a wearable system architecture, including data flow, leveraging fog and edge computing in accordance with some embodiments of the present disclosure, and FIG. 38B is a flow diagram illustrating operations performed by the system of FIG. 38A in accordance with some implementations.

[0551] The system of FIG. 38A includes a wearable device 10, one or more edge services 3810, a fog service 3820, a gateway 3830 that can mediate communications between the edge server 3810 and the fog service 3820, and one or more cloud computing devices 3840. In some implementations, the functionality of the edge server 3810 may be implemented in the wearable device 10 or a mobile device 20 communicatively coupled to the wearable device 10.

[0552] Operation 3901 includes the wearable device 10 collecting sensor data. For example, one or more sensors of the wearable device 10 can be used to record sensory input stimuli to a user. This can include sensing ecological and physiological / psychophysiological data as described above. In some implementations, other devices (e.g., mobile device 20) other than the wearable device 10 and in the same environment as the wearable device 10 collect the sensor data.

[0553] Operation 3902 includes one or more fog nodes of fog service 3820 processing, storing, and / or managing the collected sensor data. In some implementations, the one or more fog nodes include a data store that stores and / or manages the sensor data. In some implementations, the one or more fog nodes include a data store that stores one or more sensory thresholds specific to a user of the wearable device 10 (e.g., one or more sensory thresholds selected from auditory, visual, or physiological sensory thresholds). In some implementations, the one or more fog nodes compare the sensory input stimuli to the one or more sensory thresholds specific to the user to determine that intervention may be required.

[0554] Operation 3904 includes the edge server 3810 encrypting and uploading the data to one or more cloud computing devices 3840. For example, if the fog service 3820 determines that a threshold has been met after considering a subset of the sensor data, the subset of sensor data that triggered the determination may be encrypted by the edge server 3810 and uploaded to the cloud.

[0555] Operation 3905 includes applying a data processing and machine learning pipeline / process. The pipeline can be executed using at least one or more cloud computing devices. Operation 3907 includes presenting an intervention / mediation to the user. As an example, the user is visually unable to focus, causing a change in pupil measurement. The updated pupil measurement can result in a threshold being met that causes a mediation / intervention to be presented to the user (e.g., alerting the user to change focus). Mediation can be triggered as follows: During a pupil measurement, an outward-facing camera (e.g., integrated into the wearable device) images a distracting object (e.g., the camera captures an image in the direction of the pupil gaze). If the image matches or is sufficiently similar (e.g., determined by calculating a similarity score based on image features) to a published and stored image on the cloud of the same / similar object that was previously tagged as a personalized trigger for the user as a distracting cue, then mediation can be triggered. Using a machine learning pipeline, the captured images can be collated against a data store in the cloud, and the images can be checked against each other as cluttered images are processed based on different parameters, including color, shape, edge detection, etc., to make the comparison.

[0556] In some implementations, the FBDL model can be used to generate customized mediations given data from one or more sensors as input. Deep learning is a machine learning category that uses neural network algorithms to store data for analysis and prediction. Neural networks use hidden layers to obtain features by connecting with each other for replicable results (output layer). FBDL restricts the connections between the input layer and the hidden layer so that every veiled unit is attached to a subsection of its corresponding input. Thus, lower dimensional features can be derived by arbitrarily sampling the big data. FIG. 47 illustrates an example of an FBDL model, according to some implementations of the present disclosure. As shown, the input can be one or more different types of sensor data, including audio data, pupil data, IMU data, GSR data, optical data (e.g., image data), temperature data, etc. Through multiple layers, FBDL can be trained to recognize specific / personalized mediation types based on the input data and depending on the recognition, the mediation can be an alert, a filter, guidance, or a combination thereof.

[0557] Open Learner Model

[0558] Various implementations of the technologies described herein can leverage OLM to graphically represent (e.g., using a GUI) the current progress of a user (e.g., a neurodiverse individual) of a multisensory assisted wearable technology so that the user or other interested parties (e.g., therapists) can visualize, track, and / or reflect on their progress. OLM components can be incorporated into the systems and methods described herein to improve an individual's super, low, and sensory-seeking tasks, which can impact task accuracy (i.e., performance) and mental health (i.e., calmness and attention), especially when distracted by ecological or psychophysiological cues. Fully transparent mediation can provide better results than those that limit user data access, direct system control, reliability, and technology trustworthiness.

[0559] While most of the OLM data collected and used in some prior art is fully open (e.g., disclosed) to many individuals with minimal safeguards for security and privacy, the technology described herein incorporates OLM into a system designed to promote security, privacy, and efficient use of computational resources, including bandwidth and storage. For example, the systems and methods described herein can utilize OLM in conjunction with fog and edge networking, as previously described. Additionally, access control can be configured using a GUI associated with the OLM model.

[0560] 39-45 are diagrams illustrating an OLM framework according to some implementations of the present disclosure. The illustrated OLM framework includes three tables (FIGS. 40, 42, 44) and four flow charts (FIGS. 39, 41, 43, 45). The OLM framework illustrates custom labeled characteristics related to the securing of data by individual users / wearers and their support (e.g., therapists, family, etc.). In operation, the wearer can select / actuate controls on the GUI to determine the amount of data that can be sensed, collected, processed, and / or shared per characteristic. As illustrated, the characteristics or features can be divided into which elements are important and should be sensed, mediated, and / or stored, how this is accomplished, and the access permissions to review and manage these functions.

[0561] Figure 39 is a high-level flow diagram of the OLM framework. This exemplary OLM framework defines eleven elements in three categories (i.e., model accessibility, presentation, access methods, accessibility control, etc.), their corresponding characteristics (i.e., complete, partial, present, future, etc.), and their descriptions (i.e., textual descriptions of each objective element) that are used in defining a particular OLM (see the tables from left to right in Figures 40, 42, and 44). These characteristics signify the levels of accessibility objective elements across eleven aspect columns (i.e., from left to right, including right to access, control by trust, evaluation, etc.). Each aspect uses ranking levels to distinguish elements from one another (i.e., those considered critical or particularly critical are marked with X and XX, respectively; those considered debatable are marked with the following; those not relevant are left blank).

[0562] One of the OLM maps describes "what is available" (Figures 40-41) by addressing the degree of model accessibility, access to the underlying representation, access to uncertainty, role of time, access to source issues, and access to personalization of the model. As an example, the degree of model accessibility (item #1) is primarily "fully" open across the board, with significant availability to nearly all stakeholders.

[0563] One of the OLM maps describes "how the model is presented" to stakeholders, including friends and acquaintances (Figures 42-43). Presentation details (i.e., word cloud, skill meter, radar plot, etc.), access methods (i.e., inspectable, editable, user-to-system persuasive, etc.), and flexibility of access are included. Compared to the "What's Available" table, this table includes elements tagged with rankings of importance, especially importance.

[0564] One of the OLM maps describes "Who controls access" (Figures 44-45) and discloses two objective elements that map focus (i.e., whose accessibility comes from) and governing access (i.e., who controls access to others).

[0565] Some implementations of the multisensory assisted wearable technology described herein may leverage the AR-supported framework of development, analysis, and evaluation criteria. In this case, AR support may refer to the use of AR to ultrasonically or visually replace certain auditory or visual information presented to the user, such as blurring, pushing away, or erasing unpleasant images, or performing digital signal processing of the audio signal to avoid distraction. The framework may provide a mechanism for implementing an improved OLM, a Quantified Self (QS) framework, and / or a Multimodal Learning Analytics (MMLA) framework. To this end, FIG. 46 illustrates a system implementing an AR-based MMLA framework, according to some implementations of the present disclosure. As illustrated, the system is configured to perform at least three functions for the user and / or other parties of the multisensory assisted wearable technology described herein: battery, diagnostics, and personalization; purpose, goals, and iterative results; and mediation strategies and digital adaptations via the technology (e.g., using multimodal sensors and implementing intervention strategies). Depending on the configured access control, parties, including the user, may maintain accessibility throughout the framework.

[0566] As shown in FIG. 46, a general practitioner, therapist, instructor, employer, caregiver, and / or educator can help guide the three functions mentioned above. Sensory sensitivity, attention, mental health objectives, goals, and recurring results can be augmented using AR-augmented and / or OLM applications. The individual can be a neurodiverse or potentially neurodiverse individual who may be at risk in a social environment, an employee in the workplace, and / or a student in a higher education setting. A battery and diagnostics (i.e., MaRs-IB, ASRS, AQ-50, etc.) including an Individual Sensitivity and Mediation Profile (ISIP) can be developed for each user. As described herein, an ISIP can refer to a data profile provisioned by one or more mobile applications and used daily by a user to customize their responsiveness and provide behavior modification (e.g., using a Distraction Intervention Desire Questionnaire). The ISIP can help personalize a user's specific sensory threshold and / or sensory resolution as described herein, which can affect the alerts, filtering, and guidance interventions provided to a given user.

[0567] In some embodiments, the ISIP can utilize state-based anxiety and fatigue monitoring (SAFE) and randomized periodic feedback (FADE) to ensure ethical compliance, effectiveness, and user satisfaction. Data can be stored on the wearable device (e.g., wearable device 10), mobile device (e.g., mobile device 20), and / or other devices within a LAN or ad-hoc network of wearable / mobile devices and is available for OLM application analysis or stakeholder review. Due to the sensitive, contextual, and personalized nature of the data, the majority of the information can be localized and processed, whenever possible, using only edge and fog transmission as previously described with reference to Figures 37 and 38A-38B. Cloud processing, transmission, and storage can be minimized or avoided entirely to protect privacy / security and ensure ethical robustness. Further security can be enabled by encryption / decryption policies that provide an additional layer of protection whenever stakeholders review or process sensitive data.

[0568] FIG. 48 is a high-level flow diagram conceptually illustrating the operation of a multisensory assistance wearable system according to some embodiments of the present disclosure. The individual / wearer reacts to the environment. MMLA sensors (e.g., incorporated in the wearable device and / or some other device in the user's environment) collect data corresponding to ecological cues (e.g., temperature data, image data, etc.) and psychophysiological cues (e.g., pupil data, GSR data, heart rate data, etc.). The user's sensitivity profile, which may include thresholds and intervention / mediation preferences, is used to determine the intervention / medium, which may be an alert, a filter, guidance, or a combination thereof. The user's response continues to be measured after mediation. In this way, a feedback loop allows a constant consistent path to be developed, whereby the individual's response is weighted against the ecological and psychophysiological responses. When the personalization threshold is exceeded, an assistance or mediation event occurs and the system again monitors the individual's response and weighs it against the current sensory input.

[0569] In some implementations, the individual may receive mediation as long as the system determines that there is a mismatch in the signals that align with personalization goals that can be accommodated via the assistance means and that these address sensory, attentional, and / or mental health indicators. In some implementations, if the mediations are no longer effective or do not enhance the user's experience, they may be disabled by either the individual (e.g., via a user interface of the wearable or mobile device) or by an interested party.

[0570] In some implementations, cues that are no longer bothersome can be removed from the identification process.

[0571] Tables 47A-47B show some example design specifications, including latency parameters, for implementing auditory sensing, physiological / psychophysiological sensing, and transmission according to some implementations of the present disclosure. It should be understood that system specifications may vary depending on available hardware.

[0572] [Table 47A]

[0573] [Table 47B]

[0574] Purpose

[0575] The multi-sensory assistive wearable technology described herein can be utilized across a myriad of applications to provide a myriad of potential benefits. For example, in employment applications, the technology described herein can potentially reduce distraction, improve attention and performance, reduce anxiety, and / or increase employee output and / or satisfaction. Metrics that could potentially be improved in employment applications include improved onboarding and training of neurodiverse, autistic, and neurotypical applicants and new hires, reduced employee turnover, increased productivity rates, diversity and / or inclusion, increased benefits per employee, reduced healthcare costs and / or ROI, employee net promoter score, HR costs per employee, employee referrals, combinations thereof, and the like.

[0576] In academic applications, the techniques described herein can potentially increase focus and / or comprehension, reduce, minimize and / or substantially eliminate hesitation, and / or increase, enhance and / or increase comfort. Metrics that can potentially be improved in academic applications include retention rates (next tenure versus leaving), student grades, time to degree completion and / or credits to degree award, grades, educational goal tracking, academic reputation, and / or recent student underemployment.

[0577] In social applications, the techniques described herein can potentially increase participation and / or motivation and reduce anxiety. Metrics that may potentially be improved in social applications include primary socialization (learning attitudes, values, and / or behaviors appropriate to the individual and culture), secondary socialization (learning behaviors of smaller groups within society), developmental socialization (learning behaviors in social institutions and / or developing social skills), anticipatory socialization (rehearsing future positions, occupations, and / or relationships), and resocialization (discarding previous behaviors and / or adopting new patterns as part of a life transition).

[0578] In transport truck / trucking applications, the techniques described herein may increase and / or improve attention and / or performance, reduce fatigue, improve response times, etc. Potential improved metrics in transport truck / trucking applications include increased safety and / or productivity (such as shutting down engines, recommending breaks, crash data statistics and / or analysis), reduced logistical and / or financial burdens (reduced shipping, delivery times, and / or transportation costs), and logistical benefits including effective planning, dispatching, and / or scheduling.

[0579] In transport aircraft applications, the techniques described herein can potentially increase concentration and / or performance and reduce fatigue and / or anxiety. Metrics that may be improved in a transport aircraft setting include safety (e.g., fatality and / or accident rates, system risk events, runway incursions, hazard risk mitigation, commercial space launch accidents, global fatality rates), efficiency (tax in / out times, gate arrival / delay, gate to gate times, distance during level flight descent, flight vs. submit flight time, average flight distance, total arrival and / or departure delays, number of operations, on-time arrival, average fuel burn), capacity (average daily capacity and daily operations, runway pavement condition, NAS reliability), environmental (noise exposure, renewable jet fuel, NAW wide energy efficiency, emissions exposure), and / or cost-effectiveness (cost per unit of operation).

[0580] In IoT applications, the techniques described herein can potentially integrate with mechanical and digital machines, objects, animals, and / or people (each with a unique identifier) ​​that receive information transferred from the wearable so that actionable commands and / or analysis can be performed. Metrics that could potentially be improved in IoT applications include increased physiological / psychophysiological activity and can provide alerts to parents, caregivers, and / or professionals (such as paramedics) if wearable thresholds are exceeded. Integration into environmental control units (ECUs) can be bridged between the wearable and devices, including but not limited to TVs, radios, lights, VCRs, motorized drapes and / or motorized hospital beds, heating and / or ventilation units (air conditioners), washers and / or dryers.

[0581] In performance enhancement applications, the techniques described herein can potentially improve backlog, mental health, fatigue, anxiety, and / or focus. Metrics that can potentially be improved in performance enhancement applications include testing (logical processing, advocacy, curiosity, technical acumen, and / or tenacity), leadership (mentorship, subject matter expertise, team spirit, interpersonal skills, reliability), strategy and planning (desire, quality, community, knowledge and capabilities), intangibles (communication, diplomacy, negotiation, self-starter, confidence, maturity, and self-sacrifice).

[0582] In telemedicine, emergency medicine, and healthcare applications, the technology described herein can potentially improve the ability of medical and healthcare practitioners to share data with wearable users to assist with fine-tuning of treatment, Rx, dispatch for emergency assistance, monitoring and / or optimization of operating rooms, work schedules, and / or logistics strategies, pupillometry indicating unsafe conditions, unsafe warnings when thresholds are crossed (performance or physiological / psychophysiological). Metrics that may potentially be improved in telemedicine, emergency medicine, and / or healthcare applications include telemedicine metrics (e.g., consultation time, diagnostic accuracy, admission rates, quality of service / skills, patient and / or clinician retention, time and / or trips saved, adherence to treatment plan, patient referrals), surgical metrics (e.g., first case start, turnover time, space use / time, complications, value-based purchasing, consistency of service, outcomes), emergency metrics (e.g., average patient flow per hour, length of procedure / stay, relative value units vs. time, patient visibility, RVUs generated, current procedure terminology (CPT) performance, average rating and management distribution rates, total number of missing charts).

[0583] In parental, guardian, and / or educational monitoring applications, the technology described herein can potentially discourage metric planning (and protection), thereby enabling work-life balance by meeting actionable and measurable goals and deadlines for improving family dynamics, including being more present, aware, and / or tracking children's engagement (especially for exceptional individuals, but not limited to all abled neurodiverse children). Metrics that could potentially be improved in parental, guardian, and / or educational monitoring applications include family time, engagement, academic improvement, reduction in digital media technologies, screen time, online and console gaming, schedule adherence, nutritional fidelity, safety and / or substance abuse, exposure to seizures and / or location monitoring.

[0584] Since various changes can be made in the above-described systems, devices, and methods without departing from the scope of the present invention, all matters contained in the above description should be construed as illustrative and not limiting. Any numbers expressing quantities of ingredients, components, reaction conditions, and the like used herein should be construed to encompass the exact numerical values ​​specified herein and are modified in all cases by the term "about". Notwithstanding the numerical ranges and parameters described, the broad scope of the subject matter presented herein is approximate, and the numerical values ​​described are indicated as precisely as possible. However, any numerical values ​​may inherently contain certain errors or imprecision as evidenced by the standard deviation found in their respective measuring techniques. None of the features recited herein should be construed as invoking 35 U.S.C. § 112, paragraph 6, unless the term "means" is explicitly used.

[0585] As used herein, the terms "machine-readable medium", "computer-readable medium" and similar terms are generally used to refer to a non-transitory medium, volatile or non-volatile, that stores data and / or instructions that cause a machine to operate in a specific manner. Common forms of machine-readable media include, for example, hard disks, solid-state drives, magnetic tapes or any other magnetic data storage media, optical disks or any other optical data storage media, any physical media with a pattern of holes, RAM, PROM, EPROM, FLASH-EPROM, NVRAM, any other memory chip or cartridge, and network versions thereof.

[0586] These and various other forms of computer readable media may be involved in carrying one or more sequences of one or more instructions to a processing device for execution. Such instructions embodied on the media are generally referred to as "instructions" or "code." The instructions may be grouped in the form of a computer program or other grouping. When executed, such instructions may enable the processing device to perform the features or functions of the present application as described herein.

[0587] As used herein, a "processing device" may be implemented as a single processor that performs processing operations, or as a combination of dedicated and / or general-purpose processors that perform processing operations. A processing device may include a CPU, GPU, APU, DSP, FPGA, ASIC, SOC, and / or other processing circuitry.

[0588] Various embodiments described herein are described with reference to exemplary block diagrams, flow charts, and other diagrams. As will be apparent to one of ordinary skill in the art after reading this specification, the illustrated embodiments and their various alternatives can be implemented without being limited to the illustrated examples. For example, block diagrams and their accompanying description should not be construed as mandating a particular architecture or configuration.

[0589] Each of the processes, methods, and algorithms described in the preceding sections may be embodied in, and fully or partially automated by, instructions executed by one or more computer systems or computer processors, including computer hardware. The processes and algorithms may be implemented in part or in whole in application specific circuits. The various features and processes described above may be used independently of one another or may be combined in various ways. Different combinations and subcombinations are intended to fall within the scope of the present disclosure, and in some implementations, certain method or process blocks may be omitted. Furthermore, unless the context dictates otherwise, the methods and processes described herein are also not limited to any particular sequence, and the blocks or states associated therewith may be performed in other appropriate sequences, or in parallel, or in some other manner. Blocks or states may be added or deleted from the disclosed exemplary embodiments. The performance of a particular operation or process may be distributed among computer systems or computer processors that are not only present in a single machine, but also spread across multiple machines.

[0590] As used herein, the term "or" may be interpreted in an inclusive or exclusive sense. Additionally, descriptions of resources, acts, or structures in the singular should not be read to exclude the plural. In particular, conditional language such as "can," "could," "might," or "may" is generally intended to convey that certain embodiments include certain features, elements, and / or steps, while other embodiments do not, unless otherwise stated or understood otherwise within the context in which it is used.

[0591] Terms and phrases used herein, and variations thereof, unless expressly stated otherwise, should be construed as open ended rather than limiting. Modifiers such as "conventional," "traditional," "normal," "standard," "known," and terms of similar meaning should not be construed as limiting the described items to those available during a given period or at a given time, but should instead be read to embrace conventional, conventional, ordinary, or standard technology that may be available or known now or at any time in the future. In some cases, the presence of broadening words and phrases, such as "one or more," "at least," "but not limited to," or other similar phrases, should not be construed to imply that a narrower case is intended or required in the case where such broadening phrase is not present. [Prior art documents] [Non-patent literature]

[0592] [Non-Patent Document 1] Leekam,SR,Nieto,C.,Libby,SJ,Wing,L.,&Gould,J.(2007).Describing the Sensory Abnormalities of Children and Adults with Autism.Journal of Autism and Developmental Disorders,37(5),894-910 [Non-Patent Document 2] Maltz, S. (2019). Autistica Action Briefing: Employment-Harper G, Smith E, Heasman B, Remington A, Girdler S, Appleton VJ, Cameron C, Fell C [Non-Patent Document 3] Cusack, J., & Sterry, R. (December 2019). Autistica's top 10 research priorities [Non-Patent Document 4] Knapp, M., Romeo, R., & Beecham, J. (2009). Economic cost of autism in the UK. Autism, 13(3), 317-336; (London School of Economics, 2014). Autism is the UK's most expensive medical condition.

Claims

1. a wearable device comprising one or more sensors; one or more processors; one or more non-transitory computer readable media having executable instructions stored thereon; A system comprising: The executable instructions, when executed by the one or more processors, provide the system with: acquiring user sensory sensitivity data corresponding to a user input indicating whether a user of the wearable device is visually sensitive, ultrasonically sensitive, or interoceptively sensitive; using at least the user sensory sensitivity data to determine one or more sensory thresholds specific to the user and mediation data corresponding to one or more mediations specific to the user, the one or more sensory thresholds being selected from auditory, visual, or physiological sensory thresholds; storing the one or more sensory thresholds and the mediation data; recording a sensory input stimulus to the user using the one or more sensors; comparing the sensory input stimulus to the one or more sensory thresholds specific to a user; determining a mediation to be provided to the user based at least on the mediation data in response to comparing the sensory input stimulus to the one or more sensory thresholds, the mediation being configured to provide the user with relief from distraction, inattention, anxiety, fatigue, or sensory issues; providing the mediation to the user, the mediation including a warning mediation, a guidance mediation, or a filter mediation; A system that causes a system to perform an operation including:

2. the operations further include storing a first identifier indicative of whether the user is neurodiverse or neurotypical; determining mediation data corresponding to one or more sensory thresholds specific to the user and one or more mediations specific to the user includes determining the one or more sensory thresholds and the mediation data using at least the first identifier and the user sensory sensitivity data; The system of claim 1 .

3. The operations further include receiving user demographic data corresponding to a user input indicating an age, education level, or gender of the user; determining one or more sensory thresholds specific to the user and mediation data corresponding to one or more mediations specific to the user includes determining the one or more sensory thresholds and the mediation data using at least the first identifier, the user sensory sensitivity data, and the user demographic data; The system of claim 2.

4. The system of claim 2 , wherein the first identifier indicates whether the user has autism.

5. The system of claim 4 , wherein the first identifier indicates that the user has autism.

6. the mediation is configured to provide relief to the user from fatigue; the mediation includes the filter mediation; The filter mediation includes filtering in real time an audio signal presented to the user or a light signal presented to the user. The system of claim 5.

7. The system of claim 5 , wherein the mediation is configured to provide the user with relief from the distraction by increasing the user's response time to the distraction.

8. obtaining the sensory sensitivity data of the user includes receiving, in a graphical user interface, one or more first responses by the user to one or more first prompts indicating whether the user is visually sensitive, ultrasonically sensitive, or interoceptively sensitive; The operations include: receiving, in the graphical user interface, one or more second responses by the user to one or more second prompts indicative of an anxiety level of the user; deriving one or more sensory sensitivity scores based on the sensory sensitivity data, including a visual sensitivity score, a sound sensitivity score, or an interoceptive sensitivity score; deriving an anxiety score based on the one or more second responses; predicting that the user has autism based on at least the anxiety score and the one or more sensory sensitivity scores using a model that predicts the probability of autism based on at least the anxiety level and the one or more sensory sensitivity levels. The system of claim 5 , further comprising:

9. obtaining the user's sensory sensitivity data includes receiving, in a graphical user interface, one or more first responses by the user to one or more first prompts indicating whether the user is visually sensitive, ultrasonically sensitive, or interoceptively sensitive; The operations include: receiving, at the graphical user interface, one or more second responses by the user to one or more second prompts indicative of a fatigue level of the user; deriving one or more sensory sensitivity scores based on the sensory sensitivity data, including a visual sensitivity score, a sound sensitivity score, or an interoceptive sensitivity score; deriving a fatigue score based on the one or more second responses; predicting that the user has autism based on at least the fatigue score and the one or more sensory sensitivity scores using a model that predicts the probability of autism based on at least the fatigue level and the one or more sensory sensitivity levels. The system of claim 5 , further comprising:

10. 2. The system of claim 1, wherein obtaining the user sensory sensitivity data further comprises recording responses by the user to visual, acoustic, or physiological stimuli using at least the one or more sensors.

11. The system of claim 1 , wherein the mediation comprises a combination mediation of at least two mediations selected from the warning mediation, the guidance mediation, and the filter mediation.

12. The system of claim 11 , wherein the combination mediation includes the alert mediation followed by the filter mediation.

13. the alert mediation includes alerting the user about a visual or audio distraction; The filter mediation includes filtering in real time an audio or optical signal presented to the user, the audio or optical signal being associated with a distraction. The system of claim 12.

14. 2. The system of claim 1 , further comprising one or more fog nodes configured to locally store sensor data collected by the one or more sensors, the sensor data including first sensor data associated with the sensory input stimulus.

15. Storing the one or more sensory thresholds and the mediation data includes: using the one or more fog nodes to locally store the one or more sensory thresholds and the mediation data; comparing the sensory input stimulus to the one or more sensory thresholds includes comparing the sensory input stimulus to the one or more sensory thresholds using the one or more fog nodes. The system of claim 14.

16. 15. The system of claim 14, further comprising one or more edge nodes configured to communicatively couple to the one or more fog nodes and to a cloud server located remotely from the wearable device.

17. The one or more edge nodes: encrypting the first sensor data associated with the sensory input stimulus to obtain encrypted data; Sending the encrypted data to the cloud server; receiving a response from the cloud server; The system of claim 16 configured to:

18. 17. The system of claim 16, wherein the one or more fog nodes and the one or more edge nodes are present on a local area network (LAN) that includes the wearable device, an ad-hoc network that includes the wearable device, a LAN of a mobile device directly coupled to the wearable device, or the ad-hoc network of the mobile device.

19. the sensor data includes second sensor data that does not trigger mediation; The system is configured such that the second sensor data that does not trigger mediation is not made available to any cloud server located remotely from the wearable device.

17. The system of claim 16.

20. the mediation includes filtering the optical signal presented to the user in real time; the first sensor data associated with the sensory input stimulus includes first image data; the one or more edge nodes or the one or more fog nodes are configured to determine whether the first image data is sufficiently similar to second image data stored in the cloud server; determining the mediation to be provided to the user includes determining the filter mediation in response to determining that the first image data is sufficiently similar to the second image data.

17. The system of claim 16.

21. the mediation includes a filter mediation including filtering in real time an audio signal presented to the user; the first sensor data associated with the sensory input stimulus includes first audio data; the one or more edge nodes or the one or more fog nodes are configured to determine whether the first audio data is sufficiently similar to second audio data stored in the cloud server; determining the mediation to be provided to the user includes determining the filter mediation in response to determining that the first audio data is sufficiently similar to the second audio data.

17. The system of claim 16.

22. The operation includes: presenting to the user, in a graphical user interface, one or more access controls for controlling user data made available to one or more other users, the user data including sensor data collected by the one or more sensors, the one or more sensory thresholds, the mediation data, or a record of mediation presented to the user; receiving data corresponding to a user input selecting the one or more access controls; The system of claim 1 further comprising:

23. The operation includes: presenting to the user in a graphical user interface one or more access controls that allow or deny access to one or more other users to affect mediations presented to the user; receiving data corresponding to a user input that activates the one or more access controls; The system of claim 1 further comprising:

24. 2. The system of claim 1, wherein the operations further include presenting to the user in a graphical user interface a graphical summary of the user's progress through using the wearable device, the graphical summary including a running average of time between mediations.

25. the one or more sensors comprise a plurality of sensors of different types, the plurality of sensors comprising an auditory sensor, a galvanic skin sensor, a pupil sensor, a body temperature sensor, a head bobbing sensor, or an inertial motion unit; Recording a sensory input stimulus to the user includes obtaining first sensory data corresponding to a first sensory input stimulus from a first sensor of the plurality of sensors and second sensory data corresponding to a second sensory input stimulus from a second sensor of the plurality of sensors; determining the mediation to be provided to the user includes inputting at least the first sensory data and the second sensory data into a fusion-based deep learning (FBDL) model, the FBDL model outputting an identification of the mediation to be provided to the user; The system of claim 1 .

26. 26. The system of claim 25, wherein determining the mediation to be provided to the user includes inputting at least the first sensory data, the second sensory data, and the mediation data into the FBDL model, which outputs the identification information of the mediation to be provided to the user.

27. acquiring, in a wearable device system, user sensory sensitivity data corresponding to a user input indicating whether a user of a wearable device of the wearable device system is visually sensitive, ultrasonically sensitive, or interoceptively sensitive; determining, in the wearable device system, one or more sensory thresholds selected from an auditory, visual, or physiological sensory threshold using at least the user sensory sensitivity data, one or more sensory thresholds specific to the user, and mediation data corresponding to one or more mediations specific to the user; storing the one or more sensory thresholds and the mediation data in a storage device of the wearable device system; recording a sensory input stimulus to the user using one or more sensors of the wearable device system; comparing the sensory input stimulus to the one or more sensory thresholds specific to the user in the wearable device system; determining a mediation to be provided to the user based at least on the mediation data in response to comparing the sensory input stimulus with the one or more sensory thresholds, the mediation being configured to provide the user with relief from distraction, inattention, anxiety, fatigue, or sensory issues; providing the mediation to the user using at least the wearable device, the mediation comprising: alert mediation, guidance mediation, or filter mediation; The method includes:

28. The method further includes storing a first identifier indicative of whether the user is neurodiverse or neurotypical; determining mediation data corresponding to one or more sensory thresholds specific to the user and one or more mediations specific to the user includes determining the one or more sensory thresholds and the mediation data using at least the first identifier and the user sensory sensitivity data; 28. The method of claim 27.

29. The method further includes receiving user demographic data corresponding to a user input indicating an age, education level, or gender of the user; determining one or more sensory thresholds specific to the user and mediation data corresponding to one or more mediations specific to the user includes determining the one or more sensory thresholds and the mediation data using at least the first identifier, the user sensory sensitivity data, and the user demographic data; 29. The method of claim 28.

30. 30. The method of claim 28, wherein the first identifier indicates whether the user has autism.

Images