Systems and methods for assessing neurobehavioral traits

Abstract

A method for analyzing a user's behavioral, self-reported, and neural data to derive said user's neurobehavioral traits is disclosed. The method comprises: obtaining data associated with said user from a plurality of sources, wherein said data associated with said user comprises electroencephalography (EEG) data and said user's responses to one or more questionnaires; processing said data associated with said user to assess a range of functional brain measures or metrics by generating a model that is specific to said user's neurotype, wherein said neurotype comprises time-domain and time-frequency representations of said user's functional neural activity; and using said model to generate one or more predictive insights indicative of said user's neurobehavioral traits.

Description

SYSTEMS AND METHODS FOR ASSESSING NEUROBEHAVIORAL TRAITSCROSS-REFERENCE

[0001] This application claims the benefit of U.S. Provisional Application No. 63 / 554,405, filed February 16, 2024, and U.S. Provisional Application No. 63 / 735,098, filed December 17, 2024, which applications are incorporated herein by reference.BACKGROUND

[0002] The brain is a complex organ in the body made up of billions of neurons, which communicate both electrically and chemically. An individual’s psychological and cognitive state can be linked to the biological state of their brain; however, this linkage is not always readily apparent. For example, mental health treatment is typically based only on a user’s self-reported experiences and not on more biologically-based methods. For other medical treatments, healthcare providers may do basic mental health screenings through questionnaires only. Healthcare providers and individuals thus lack the complex portrait of brain function that biologically-based methods can add to a mental health assessment.SUMMARY

[0003] The present disclosure provides a method for analyzing a user’s behavioral, selfreported, and neural data to derive the user's neurobehavioral traits, comprising: obtaining data associated with the user from a plurality of sources, wherein the data associated with the user comprises electroencephalography (EEG) data and the user's responses to one or more questionnaires; processing the data associated with the user to assess a range of functional brain measures or metrics by generating a model that is specific to the user's neurotype, wherein the neurotype comprises time-domain and time-frequency representations of the user's functional neural activity; and using the model to generate one or more predictive insights indicative of the user's neurobehavioral traits. In some embodiments, the EEG data is collected using an EEG headset. In some embodiments, the EEG data is collected while the user is completing or performing a series of tasks on a device. In some embodiments, the EEG data comprises event-related potential (ERP) data. In some embodiments, the series of tasks comprises memory-related tasks, attention-related tasks, speed response tasks, gambling-style tasks, or viewing pictures. In some embodiments, the functional brain metrics span a range of domains. In some embodiments, the domains comprise attention, emotional reactivity, working memory, memory, stimulus categorization, decision-making speed, task engagement, reward sensitivity, emotional processing, emotional salience, error processing,or anticipation. In some embodiments, the method further comprises: comparing the user’s neurobehavioral traits with neurob ehavi oral traits of one or more other users, based at least in part on measured differences between the user's functional brain metrics and the one or more other users’ functional brain metrics. In some embodiments, the predictive insights further comprise a comparison of the user’s neurobehavioral traits with the neurobehavioral traits of the one or more other users. In some embodiments, the questionnaires are designed to elicit responses to a plurality of demographic and individual-specific variables. In some embodiments, the questionnaires comprise topics relating to anxiety, depression, stress exposure, sleep quality, or occupational activity. In some embodiments, the data further comprises data collected using one or more wearable devices. In some embodiments, the data collected using the one or more wearable devices comprises movement data, sleep data, heart rate data, blood oxygen data, or electrocardiogram (ECG) data. In some embodiments, the one or more wearable devices comprise a smartphone, a sleep monitoring device, a smart watch, a fitness tracker, smart glasses, smart jewelry, or smart clothing. In some embodiments, the questionnaires are dynamically generated on a predefined schedule. In some embodiments, the questionnaires are dynamically generated on a randomized schedule. In some embodiments, the EEG data is collected on a predefined schedule. In some embodiments, the EEG data is collected on a randomized schedule. In some embodiments, the EEG data is collected on a schedule that is based at least in part on a sleep quality of the user. In some embodiments, the questionnaires are dynamically generated on a schedule that is based at least in part on time of year, menstrual cycle schedule, time change, following important world events, on holidays, or after holidays. In some embodiments, the EEG data is collected on a schedule that is based at least in part on time of year, menstrual cycle schedule, time change, following important world events, on holidays, or after holidays. In some embodiments, the user's neurotype comprises associations between the user's self-reported measures, behavior, brain function, and one or more biomarkers of interest. In some embodiments, the method further comprises generating a time-series graphical representation of the functional brain metrics. In some embodiments, the data is analyzed using principal component analysis to reduce dimensionality within the questionnaire data, the EEG data, or the wearable data. In some embodiments, latent profile analysis is performed on factor scores from the principal component analysis. In some embodiments, the EEG data is filtered. In some embodiments, the EEG data is corrected for blinks. In some embodiments, the EEG data is scored in a time domain. In some embodiments, the EEG data is scored in a frequency domain. In some embodiments, results from the time domain and the frequency domain are averaged together. In some embodiments, the questionnaires are generated using generativeartificial intelligence (Al). In some embodiments, the method further comprises: using the model to recommend clinical interventions for the user based on the user’s neurotype or the predictive insights. In some embodiments, the method further comprises: using the model to recommend suggestions for the user to improve the user’s well-being. In some embodiments, the method further comprises: using the model to predict treatment responses for a health condition or disorder that the user is having or suspected to have. In some embodiments, the data is pre-processed and analyzed to create data quality metrics, wherein the data quality metrics are measured or based on signal and noise indices, internal reliability, or comparisons between odd and even halves of collected data. In some embodiments, the data quality metrics are compared to previous data collected to create internal norms. In some embodiments, the data quality metrics are used to determine when sufficient data has been collected, wherein the metric used to determine whether sufficient data has been collected is whether a difference between odd and even trials on a given neural metric falls within one standard deviation of the internal norms. In some embodiments, the data comprises time series data, structured data, or unstructured data. In some embodiments, the EEG data is collected in a single session. In some embodiments, the EEG data is collected over multiple sessions. In some embodiments, the ERP data is collected in a single session. In some embodiments, the ERP data is collected over multiple sessions. In some embodiments, the EEG data collected in a single session is used to create a neuropsychiatric profile for a subject. In some embodiments, the ERP data collected in a single session is used to create a neuropsychiatric profile for a subject. In some embodiments, the EEG data collected over multiple sessions is used to build a within-subject model for a subject. In some embodiments, the ERP data collected over multiple sessions is used to build a within-subject model for a subject.INCORPORATION BY REFERENCE

[0004] All publications, patents, and patent applications mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference.BRIEF DESCRIPTION OF THE DRAWINGS

[0005] The novel features of the invention are set forth with particularity in the appended claims. A better understanding of the features and advantages of the present invention will be obtained by reference to the following detailed description that sets forth illustrativeembodiments, in which the principles of the invention are utilized, and the accompanying drawings of which:

[0006] FIG. 1 shows a basic diagram of how data is incorporated to build a model of a user’s neurobehavioral traits in some embodiments of the present disclosure.

[0007] FIG. 2 shows some of the different functional brain metrics that may be measured in some embodiments of the present disclosure.

[0008] FIGs. 3A-3B displays views of the electroencephalography headset utilized. FIG. 3A displays a view of the electroencephalography headset. FIG. 3B displays a view of an individual wearing the electroencephalography headset shown in FIG. 3A.

[0009] FIGs. 4A-4D show an example of what an event-related potential graph looks like for a specific cognitive task trial. In FIG. 4A, response-locked ERPs at the frontocentral midline (FCz) for a single correct trial and an error trial from a single participant is shown. In FIG. 4B, response-locked ERPs from the same participant after 20 error trials and 20 correct trials were added to the ERP average shown in FIG. 4A. In FIG. 4C, ERPs from 20 participants for correct and error trials are presented. In FIG. 4D, the scalp distribution of the difference between error and correct trials in the highlighted time window (e.g. 0-100 ms) of the error- related negativity is shown.

[0010] FIG. 5 shows a computer system that may be used in this method.

[0011] FIGs. 6A-6B show diagrams of single-session vs. multi-session EEG measurements.

[0012] FIG. 7 shows a chart of drug development success rate by disease.

[0013] FIG. 8 shows a schematic of challenges for drug development in psychiatry.

[0014] FIG 9 shows a schematic of effects of a lack of clinically reliable and objective indicators in psychiatry.

[0015] FIG. 10 shows a schematic of the neurotyping approach for personalized psychiatry.

[0016] FIG. 11 shows a schematic of the event-related potential (ERP)-based biotyping method.

[0017] FIG. 12A shows an example ERP.

[0018] FIG. 12B shows different domains that may be characterized using ERPs.

[0019] FIG. 13 shows a schematic of the role ERP -based neurotyping could play in clinical trials.

[0020] FIG. 14 shows how neurotyping could be applied to a range of neurological diseases and disorders.

[0021] FIG. 15 shows the ERP from the picture-viewing paradigm, averaged across neutral and pleasant pictures, for healthy controls, patients with MDD who would remit, and patients with MDD who would not remit.DETAILED DESCRIPTION

[0022] The following description provides specific details for a comprehensive understanding of, and enabling description for, various embodiments of the technology. It is intended that the terminology used be interpreted in its broadest reasonable manner, even where it is being used in conjunction with a detailed description of certain embodiments.

[0023] Before describing the present teachings in detail, it is to be understood that the disclosure is not limited to specific compositions or process steps, and as such, may vary. As used herein, the singular forms “a,” “an,” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. Furthermore, to the extent that the terms “including,” “includes,” “having,” “has,” “with,” “such as,” or variants thereof, are used in either the specification and / or the claims, such terms are not limiting and are intended to be inclusive in a manner similar to the term “comprising.” Unless specifically noted, embodiments in the specification that recite “comprising” various components are also contemplated as “consisting of’ or “consisting essentially of’ the recited components.

[0024] Where values are described as ranges, it will be understood that such disclosure includes the disclosure of all possible sub-ranges within such ranges, as well as specific numerical values that fall within such ranges irrespective of whether a specific numerical value or specific sub-range is expressly stated.

[0025] Provided herein is a method for analyzing a user’s behavioral, self-reported, and neural data to derive said user's neurob ehavi oral traits. The present disclosure describes a method for combining EEG data with other types of data to provide insight about a user. This may allow healthcare providers to better personalize treatment for patients. It may also allow individuals to better understand their own traits for the purpose of improving their mental health and wellbeing.

[0026] In one aspect, the present disclosure provides a method for analyzing a user’s behavioral, self-reported, and neural data to derive the user’s neurob ehavi oral traits, comprising: (a) obtaining data associated with the user from a plurality of sources, wherein the data associated with the user comprises electroencephalography (EEG) data and the user’s responses to one or more questionnaires; (b) processing the data associated with the user to assess a range of functional brain measures or metrics by generating a model that is specific to the user’s neurotype, wherein the neurotype comprises time-domain and time-frequency representations of the user’s functional neural activity; and (c) using the model to generate one or more predictive insights indicative of the user’s neurob ehavi oral traits.Electroencephalography

[0027] In one aspect, the present disclosure provides a method that utilizes electroencephalography (EEG), a method that can be used to record the spontaneous electrical activity of the brain. EEG is frequently used to diagnose epilepsy, a disease where a patient suffers from repeated seizures. EEG can also be useful for studying other disorders that affect brain electrical activity. An EEG can be conducted by placing one or more electrodes on the scalp, wherein the one or more electrodes are coupled to a controller or other output device using conductive wires. The one or more electrodes may comprise an electrode body that is conductive at least partially. The one or more electrodes may comprise one or more members extending from the electrode body wherein each of the one or more members defines a lumen and a distal opening. The one or more electrodes may comprise a reservoir with a compressible structure that contains a conductive fluid or gel in fluid communication with the one or more members. The one or more electrodes may comprise a backing that supports the electrode body and reservoir. Each electrode may be connected to one input of a differential amplifier, with one amplifier per electrode pair. In some embodiments, the electrodes may have electrical communication with a controller and / or output device that is configured to receive electrical signals from the one or more electrodes. This controller and / or output device may record and / or output a response. In some cases, a common system reference electrode may be connected to the other input of each differential amplifier. These amplifiers may amplify the voltage between the active electrode and the reference electrode. In digital EEG, the amplified signal may be digitized via an analog-to- digital converter, after being passed through an anti-aliasing filter.

[0028] In some cases, different activation procedures may be used in an EEG to stimulate the subject. These activation procedures may include hyperventilation, photic stimulation (e.g. with a strobe light), eye closure, mental activity, sleep, and sleep deprivation.

[0029] In some cases, the electrodes may be coupled wirelessly to the controller or other output device. The electrodes may be contained within an electrode carrier that can wrap around the head of a subject. In some embodiments, the electrode carrier may be configured as a headband or headset. When the electrode carrier is configured as a headband or headset, the electrodes may be spaced apart from each other and aligned on the subject’s head for optimal reception of EEG signals. In some embodiments, the EEG data may be collected using an EEG headset.

[0030] The controller and / or output device may comprise any number of devices for receiving electrical signals. Such devices may include electrophysiological monitoring devices.

[0031] In some cases, an EEG device may have active electrodes. In other cases, an EEG device may have passive electrodes. In some cases, an EEG device may have wet electrodes. In some cases, an EEG device may have dry electrodes. In some cases, an EEG device may have passive, dry electrodes. In one embodiment, the EEG device may be configured as a headset, and the headset may have passive, dry electrodes.

[0032] In some cases, the EEG electrodes used in an EEG device may include, but are not limited to, Fpl, Fp2, F7, F3, FZ, F4, F8, FT9, FT10, FC5, FC1, FC2, FC6, T7, C3, CZ, C4, T8, CP5, CPI, CP2, CP6, P7, P3, PZ, P4, P8, TP9, TP10, 01, OZ, and 02. In one embodiment, the EEG electrodes (or channels) used in the EEG device may be Fpl, Fp2, F3, F4, FZ, CZ, PZ, and a ground or reference electrode.

[0033] In some cases, the measurement range of the EEG device may be about ±20 mV, about ±40 mV, about ±60 mV, about ±80 mV, about ±100 mV, about ±200 mV, about ±300 mV, about ±500 mV, about ±700 mV, about ± 1 V, about ±1.125 V, about ±1.25 V, about ±1.5 V, about ±2 V, about ±2.5 V, or any range therein. In one embodiment, the measurement range of the EEG device may be about ±100 mV.

[0034] In some cases, the input-referred noise of the EEG device may be less than about 1 pVpp. In one embodiment, the input-referred noise may be about 0.8 pVpp, e.g. about 0.01 Hz to about 65 Hz at about a 250 Hz sample rate.

[0035] In some embodiments, the input impedance of the EEG device may be greater than or equal to about 100 MOhm. In some embodiments, the input impedance of the EEG device may be greater than about 200 MOhm. In some embodiments, the input impedance of the EEG device may be greater than about 300 MOhm. In some embodiments, the input impedance of the EEG device may be greater than about 400 MOhm. In some embodiments, the input impedance of the EEG device may be greater than about 500 MOhm. In some embodiments, the input impedance of the EEG device may be greater than about 600 MOhm. In some embodiments, the input impedance of the EEG device may be greater than about 700 MOhm. In some embodiments, the input impedance of the EEG device may be greater than about 800 MOhm. In some embodiments, the input impedance of the EEG device may be greater than about 900 MOhm. In some embodiments, the input impedance of the EEG device may be greater than about 1000 MOhm. In some embodiments, the input impedance of the EEG device may be greater than about 1100 MOhm. In some embodiments, the input impedance of the EEG device may be greater than about 1200 MOhm. In some embodiments, the input impedance of the EEG device may be greater than about 1300 MOhm. In some embodiments, the input impedance of the EEG device may be greater than about 1400MOhm. In some embodiments, the input impedance of the EEG device may be greater than about 1500 MOhm.

[0036] In some embodiments, the common mode rejection ratio of the EEG device may be greater than about 80dB at 50 / 60 Hz. In some embodiments, the common mode rejection ratio of the EEG device may be greater than about 90dB at 50 / 60 Hz. In some embodiments, the common mode rejection ratio of the EEG device may be greater than about lOOdB at 50 / 60 Hz. In some embodiments, the common mode rejection ratio of the EEG device may be greater than about 1 lOdB at 50 / 60 Hz. In some embodiments, the common mode rejection ratio of the EEG device may be greater than about 120dB at 50 / 60 Hz.

[0037] In some embodiments, the isolation mode rejection ratio of the EEG device may be greater than about 20 dB. In some embodiments, the isolation mode rejection ratio of the EEG device may be greater than about 40 dB. In some embodiments, the isolation mode rejection ratio of the EEG device may be greater than about 60 dB. In some embodiments, the isolation mode rejection ratio of the EEG device may be greater than or equal to about 80 dB. In some embodiments, the isolation mode rejection ratio of the EEG device may be greater than or equal to about 100 dB. In some embodiments, the isolation mode rejection ratio of the EEG device may be greater than or equal to about 120 dB. In some embodiments, the isolation mode rejection ratio of the EEG device may be greater than or equal to about 140 dB. In some embodiments, the isolation mode rejection ratio of the EEG device may be greater than or equal to about 160 dB. In some embodiments, the isolation mode rejection ratio of the EEG device may be greater than or equal to about 180 dB.

[0038] In some embodiments, the A / D conversion of the EEG channel in the EEG device may be about 8 bits. In some embodiments, the A / D conversion of the EEG channel in the EEG device may be about 10 bits. In some embodiments, the A / D conversion of the EEG channel in the EEG device may be about 12 bits. In some embodiments, the A / D conversion of the EEG channel in the EEG device may be about 14 bits. In some embodiments, the A / D conversion of the EEG channel in the EEG device may be about 16 bits. In some embodiments, the A / D conversion of the EEG channel in the EEG device may be about 18 bits. In some embodiments, the A / D conversion of the EEG channel in the EEG device may be about 20 bits. In some embodiments, the A / D conversion of the EEG channel in the EEG device may be about 22 bits. In some embodiments, the A / D conversion of the EEG channel in the EEG device may be about 24 bits. In some embodiments, the A / D conversion of the EEG channel in the EEG device may be about 26 bits. In some embodiments, the A / D conversion of the EEG channel in the EEG device may be about 28 bits. In some embodiments, the A / D conversion of the EEG channel in the EEG device may be about 30bits. In some embodiments, the A / D conversion of the EEG channel in the EEG device may be about 32 bits.

[0039] In some embodiments, the hardware sample rate of the EEG device may be about 250Hz. In some embodiments, the hardware sample rate of the EEG device may be about 500Hz. In some embodiments, the hardware sample rate of the EEG device may be about 750Hz. In some embodiments, the hardware sample rate of the EEG device may be about 100Hz.Electroencephalography Data Analysis

[0040] In one aspect, the present disclosure utilizes data from EEG to build a model of the user’s neurob ehavi oral traits. EEG may be used in the framework of neuroscience, cognitive science, cognitive psychology, psychiatry, neurolinguistics, and psychophysiology. EEG can have the advantage over some other methods of studying brain function in that the equipment needed for it is relatively small and portable. It also is relatively safe and generally does not involve exposure to high-intensity magnetic fields or radiation. It also is generally relatively non-invasive.

[0041] EEG reports generally comprise graphs of the electrical signal at each electrode over time. Such electrical signals may have a wave-like structure. An EEG may be described in terms of rhythmic activity and transients. The rhythmic activity may be divided by frequency into bands. These frequency bands may have different nomenclatures such as “alpha”, “beta”, “theta”, “delta”, “gamma”, and “mu”. These frequency bands can have a certain distribution over the scalp or a certain biological significance.Event-Related Potentials

[0042] In one aspect, the present disclosure provides a method for analyzing data from a particular type of EEG measurement known as an event-related potential (ERP). In some cases, a specific event may cause the brain’s neurons to become active at the same time, creating a summed electrical signal that propagates throughout the brain. Because an EEG reflects all brain activity simultaneously, the brain’s response to a single stimulus or event of interest may not be visible in a single trial of an EEG recording. In order to see the brain’s response to a specific stimulus, the experimenter may conduct many trials and then average them together, causing random brain activity to be averaged out while only the stimulus- related waveform, e.g. the ERP, remains. The ERP signal has a signal-to-noise ratio, wherein the random background brain activity is the noise contribution that obscures the signal of interest. Averaging many trials together can increase the signal-to-noise ratio of the ERPs, making them more discernible and allowing them to be interpreted.

[0043] ERPs can have several advantages, such as being evident in a single participant through these averages. This may make them robust and reproducible. ERPs can also have strong temporal resolution, due to the electrical activity being sampled very frequently, e.g. every millisecond.

[0044] An error-related negativity, or ERN, can make up one component of an ERP. In clinical psychology, an ERN may be a distinct neural response to making a mistake. Different individuals can have different sizes of ERNs, and this variability can be clinically meaningful. For example, the ERN may be increased among subjects diagnosed with obsessive-compulsive disorder (OCD), generalized anxiety disorder (GAD), or social anxiety disorder (SAD). The impact of major depressive disorder (MDD) on the ERN can be unclear; in some cases, GAD subjects with comorbid MDD may not have an ERN different from healthy controls, whereas GAD subjects without comorbid MDD may have an increased ERN. This may suggest that MDD and GAD are distinct neural processes with different effects on the brain.Tasks Performed During EEG

[0045] In some embodiments, the EEG is performed while a user is performing a specific task. In some embodiments, the EEG is performed while a user is completing or performing a series of tasks on an EEG device. Such tasks may be memory-related tasks, attention-related tasks, speed response tasks, gambling-style tasks, or viewing pictures. For example, a memory-related task that may be used is the N-back task, which is a continuous performance task where subjects are presented with a series of stimuli and have to indicate whether the current stimulus is presented N steps before or not. Increasing or decreasing N, the number of steps, may allow the EEG technician to adjust the difficulty of the task. Increases in mental workload may result in an increase in the frontal part of the brain’s theta wave activity. Increases in mental workload may result in a decrease in the parietal part of the brain’s alpha wave activity.

[0046] In some cases, users may complete a visual go / nogo computer task. A go / nogo task is an experimental paradigm where a user is required to press a button when they see a “go” signal, and to not respond when they see a “no go” signal. A key behavior measured by this task is the user’s ability to withhold a response when they see a “no go” signal. For example, a user may be presented with a stream of continuous stimuli and asked to use a mouse click to blast aliens and asteroids, but not astronauts, wherein the three different targets have different frequencies with which they appear.

[0047] Go-nogo tasks can be used to measure the P300 component of an ERP. The P300 component of an ERP is a component elicited during the decision-making process. When it isrecorded by EEG, the P300 may appear as a positive deflection in voltage. The P300 may have a latency (e.g. delay between stimulus and response) of about 250 ms to about 500 ms. Increased P300 latency may be associated with Alzheimer’s disease.

[0048] In some cases, participants may complete an affective picture task. In some embodiments, the EEG technician may show the participant an affective photograph from an established database of known affective photographs (e.g. the International Affective Picture System). Such affective photographs can be standardized. Such affective photographs can be emotionally evocative. All stimuli of the database may be rated along the dimensions of valence and arousal, as described in the two-dimensional circumplex model of emotion. The valence dimension reflects the pleasantness of a situation and ranges from sadness to happiness. The arousal dimension reflects the responsiveness of the participant and ranges from sleep to frenzied excitement. Affective states may be measured using the late positive potential (LPP) and / or frontal alpha asymmetry (FAA). The LPP is an EEG feature seen in the time domain, which represents a positive deflection in the ERP curve. The LPP may reflect activity related to the arousal dimension. The FAA represents contributions from the brain’s individual hemispheres. The FAA may be related to the valence dimension.

[0049] In some embodiments, a gambling task such as the Iowa Gambling Task may be performed. In the IGT, a participant may be presented with four decks of cards, with each card containing a reward and a punishment. The participant may be asked to pick cards from these decks to earn as much money as possible. Two of said decks may contain a large win each time and sometimes an even larger loss; in the long term, the participant would lose money if these decks continued to be selected (the disadvantageous decks). The other two of said decks may contain a small win each time and sometimes a small loss; in the long term, the participant would earn money if these decks continued to be selected (the advantageous decks).

[0050] In some embodiments, a speed response task may be performed. Speeded response tasks may be performed on a computer in order to elicit error-related brain activity. In some cases, the speeded response task is a flankers task. For example, in an arrowhead version of a flankers task, a participate can respond to the direction of a central arrowhead flanked by either compatible (e.g. “< < < < <” or “> > > > >”) or incompatible (e.g. “< < > < <” or “ > > < > >”) arrowheads. Stimuli can be presented briefly (e.g. for 200 ms). Participants may perform between about 300 and about 400 trials over about 10 minutes.Neurob ehavi oral Traits

[0051] In some embodiments, the method comprises generating functional brain metrics in a range of domains. In some embodiments, such domains comprise attention, error processing,emotional reactivity, working memory, memory, stimulus categorization, decision-making speed, task engagement, reward sensitivity, emotional processing, emotional salience, or anticipation.

[0052] In FIG. 2, a diagram of some of the different possible functional brain metrics is shown. These functional brain metrics may comprise attention, anticipation, stimulus categorization, working memory, emotional reactivity, and error processing.

[0053] In some cases, different components of an ERP may be analyzed to determine functional brain metrics. In some embodiments, the P300 as previously described may be used to generate functional brain metrics related to decision-making speed, task engagement, or categorization abilities.

[0054] Another component of an ERP that may be used to generate functional brain metrics is Reward Positivity (RewP). The RewP is a frequently used EEG measure of neural response to rewards. The RewP is commonly measured at frontocentral sites of the brain. It is often thought to be maximal at what is called the FCz electrode, which is generally placed the midline of the brain between the frontal and central parts of the brain. In some embodiments, the RewP may be used to generate functional brain metrics related to reward sensitivity.

[0055] Another component of an ERP that may be used to generate functional brain metrics is the LPP as previously described. The LPP generally reflects facilitated attention to emotional stimuli. The magnitude of the LPP is often greater when an individual views emotionally arousing pictures, as compared to neutral pictures. In some embodiments, the LPP may be used to generate functional brain metrics related to emotional processing or emotional salience.

[0056] Another component of an ERP that may be used to generate functional brain metrics is the ERN as previously described. In some embodiments, the ERN may be used to generate functional brain metrics related to error processing.

[0057] Another component of an ERP that may be used to generate functional brain metrics is the stimulus-preceding negativity (SPN). The SPN is a slow cortical potential that is often seen in anticipation of motivational stimuli, such as aversive stimuli like an electric shock. In some embodiments, the SPN may be used to generate functional brain metrics related to anticipation.

[0058] Other functional brain metrics may be generated based on other components of ERP or EEG data not mentioned here.Machine Learning

[0059] As used in this specification and the appended claims, the terms “artificial intelligence,” “artificial intelligence techniques,” “artificial intelligence operation,” and“artificial intelligence algorithm” generally refer to any system or computational procedure that may take one or more actions that simulate human intelligence processes for enhancing or maximizing a chance of achieving a goal. The term “artificial intelligence” may include “generative modeling,” “machine learning” (ML), or “reinforcement learning” (RL).

[0060] As used in this specification and the appended claims, the terms “machine learning,” “machine learning techniques,” “machine learning operation,” and “machine learning model” generally refer to any system or analytical or statistical procedure that may progressively improve computer performance of a task. In some cases, ML may generally involve identifying and recognizing patterns in existing data in order to facilitate making predictions for subsequent data. ML may include a ML model (which may include, for example, a ML algorithm). Machine learning, whether analytical or statistical in nature, may provide deductive or abductive inference based on real or simulated data. The ML model may be a trained model. ML techniques may comprise one or more supervised, semi-supervised, selfsupervised, or unsupervised ML techniques. For example, an ML model may be a trained model that is trained through supervised learning (e.g., various parameters are determined as weights or scaling factors). ML may comprise one or more of regression analysis, regularization, classification, dimensionality reduction, ensemble learning, meta learning, association rule learning, cluster analysis, anomaly detection, deep learning, or ultra-deep learning. ML may comprise, but is not limited to: k-means, k-means clustering, k-nearest neighbors, learning vector quantization, linear regression, non-linear regression, least squares regression, partial least squares regression, logistic regression, stepwise regression, multivariate adaptive regression splines, ridge regression, principal component regression, least absolute shrinkage and selection operation (LASSO), least angle regression, canonical correlation analysis, factor analysis, independent component analysis, linear discriminant analysis, multidimensional scaling, non-negative matrix factorization, principal components analysis, principal coordinates analysis, projection pursuit, Sammon mapping, t-distributed stochastic neighbor embedding, AdaBoosting, boosting, gradient boosting, bootstrap aggregation, ensemble averaging, decision trees, conditional decision trees, boosted decision trees, gradient boosted decision trees, random forests, stacked generalization, Bayesian networks, Bayesian belief networks, naive Bayes, Gaussian naive Bayes, multinomial naive Bayes, hidden Markov models, hierarchical hidden Markov models, support vector machines, encoders, decoders, auto-encoders, stacked auto-encoders, perceptrons, multi-layer perceptrons, artificial neural networks, feedforward neural networks, convolutional neural networks, recurrent neural networks, long short-term memory, deep belief networks, deepBoltzmann machines, deep convolutional neural networks, deep recurrent neural networks, or generative adversarial networks.

[0061] Training the ML model may include, in some cases, selecting one or more untrained data models to train using a training data set. The selected untrained data models may include any type of untrained ML models for supervised, semi-supervised, self-supervised, or unsupervised machine learning. The selected untrained data models may be specified based upon input (e.g., user input) specifying relevant parameters to use as predicted variables or other variables to use as potential explanatory variables. For example, the selected untrained data models may be specified to generate an output (e.g., a prediction) based upon the input. Conditions for training the ML model from the selected untrained data models may likewise be selected, such as limits on the ML model complexity or limits on the ML model refinement past a certain point. The ML model may be trained (e.g., via a computer system such as a server) using the training data set. In some cases, a first subset of the training data set may be selected to train the ML model. The selected untrained data models may then be trained on the first subset of training data set using appropriate ML techniques, based upon the type of ML model selected and any conditions specified for training the ML model. In some cases, due to the processing power requirements of training the ML model, the selected untrained data models may be trained using additional computing resources (e.g., cloud computing resources). Such training may continue, in some cases, until at least one aspect of the ML model is validated and meets selection criteria to be used as a predictive model.

[0062] In some cases, one or more aspects of the ML model may be validated using a second subset of the training data set (e.g., distinct from the first subset of the training data set) to determine accuracy and robustness of the ML model. Such validation may include applying the ML model to the second subset of the training data set to make predictions derived from the second subset of the training data. The ML model may then be evaluated to determine whether performance is sufficient based upon the derived predictions. The sufficiency criteria applied to the ML model may vary depending upon the size of the training data set available for training, the performance of previous iterations of trained models, or user- specified performance requirements. If the ML model does not achieve sufficient performance, additional training may be performed. Additional training may include refinement of the ML model or retraining on a different first subset of the training dataset, after which the new ML model may again be validated and assessed. When the ML model has achieved sufficient performance, in some cases, the ML may be stored for present or future use. The ML model may be stored as sets of parameter values or weights for analysis of further input (e.g., further relevant parameters to use as further predicted variables, furtherexplanatory variables, further user interaction data, etc.), which may also include analysis logic or indications of model validity in some instances. In some cases, a plurality of ML models may be stored for generating predictions under different sets of input data conditions. In some embodiments, the ML model may be stored in a database (e.g., associated with a server).Questionnaires

[0063] In one aspect, the present disclosure provides a method for incorporating data from self-reported questionnaires about the user into the model of the user’s neurob ehavi oral traits. In some cases, a subject’s mental health can be monitored using self-report questionnaires. For example, the Patient Health Questionnaire 9 (PHQ-9) or the Generalized Anxiety Disorder 7 (GAD-7) may be used. Other questionnaires may use free response data about a person’s psychological state rather than the discrete numerical measures used in the PHQ-9 or the GAD-7. Such questionnaires may be incorporated into the model. In some embodiments, said questionnaires may be designed to elicit responses to a plurality of demographic and individual-specific variables.

[0064] In some embodiments, other questionnaires assessing a subject’s psychological state may be used, such as questionnaires assessing sleep quality, stress exposure, or occupational activity. These questionnaires may comprise a plurality of demographic and individualspecific variables.

[0065] In some embodiments, the questionnaires may be generated using generative artificial intelligence (Al).Model Building from Questionnaire Data

[0066] In one aspect, the present disclosure provides a method for utilizing machine learning to generate predictive insights about a user’s neurob ehavi oral traits. In the case of the questionnaire data described herein, models may utilize different methods. For example, with questionnaires that use categorical data, such as the PHQ-9 or GAD-7, it may be relatively straightforward to encode these categories numerically to be used by a machine learning model. In some embodiments, a questionnaire that uses open-ended free responses about the user’s mental health may be analyzed using natural language processing (NLP). The meanings of words in the open-ended free responses may be represented as vectors, with words with similar meanings clustering closer to each other in vector space.

[0067] In some embodiments, data used in the method may further comprise data collected using one or more wearable devices. In some embodiments, data from a wearable device, such as a smartphone, a sleep monitoring device, a smart watch, a fitness tracker, smart glasses, smart jewelry, or smart clothing, may be incorporated into the model. Such wearabledevices may comprise an accelerometer, a gyroscope, a heart rate monitor, a blood pressure monitor, a GPS sensor, a barometer, or an electrocardiogram (ECG) device. In some embodiments, the data collected using the one or more wearable devices comprises movement data, sleep data, heart rate data, blood oxygen data, or electrocardiogram (ECG) data. Metrics from such wearable devices may comprise distance traveled, steps taken, calories burned, sleep duration, sleep quality, heart rate, blood pressure, time spent active, time spent sitting or resting, skin perspiration, oxygen saturation level, ECG graph, or other metrics. In some embodiments, numerical features may be extracted from the data from such wearable devices. These features may also be incorporated into the model describe herein. Model Building from EEG Data

[0068] Machine learning may also be performed on EEG data to derive predictive insights. In some cases, EEG signal can be divided into one or more temporal segments, wherein each temporal segment corresponds to a time epoch, wherein a time epoch comprises a start time and a duration. Features from the temporal segments may be extracted. In some cases, multichannel features may be extracted that quantify the correlation between different pairs of temporal segments from different EEG signals in the same time epoch. Time epochs may vary in duration. The duration of the time epochs may be about 10 seconds, about 20 seconds, about 30 seconds, about 40 seconds, about 50 seconds, about 60 seconds, about 2 minutes, about 5 minutes, or about 10 minutes. Successive time epochs may or may not overlap. Overlapping time epochs may overlap by 50% or less.

[0069] In some cases, the features extracted from the EEG data may comprise one or more time-domain features, one or more frequency domain features, or features that correlate different time-based segments from another EEG signal that is simultaneously collected. Simultaneously-collected EEG signals may be collected from a same hemisphere of the brain or from different hemispheres of the brain.

[0070] In some cases, machine learning can be performed on ERP data.

[0071] In some embodiments, the EEG data may be pre-processed. In some embodiments, the EEG data may be filtered. In some embodiments, the EEG data may be corrected for blinks. In some embodiments, the EEG data may be scored in the time domain. In some embodiments, the EEG data may be scored in the frequency domain. In some embodiments, the results from the EEG data scored in the time domain may be averaged together. In some embodiments, the results from the EEG data scored in the frequency domain may be averaged together.Building a Model

[0072] In one aspect, the method provided herein combines data from EEG and self-reported questionnaires to create a model specific to a user’s “neurotype”. The neurotype comprises one or more associations between the user's self-reported measures, behavior, brain function, and one or more biomarkers of interest. In one aspect, the method comprises processing the data to assess a range of functional brain measures or metrics by generating a model that is specific to the user’s neurotype, wherein the neurotype comprises time-domain and timefrequency representations of the user’s functional neural activity. In some embodiments, the method provided herein may use time series data, structured data, or unstructured data.

[0073] In some embodiments, the method further comprises data from wearable devices.

[0074] In some embodiments, the data is pre-processed and analyzed to create data quality metrics. In some cases, the data quality metrics may be measured or based on signal and noise indices. In some cases, the data quality metrics may be measured or based on internal reliability. In some bases, the data quality metrics may be measured or based on comparisons between odd and even halves of collected data. In some cases, the data quality metrics are compared to previous data collected to create internal norms. The data quality metrics may also be used to determine when sufficient data has been collected. In some cases, the metric used to determine whether sufficient data has been collected is whether the difference between odd and even trials on a given neural metric fall within one standard deviation of the internal norms.

[0075] In some embodiments, the model uses dimensionality reduction methods to reduce the dimensionality of the EEG data, questionnaire data, or wearable device data. In some embodiments, said dimensionality reduction methods may comprise principal component analysis (PCA), random forest, independent component analysis (ICA), t-distributed stochastic neighbor embedding (t-SNE), or uniform manifold approximation and projection (UMAP). Principal component analysis may be used to reduce dimensionality within the questionnaire data, the EEG data, or the wearable data.

[0076] In some embodiments, the model may comprise performing latent class analysis on data from said PCA. Latent class analysis is a probabilistic modeling algorithm that allows clustering of data and statistical inference. Said latent class analysis may utilize PCA factor scores.

[0077] In some embodiments, the model input may change with additional user data. In some cases, the questionnaires may be dynamically generated on a predetermined schedule. In some cases, the questionnaires may be dynamically generated on a random schedule. In some cases, the questionnaire data may be collected on a schedule that is based at least in part on asleep quality of the user. In some cases, the questionnaires may be dynamically generated on a schedule that is based at least in part on time of year, menstrual cycle schedule, time change, following important world events, on holidays, or after holidays.

[0078] In some cases, the EEG data may be collected on a predefined schedule. In some cases, the EEG data may be collected on a randomized schedule. In some cases, the EEG data may be collected on a schedule that is based at least in part on a sleep quality of the user. In some cases, the EEG data may be collected on a schedule that is based at least in part on time of year, menstrual cycle schedule, time change, following important world events, on holidays, or after holidays.

[0079] In some cases, the wearable device data may be collected on a predefined schedule. In some cases, the wearable device data may be collected on a randomized schedule. In some cases, the wearable device data may be collected on a schedule that is based at least in part on a sleep quality of the user. In some cases, the wearable device data may be collected on a schedule that is based at least in part on time of year, menstrual cycle schedule, time change, following important world events, on holidays, or after holidays.

[0080] In some embodiments, the EEG data may be collected in a single session. In some embodiments, the ERP data may be collected in a single session. In some embodiments, the EEG data collected in a single session may be used to create a neuropsychiatric profile for a subject. In some embodiments, the ERP data collected in a single session may be used to create a neuropsychiatric profile for a subject. Such single-session profiles may be useful in clinical trials.

[0081] In some embodiments, the EEG data may be collected over multiple sessions. In some embodiments, the ERP data may be collected over multiple sessions. In some embodiments, the EEG data collected over multiple sessions may be used to build a within-subject model for a subject. In some embodiments, the ERP data collected over multiple sessions may be used to build a within-subject model for a subject. Such within-subject models may be useful in generating deeper insights about a subject.

[0082] In FIG. 6A, a diagram of single session is shown. In a single session setup, data from ERPs is translated into brain function measures, as shown on the left. These brain function measures may be combined with data from wearables and questionnaires. Brain function measures alone can biotype the subjects. Providers can use data from wearables, questionnaires, and ERPs, but it is not necessary to use all three at once. Specific ERPs may be able to suggest the best treatment options.

[0083] In FIG. 6B, a diagram of a multi-session setup is shown. The basic setup may be similar to a single-day setup, but is performed over multiple days. Performing themeasurements over multiple days may allow a provider to predict mental health problems. Measurements over multiple days may also be able to suggest the best treatment options. This can be a more precise model for individuals.Comparisons Between Users

[0084] In some embodiments, the method may comprise gathering data from at least one user. In some cases, the method comprises gathering data from about 100 users, about 200 users, about 300 users, about 400 users, about 500 users, about 600 users, about 700 users, about 800 users, about 900 users, about 1,000 users, about 2,000 users, about 5,000 users, about 10,000 users, or more. In some embodiments, the method further comprises comparing the data and / or neurotype model of one user with the data and / or neurotype model of another user. In some cases, the method comprises comparing the neurob ehavi oral traits of one user with the neurob ehavi oral traits of one or more other users, based at least in part on measured differences between the users’ functional brain metrics.Drug Development

[0085] In some cases, the methods described herein may be used in developing psychiatric drugs. FIG. 7 shows a chart of drug development success rate by disease, from Phase I to approved (Office of Pharmaceutical Industry Research). Drug development success rate may vary by disease. Surgery, for example, may have a success rate close to 30%, while hepatic and biliary diseases may have a success rate close to 7%. The average success rate for all diseases is 13%. Psychiatry, however, has a success rate of only 6%. As shown in FIG. 8, drug development in psychiatry may be difficult due to patient heterogeneity, a lack of biomarkers and / or no treatment outcome predictors. These aspects may lead to a lack of clinically reliable and objective indicators of disease subtypes and response prognosis.

[0086] A lack of clinically reliable and objective indicators of disease subtypes and response prognosis in psychiatry may have broad implications for pharmaceutical companies, healthcare providers, and patients / subjects. This is shown in FIG. 9. For example, pharmaceutical companies may be impacted by patient selection criteria in trial recruitment, a high failure rate of drug trials, or a low rate of economic value data for payers. For instance, two patients with depression may have quite different symptoms. The goal of the technology described herein is to reduce inter-patient variability through neurotyping. A healthcare provider, such as a behavioral health specialist, may be impacted by a lack of objective data in assessing a proper treatment plan for a subject. A trial and error approach may also cause a subject to suffer. Finally, subjects may be impacted by an extended time necessary to find a response to drug therapy, a failure of up to 3 drugs to treat depression, a loss of work, or high annual out of pocket costs in finding a treatment that works.

[0087] The present disclosure provides methods of neurotyping, which may change patient outcomes in the future. Neurotyping may be a form of precision psychiatry. A schematic of this approach is shown in FIG. 10. A personalized psychiatry approach may integrate symptoms, circuits, physiology, cognition, labs, genetics, and / or life experience to determine subjects’ clinical biotypes for a variety of different psychiatric disorders. Knowledge of these clinical biotypes can lead to personalized treatment approaches.

[0088] The present disclosure provides a configurable platform that enables ERP -based biotyping. This biotyping platform can neurotype subjects to inform actions that can predict outcomes. ERP -based biotyping may be used to understand disease subtypes and predict treatment response. A schematic of this approach is shown in FIG. 11. An EEG headset may be used to measure functional neural activity in a subject. This EEG headset may be a high- quality, accessible EEG wearable. The EEG headset may allow for active EEG data capture. An EEG headset may be used to measure one or more ERPs. The ERPs may be determined while the user is performing a series of tasks. These measurements may be used to determine a subject’s neurob ehavi oral traits, such as effort, attention, negative bias, emotion, anticipation, and memory. The ERP data may be used to create a biotype or neurotype. These biotypes or neurotypes may be used in algorithm-driven software that can use predictive analysis for intervention or decision making.

[0089] In FIGs. 3A-3B, views of the EEG headset used in this method are shown. FIG. 3A displays a view of the EEG headset used. FIG. 3B displays a view of an individual wearing this headset.

[0090] In FIG. 4, ERPs from an arrows version of a flankers task as described herein are shown. In FIG. 4A, response-locked ERPs at the frontocentral midline (FCz) for a single correct trial and an error trial from a single participant is shown. In FIG. 4B, response-locked ERPs from the same participant after 20 error trials and 20 correct trials were added to the ERP average shown in FIG. 4A. In FIG. 4C, ERPs from 20 participants for correct and error trials are presented. In FIG. 4D, the scalp distribution of the difference between error and correct trials in the highlighted time window (e.g., 0-100 ms) of the error-related negativity is shown.

[0091] FIGs. 12A-12B illustrates more details of the methods described herein. FIG. 12A shows an example event-related potential (ERP). The method may leverage multiple ERPs. An ERP can have different peaks and valleys, herein labeled as Pl, P2, P3, Nl, and N2. These peaks and valleys may represent different psychological traits when measured when a subject is performing specific tasks. As shown in FIG. 12B, ERPs may be used to characterize neural function across different domains, such as reward, emotion, attention,negative bias, and working memory. Different subjects may have different levels of neural function in these different domains.

[0092] Patterns of brain function may be used to subtype disorders to improve or de-risk clinical trials. For example, as shown on the left side of FIG. 13, two different subjects, a drug responder and a drug non-responder, may have different levels of neural function in the different domains. These neural function measurements can provide additional insight during a clinical trial. A configurable platform as described herein may be customized to measure specific domains for different use cases. These methods may be used to simultaneously leverage multiple functional neural, behavioral, and self-report measures to subtype depression based on interpretable brain measures.

[0093] Neurotyping may also be applied to a broader subset of neurological disorders. As shown in FIG. 14, neurotyping may be applied to such neurological diseases and disorders as major depressive disorder (MDD), anxiety, treatment-resistant depression (TRD), obsessive compulsive disorder (OCD), substance use disorders (SUD), chronic insomnia, bipolar disorder, schizophrenia, post-traumatic stress disorder (PTSD), Alzheimer’s disease, Parkinson’s disease, other neurodegenerative disorders, neurotoxicity detection, and / or neurotoxicity impact prediction. In a drug development context, neurotyping may be used in placebo response prediction, treatment resistance prediction or analysis, patient selection criteria determination, understanding the impact of the drug on cognitive or brain function, treatment response prediction or analysis, placebo response analysis, treatment outcome prediction or analysis, or drug response prediction or analysis.Computer Systems

[0094] The present disclosure provides computer systems that are programmed to implement methods of the disclosure. FIG. 5 shows a computer system that is programmed or otherwise configured to provide multi-stage online ordering between buyers and suppliers. The computer system 1901 can regulate various aspects of the present disclosure, for example, for classifying order information and identifying matching business partners. The computer system 1901 can be an electronic device of a user or a computer system that is remotely located with respect to the electronic device.

[0095] The computer system 1901 includes a central processing unit (CPU, also “processor” and “computer processor” herein) 1905, which can be a single core or multi core processor, or a plurality of processors for parallel processing. The computer system 1901 also includes memory or memory location 1910 (e.g., random-access memory, read-only memory, flash memory), electronic storage unit 1915 (e.g., hard disk), communication interface 1920 (e.g., network adapter) for communicating with one or more other systems, and peripheral devices1925, such as cache, other memory, data storage and / or electronic display adapters. The memory 1910, storage unit 1915, interface 1920 and peripheral devices 1925 are in communication with the CPU 1905 through a communication bus (solid lines), such as a motherboard. The storage unit 1915 can be a data storage unit (or data repository) for storing data. The computer system 1901 can be operatively coupled to a computer network (“network”) 1930 with the aid of the communication interface 1920. The network 1930 can be the Internet, an internet and / or extranet, or an intranet and / or extranet that is in communication with the Internet. The network 1930 in some cases is a telecommunication and / or data network. The network 1930 can include one or more computer servers, which can enable distributed computing, such as cloud computing. The network 1930, in some cases with the aid of the computer system 1901, can implement a peer-to-peer network, which may enable devices coupled to the computer system 1901 to behave as a client or a server.

[0096] The CPU 1905 can execute a sequence of machine-readable instructions, which can be embodied in a program or software. The instructions may be stored in a memory location, such as the memory 1910. The instructions can be directed to the CPU 1905, which can subsequently program or otherwise configure the CPU 1905 to implement methods of the present disclosure. Examples of operations performed by the CPU 1905 can include fetch, decode, execute, and writeback.

[0097] The CPU 1905 can execute a sequence of machine-readable instructions, which can be embodied in a program or software. The instructions may be stored in a memory location, such as the memory 1910. The instructions can be directed to the CPU 1905, which can subsequently program or otherwise configure the CPU 1905 to implement methods of the present disclosure. Examples of operations performed by the CPU 1905 can include fetch, decode, execute, and writeback.

[0098] The CPU 1905 can be part of a circuit, such as an integrated circuit. One or more other components of the system 1901 can be included in the circuit. In some cases, the circuit is an application specific integrated circuit (ASIC).

[0099] The storage unit 1915 can store files, such as drivers, libraries and saved programs. The storage unit 1915 can store user data, e.g., user preferences and user programs. The computer system 1901 in some cases can include one or more additional data storage units that are external to the computer system 1901, such as located on a remote server that is in communication with the computer system 1901 through an intranet or the Internet.

[0100] The computer system 1901 can communicate with one or more remote computer systems through the network 1930. For instance, the computer system 1901 can communicate with a remote computer system of a user (e.g., a mobile device). Examples of remotecomputer systems include personal computers (e.g., portable PC), slate or tablet PC’s (e.g., Apple® iPad, Samsung® Galaxy Tab), telephones, smart phones (e.g., Apple® iPhone, Android-enabled device, Blackberry®), or personal digital assistants. The user can access the computer system 1901 via the network 1930.

[0101] Methods as described herein can be implemented by way of machine (e.g., computer processor) executable code stored on an electronic storage location of the computer system 1901, such as, for example, on the memory 1910 or electronic storage unit 1915. The machine executable or machine readable code can be provided in the form of software. During use, the code can be executed by the processor 1905. In some embodiments, the code can be retrieved from the storage unit 1915 and stored on the memory 1910 for ready access by the processor 1905. In some embodiments, the electronic storage unit 1915 can be precluded, and machine-executable instructions are stored on memory 1910.

[0102] The code can be pre-compiled and configured for use with a machine having a processer adapted to execute the code, or can be compiled during runtime. The code can be supplied in a programming language that can be selected to enable the code to execute in a pre-compiled or as-compiled fashion.

[0103] Aspects of the systems and methods provided herein, such as the computer system 1901, can be embodied in programming. Various aspects of the technology may be thought of as “products” or “articles of manufacture” typically in the form of machine (or processor) executable code and / or associated data that is carried on or embodied in a type of machine readable medium. Machine-executable code can be stored on an electronic storage unit, such as memory (e.g., read-only memory, random-access memory, flash memory) or a hard disk. “Storage” type media can include any or all of the tangible memory of the computers, processors or the like, or associated modules thereof, such as various semiconductor memories, tape drives, disk drives and the like, which may provide non-transitory storage at any time for the software programming. All or portions of the software may at times be communicated through the Internet or various other telecommunication networks. Such communications, for example, may enable loading of the software from one computer or processor into another, for example, from a management server or host computer into the computer platform of an application server. Thus, another type of media that may bear the software elements includes optical, electrical and electromagnetic waves, such as used across physical interfaces between local devices, through wired and optical landline networks and over various air-links. The physical elements that carry such waves, such as wired or wireless links, optical links or the like, also may be considered as media bearing the software. As used herein, unless restricted to non-transitory, tangible “storage” media, terms such as computeror machine “readable medium” refer to any medium that participates in providing instructions to a processor for execution.

[0104] Hence, a machine readable medium, such as computer-executable code, may take many forms, including but not limited to, a tangible storage medium, a carrier wave medium or physical transmission medium. Non-volatile storage media include, for example, optical or magnetic disks, such as any of the storage devices in any computer(s) or the like, such as may be used to implement the databases, etc. shown in the drawings. Volatile storage media include dynamic memory, such as main memory of such a computer platform. Tangible transmission media include coaxial cables; copper wire and fiber optics, including the wires that comprise a bus within a computer system. Carrier-wave transmission media may take the form of electric or electromagnetic signals, or acoustic or light waves such as those generated during radio frequency (RF) and infrared (IR) data communications. Common forms of computer-readable media therefore include for example: a floppy disk, a flexible disk, hard disk, magnetic tape, any other magnetic medium, a CD-ROM, DVD or DVD-ROM, any other optical medium, punch cards paper tape, any other physical storage medium with patterns of holes, a RAM, a ROM, a PROM and EPROM, a FLASH-EPROM, any other memory chip or cartridge, a carrier wave transporting data or instructions, cables or links transporting such a carrier wave, or any other medium from which a computer may read programming code and / or data. Many of these forms of computer readable media may be involved in carrying one or more sequences of one or more instructions to a processor for execution.

[0105] The computer system 1901 can include or be in communication with an electronic display 1935 that comprises a user interface (UI) 1940 for providing, for example, a dashboard. Examples of UI’s include, without limitation, a graphical user interface (GUI) and web-based user interface.

[0106] Methods and systems of the present disclosure can be implemented by way of one or more algorithms. An algorithm can be implemented by way of software upon execution by the central processing unit 1905.Computer Programs

[0107] In some embodiments, the platforms, systems, media, and methods disclosed herein include at least one computer program, or use of the same. A computer program includes a sequence of instructions, executable by one or more processor(s) of the computing device’s CPU, written to perform a specified task. Computer readable instructions may be implemented as program modules, such as functions, objects, Application Programming Interfaces (APIs), computing data structures, and the like, that perform particular tasks orimplement particular abstract data types. In light of the disclosure provided herein, those of skill in the art will recognize that a computer program may be written in various versions of various languages.

[0108] The functionality of the computer readable instructions may be combined or distributed as desired in various environments. In some embodiments, a computer program comprises one sequence of instructions. In some embodiments, a computer program comprises a plurality of sequences of instructions. In some embodiments, a computer program is provided from one location. In other embodiments, a computer program is provided from a plurality of locations. In various embodiments, a computer program includes one or more software modules. In various embodiments, a computer program includes, in part or in whole, one or more web applications, one or more mobile applications, one or more standalone applications, one or more web browser plug-ins, extensions, add-ins, or add-ons, or combinations thereof.Web Application

[0109] In some embodiments, a computer program includes a web application. In light of the disclosure provided herein, those of skill in the art will recognize that a web application, in various embodiments, utilizes one or more software frameworks and one or more database systems. In some embodiments, a web application is created upon a software framework such as Microsoft .NET or Ruby on Rails (RoR). In some embodiments, a web application utilizes one or more database systems including, by way of non-limiting examples, relational, nonrelational, object oriented, associative, XML, and document oriented database systems. In further embodiments, suitable relational database systems include, by way of non-limiting examples, Microsoft SQL Server, mySQL, and Oracle. Those of skill in the art will also recognize that a web application, in various embodiments, is written in one or more versions of one or more languages. A web application may be written in one or more markup languages, presentation definition languages, client-side scripting languages, server-side coding languages, database query languages, or combinations thereof. In some embodiments, a web application is written to some extent in a markup language such as Hypertext Markup Language (HTML), Extensible Hypertext Markup Language (XHTML), or extensible Markup Language (XML). In some embodiments, a web application is written to some extent in a presentation definition language such as Cascading Style Sheets (CSS). In some embodiments, a web application is written to some extent in a client-side scripting language such as Asynchronous JavaScript and XML (AJAX), Flash ActionScript, JavaScript, or Silverlight. In some embodiments, a web application is written to some extent in a server-side coding language such as Active Server Pages (ASP), ColdFusion, Perl, Java, JavaServerPages (JSP), Hypertext Preprocessor (PHP), Python, Ruby, Tel, Smalltalk, WebDNA, or Groovy. In some embodiments, a web application is written to some extent in a database query language such as Structured Query Language (SQL). In some embodiments, a web application integrates enterprise server products such as IBM Lotus Domino. In some embodiments, a web application includes a media player element. In various further embodiments, a media player element utilizes one or more of many suitable multimedia technologies including, by way of non-limiting examples, Adobe Flash, HTML 5, Apple QuickTime, Microsoft Silverlight, Java, and Unity.Mobile Application

[0110] In some embodiments, a computer program includes a mobile application provided to a mobile computing device. In some embodiments, the mobile application is provided to a mobile computing device at the time it is manufactured. In other embodiments, the mobile application is provided to a mobile computing device via the computer network described herein.[OHl] In view of the disclosure provided herein, a mobile application is created by techniques known to those of skill in the art using hardware, languages, and development environments known to the art. Those of skill in the art will recognize that mobile applications are written in several languages. Suitable programming languages include, by way of non-limiting examples, C, C++, C#, Objective-C, Java, JavaScript, Pascal, Object Pascal, Python, Ruby, VB.NET, WML, and XHTML / HTML with or without CSS, or combinations thereof.

[0112] Suitable mobile application development environments are available from several sources. Commercially available development environments include, by way of non-limiting examples, AirplaySDK, alcheMo, Appcelerator, Celsius, Bedrock, Flash Lite, .NET Compact Framework, Rhomobile, and WorkLight Mobile Platform. Other development environments are available without cost including, by way of non-limiting examples, Lazarus, MobiFlex, MoSync, and Phonegap. Also, mobile device manufacturers distribute software developer kits including, by way of non-limiting examples, iPhone and iPad (iOS) SDK, Android SDK, BlackBerry SDK, BREW SDK, Palm OS SDK, Symbian SDK, webOS SDK, and Windows Mobile SDK.Standalone Application

[0113] In some embodiments, a computer program includes a standalone application, which is a program that is run as an independent computer process, not an add-on to an existing process, e.g., not a plug-in. Those of skill in the art will recognize that standalone applications are often compiled. A compiler is a computer program(s) that transforms sourcecode written in a programming language into binary object code such as assembly language or machine code. Suitable compiled programming languages include, by way of non-limiting examples, C, C++, Objective-C, COBOL, Delphi, Eiffel, Java, Lisp, Python, Visual Basic, and VB .NET, or combinations thereof. Compilation is often performed, at least in part, to create an executable program. In some embodiments, a computer program includes one or more executable complied applications.Software Modules

[0114] In some embodiments, the platforms, systems, media, and methods disclosed herein include software, server, and / or database modules, or use of the same. In view of the disclosure provided herein, software modules are created by techniques known to those of skill in the art using machines, software, and languages known to the art. The software modules disclosed herein are implemented in a multitude of ways. In various embodiments, a software module comprises a file, a section of code, a programming object, a programming structure, a distributed computing resource, a cloud computing resource, or combinations thereof. In further various embodiments, a software module comprises a plurality of files, a plurality of sections of code, a plurality of programming objects, a plurality of programming structures, a plurality of distributed computing resources, a plurality of cloud computing resources, or combinations thereof. In various embodiments, the one or more software modules comprise, by way of non-limiting examples, a web application, a mobile application, a standalone application, and a distributed or cloud computing application. In some embodiments, software modules are in one computer program or application. In other embodiments, software modules are in more than one computer program or application. In some embodiments, software modules are hosted on one machine. In other embodiments, software modules are hosted on more than one machine. In further embodiments, software modules are hosted on a distributed computing platform such as a cloud computing platform. In some embodiments, software modules are hosted on one or more machines in one location. In other embodiments, software modules are hosted on one or more machines in more than one location.Databases

[0115] In some embodiments, the platforms, systems, media, and methods disclosed herein include one or more databases, or use of the same. In view of the disclosure provided herein, those of skill in the art will recognize that many databases are suitable for storage and retrieval of, by way of examples, image, cell state, protocol, and culture condition information. In various embodiments, suitable databases include, by way of non-limiting examples, relational databases, non-relational databases, object-oriented databases, objectdatabases, entity-relationship model databases, associative databases, XML databases, document-oriented databases, and graph databases. Further non-limiting examples include SQL, PostgreSQL, MySQL, Oracle, DB2, Sybase, and MongoDB. In some embodiments, a database is Internet-based. In further embodiments, a database is web-based. In still further embodiments, a database is cloud computing-based. In a particular embodiment, a database is a distributed database. In other embodiments, a database is based at least in part on one or more local computer storage devices.Data Analysis

[0116] In some embodiments, the method may further comprise generating data visualizations of the users’ neurotype. In some cases, the method may comprise generating a time-series graphical representation of the functional brain metrics.

[0117] In some embodiments, the method may further comprise generating predictive insights about a user. In some cases, the method comprises using the model described herein to generate one or more predictive insights indicative of the user’s neurob ehavi oral traits. In some cases, said predictive insights may further comprise generating a prediction about a user’s neurob ehavi oral traits based on comparison between the data of said user and the data of one or more other users. In some embodiments, the method comprises comparing the neurobehavioral traits of the user with neurob ehavi oral traits of one or more other users. In some cases, this comparison is based at least in part on measured differences between the user’s functional brain metrics and the one or more other users’ functional brain metrics. In some embodiments, the method may further comprise using the model to recommend suggestions for the user to improve the user’s well-being.

[0118] In some embodiments, the method may further comprise using the model in a healthcare context to recommend clinical interventions for the user based on the user’s neurotype or the predictive insights. In some embodiments, the method may comprise using the model to predict treatment responses for a health condition or disorder that the user has or is suspected to have. In some cases, the method can be used by healthcare providers to understand the background of their patients. In some cases, the EEG-based mental status information that healthcare providers get from this invention will include data that cannot be obtained from just interviews or questionnaires. With this information, healthcare providers can focus on the patient’s specific problems and provide better treatment options.

[0119] For an individual consumer, the method described herein can be used to help the consumer better understand their own mental health and cognitive features. By better understanding their own mental and cognitive traits, the individual consumer can identify potential problems, which may help them take steps to avoid serious mental health orcognitive events. By using this invention, individuals may be able to proactively manage their well-being.

[0120] This method can also be used by enterprises to understand their employees’ mental health. Employers may be able to use the mental health information obtained from EEG data to proactively prevent mental health issues among their employees.EXAMPLES

[0121] The following examples are included for illustrative purposes only and are not intended to limit the scope of the invention.

[0122] While preferred embodiments of the present invention have been shown and described herein, it will be obvious to those skilled in the art that such embodiments are provided by way of example only. Numerous variations, changes, and substitutions will now occur to those skilled in the art without departing from the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in practicing the invention. It is intended that the following claims define the scope of the invention and that methods and structures within the scope of these claims and their equivalents be covered thereby.Example 1:

[0123] Depression is a prevalent, burdensome, and difficult mental health disorder to treat. Significant heterogeneity in clinical characteristics and course of depression may hinder treatment success. Efforts to identify more homogeneous subgroups of depression could reduce heterogeneity of depression and therefore improve treatment development and randomized clinical trial outcomes. Event-related potentials (ERPs) derived from continuous electroencephalogram (EEG) can be used to identify depression and predict course (i.e., advance precision psychiatry).

[0124] In the current study, it was demonstrated how multiple ERPs collected from the same individual across different experimental paradigms can provide insight into brain function and individual differences in depression using factor analysis. This approach for neurotyping depression exploits the high within-task and low between-task associations between ERPs to better understand brain function and depression.

[0125] Three neurotypes were observed, two of which differentiated depressed from nondepressed individuals. Only one neurotype - related to affective processing - prospectively predicted full remission. This neurotype predicted remission even when accounting for other clinical and demographic variables related to subsequent remission. The AUC of thisneurotype was acceptable (i.e., .72) in predicting remission, exceeding previous study’s measures within a single task.

[0126] Leveraging multiple ERPs derived from many tasks is an important yet underutilized approach in precision psychiatry.Introduction:

[0127] Post-COVID estimates suggest that nearly one in four individuals suffer with clinically significant levels of depression (Bueno-Notivol et al., 2021; Wu et al., 2021). Depression is associated with increased mortality, including increased risk of suicide (Cai et al., 2021) as well as morbidity from other health conditions (e.g., coronary heart disease; Meng et al., 2020). Depression is among the leading causes of disability (Vos et al., 2016), and in the US alone, the economic cost of depression exceeds 300 billion dollars annually (Greenberg et al., 2020).

[0128] The significant heterogeneity in depressive symptom presentation (Zimmerman et al., 2015) and course (Klein & Hajcak, 2015; Monroe & Harkness, 2022) can present difficulties for developing more efficacious treatments. Randomized controlled trials (RCTs) that test the efficacy of a drug for “depression” likely confound many subgroups who may not respond to the same intervention. Indeed, only about half of depressed individuals respond to antidepressant medications (Levkowitz et al, 2011); even after up to four antidepressant treatment trials, the remission rate in the STAR*D trial was 35.0% (Pigott et al., 2023). Complicating matters further, about one in three depressed individuals have a favorable response to placebo (Furukowa et al., 2016) and many untreated depressed individuals show spontaneous improvement after a year (Whiteford et al., 2013). Mirroring antidepressant treatment research, only about a third of individuals experience remission after completing psychotherapy for depression (Cuijpers et al., 2021). There is, therefore, a pressing need and opportunity to identify more homogenous subgroups of depression based on objective measures to conceptualize psychopathology more accurately so that more targeted and effective treatments can be developed.

[0129] The current paper focuses on measures of brain function derived from the electroencephalogram (EEG) called Event-Related Potentials (ERPs). ERPs are direct measures of electrical activity of the brain with excellent temporal resolution that can be leveraged to study many cognitive and emotional functions (e.g., attention, memory, cognitive control; Hajcak, MacNamara, & Olvet, 2010; Luck, 2014). Insofar as EEG hardware is cheaper and has fewer counterindications than MRI, ERP measures of brain function are ideal candidate neural biomarkers. Moreover, ERP measures of brain function have good psychometric properties, including internal and test-retest reliability (e.g., Meyer,Bress, & Proudfit, 2014; Hajcak, Meyer, & Kotov, 2017; Santopetro, Mulligan, Brush, & Hajcak, 2022). Although conventional EEG systems have been stationary (i.e., set up in a research-designated environment) and use a conducive medium (i.e., gel) between the electrode and the scalp, recent advances in EEG hardware have included EEG systems that are small and mobile (e.g., LiveAMP; Brush et al., 2021), and do not require gel (i.e., dry electrode EEG systems; Tepan, 2002). As the cost of high-quality EEG hardware comes down, there is the possibility that ERP assessments could be achieved at scale.

[0130] ERP studies have consistently found that individuals with depression are characterized by reduced neural activity related to reward (i.e., the Reward Positivity or RewP; Keren et al., 2018) and blunted emotional reactivity to positive images (the late positive potential or LPP; Proudfit et al., 2015). In these instances, depression-related effects are typically reported in the context of condition-related differences: depressed individuals are characterized by a reduced difference between ERPs following monetary gain versus loss (i.e., the RewP; Proudfit, 2015) and a reduced difference between ERPs elicited by emotional compared to neutral images (i.e., the LPP; Klawohn et al., 2021).

[0131] Moving beyond correlates of depression, it has been found that ERPs such as the RewP and LPP have predictive utility in depression: a smaller neural response to rewards (i.e., a blunted RewP) can prospectively predict first-onset depression (Nelson et al., 2018). Even among depressed adults with a blunted RewP, a smaller RewP predicts less likelihood of remission in the following nine months (Klawohn, Brush, & Hajcak, 2021). A smaller LPP has also been related to earlier-onset depression (Weinberg, Perlman, Kotov, & Hajcak, 2016), and predicts worse treatment response in depression (Barch et al., 2020).

[0132] Many studies tend to follow the typical approach of publishing separate papers based on ERPs derived from different tasks. This strategy is useful insofar as a given paper is focused on a single brain function, or a small number of functions assessed within the same task. However, it is unclear whether various ERP differences in depression derived from different tasks reflect overlapping or distinct aspects of the disorder. It was found that the LPP and RewP were independently related to depression (Klawohn et al., 2021); indeed, both the LPP and RewP also interpedently predicted remission among depressed individuals (Klawohn, Brush, & Hajcak, 2021). These data suggest that that reward insensitivity and emotional insensitivity may reflect two neurotypes (i.e., neural traits) of depression. The term neurotype may be used to describe distinct functional neural activity, analogous to personality traits — but measured in terms of specific brain functions.

[0133] In the typical approach, between-condition differences are used to identify neurotypes (e.g., examining the ERP to gain and losses to isolate reward-related neural activity [RewP],and then relating the RewP to depression). One practical issue of this approach is the high correlation between condition-related ERPs. For instance, the ERP for gain and loss trials are highly correlated with one another (e.g., .73; Thompson et al., 2023) — and the resulting difference score only demonstrates modest reliability (e.g., .28-.38; Levinson et al., 2017). Thus, although between-condition differences are useful for isolating neural activity specific to a cognitive function, the same between-condition difference scores may be poorly suited measures for studying differences between individuals (i.e., as neural biomarkers; Clayson, Baldwin, & Larson, 2021).

[0134] The existing approach based on condition-related differences misses the opportunity to leverage task-related variability to better understand brain function and individual differences. As an example of how variability common to multiple ERPs derived from the same task might reflect the same variability in depression, it was found that the stimulus- locked P300 and the response-locked error positivity (Pe) both differentiated depressed from non-depressed adults; however, it was their shared variance that related to depression (Santopetro et al., 2021) — suggesting that they reflected similar dysfunction in depression.

[0135] Here, a novel approach is presented that considers task context as an alternative source of variability in neural measures. To do this, multiple ERPs derived from three different tasks were submitted - all of which have been previously related to depression - to a factor analysis. It was considered whether factor analytically derived neurotypes based on task-related variability relate to depression: resulting factor scores would then situate depressed individuals along a dimension defined by the covariance between specific ERPs. Unlike the traditional approach of leveraging between-condition variance in the context of individual differences, this approach tests whether between-task variance might similarly be useful in creating neurotypes.

[0136] In what follows, this possible approach was considered empirically. A factor analysis was conducted on seven ERPs, obtained across three tasks. ERPs related to emotion (i.e., the LPP), reward (i.e., the RewP), attention and salience (i.e., P300), and error awareness (Pe) were included, all of which have been found to be reduced in depression in prior studies. The current factor analytic method is meant to be illustrative of a new neurotyping approach — rather than a definitive characterization of neurotyping in depression. After factor analysis, resulting neurotypes between depressed and non-depressed adults were compared. Neurotypic differences between those with MDD who would go on to fully remit in the next 9 months versus those with MDD who would not remit were then examined. It was also examined whether neurotypes could predict remission over and above demographic and clinical characteristics that also predicted remission. Overall then, it was examined whetherneurotypic variability, defined by the covariation of ERPs derived from multiple task contexts, could be useful in differentiating MDD from healthy controls, as well as predict remission among those with MDD.Methods:Participants

[0137] The current paper builds upon data previously reported on from the monetary doors and emotional picture viewing (Klawohn, Brush, & Hajcak, 2021; Klawohn et al., 2021) and flanker paradigms (Klawohn, Santopetro, Meyer, & Hajcak, 2020; Santopetro, Brush, Burani et al., 2022) from a large sample of depressed individuals. Details on clinical assessments and clinical characterization are also provided in Klawohn, Brush, & Hajcak (2021) and Klawohn et al. (2021). Prior to participation, volunteers received verbal and written explanations of aims and procedures of the study and provided informed written consent. For brevity, only key methodological aspects of the study are reported here.

[0138] Participants were included in the current depression group (i.e., MDD group) if they met diagnostic criteria for a current mood disorder (i.e., current major depressive disorder [MDD] and / or persistent depressive disorder [PDD]) and scored higher than 13 on the BDI-II (Beck et al., 1996) for current depressive symptoms in the past two weeks. Participants were included in the healthy control group (i.e., HC group) if they never met diagnostic criteria for a mood disorder, did not currently meet criteria for any other psychiatric disorder, and scored below 13 in the BDI-II. Participants in the MDD group were also invited back to the lab for a follow-up visit that occurred approximately 9 months after their initial lab visit (M = 8.8 months, SD = 1.5) and once again completed clinical interviews and self-report measures. A follow-up period of 9 months (rather than a year) was set based in part on the logistics and timeline of funding, as well as based on a larger study which found relatively high rates of remission (i.e., 25%-48%) within this briefer period among nearly 1,000 depressed adults (Fleck, Simon, Herrman, Bushnell, Martin, & Patrick, 2005). The study was conducted in accordance with the ethical guidelines of the Declaration of Helsinki and approved by the Florida State University Institutional Review Board. All participants were reimbursed for study participation (i.e., $20 per hour) and received $7.50 for completing the reward task.

[0139] The study sample included in the present project consisted of 98 total participants who had usable EEG data from all three of the experimental paradigms of interest: 64 individuals with a current depressive disorder (MDD) and 34 healthy controls (HC), 46 of the 64 MDD participants returned approximately nine months later for the follow-up visit.Clinical Measures

[0140] Diagnostic assessment was performed using the Structured Clinical Interview for DSM-5-Research Version for the Diagnostic and Statistical Manual, Fifth Edition (SCID-5- RV; First et al., 2015); the Montgomery-Asberg Depression Rating Scale (MADRS; Montgomery and Asberg, 1979; scores on this measure range from 0 to 60) was also administered to assess depression severity - and both interviews were conducted by two PhD-level clinical psychologists. It was determined whether participants achieved an episode of full remission at any time between their baseline and follow-up visit, defined as a period of at least two consecutive months during which participants did not meet diagnostic criteria for either a major depressive episode (MDE) or PDD diagnosis.EEG Recording and Processing

[0141] EEG data was acquired using a 32 channel actiChamp system (Brain Products GmbH, Gilching, Germany), and EEG data were processed using Brain Vision Analyzer, Version 2.1 (Brain Products, Gilching, Germany). Data were referenced to the average of the mastoid electrodes and filtered from 0.01 to 30 Hz (Butterworth, 4th order). For all tasks, epochs were first extracted and corrected for eye movement artifacts using the algorithm developed by Gratton and Coles (1983). Segments that contained voltage steps > 50 mV between sample points, a voltage difference of 175 mV within a 400 ms interval, or a maximum voltage difference of < 0.5 mV within 100 ms intervals were automatically rejected.Additional artifacts were identified and removed based on visual inspection. Baselinecorrection was applied using the average activity in a 200 ms window, unless otherwise noted below. All ERP measurement windows and electrode sites used were determined using a collapsed localizer approach (Luck & Gaspelin, 2017), such that data was inspected after collapsing groups; parameters were chosen in the time range and electrode sites showing maximal activity.

[0142] Continuous EEG recording took place in a single experimental session consisting of the monetary doors guessing task, flanker task, and passive emotional picture viewing task. The monetary doors task was always completed first because the primary goal of the original study related to depression and reward sensitivity — and the order of the other two tasks were counterbalanced. All tasks were administered using Presentation software (Neurobehavioral Systems, Inc., Albany, CA, USA).Doors Task

[0143] Participants completed a doors task, in which they were first presented with an image of two doors and were instructed that they would either win or lose money on each trial (Proudfit, 2015). After selecting a door with the left or right mouse button on each trial,participants were presented with either an upward green arrow or a downward red arrow indicating monetary gain (+$0.50) or loss (-$0.25), respectively, for 2,000 ms. The task consisted of 30 gain trials and 30 loss trials, presented in a pseudo-randomized order. The study focused on doors- and feedback-locked epochs that were extracted with a duration of 1000 ms, beginning 200 ms before the presentation of the doors or feedback. The doors- locked P300 was scored as the mean activity from 200 ms to 500 ms at an occipital pooling of electrode sites 01, Oz, and 02. The feedback-locked EPRs were averaged separately for gains and losses, and scored as the mean activity from 250 ms to 350 ms after feedback presentation at FCz; this activity reflected variability in the ERP that defines the RewP. Flankers Task

[0144] An arrow version of the flankers task was utilized (e.g., Hajcak & Foti, 2008;Klawohn, Santopetro, Meyer, & Hajcak, 2020; Santopetro, Mulligan, Brush, & Hajcak, 2022). On each of 330 trials, five horizontal white arrowheads were presented afor 200 ms. Participants were instructed to respond as quickly and accurately as possible in response to the direction of the center arrow using the left and right buttons on the computer mouse. EEG data epochs of 1,000 ms were extracted starting 200 ms prior to stimulus onset on correct trials and then averaged combining compatible and incompatible trials. The P300 was quantified as the mean amplitude between 300 ms to 600 ms at electrode site Pz. For analyses of the Pe, data was first segmented starting 500 ms before response errors and continuing for 1,000 ms; segments were baseline-corrected using the interval from 500 ms to 300 ms before the error. The Pe was scored at electrode site Pz as the mean amplitude from 200 ms to 400 ms after errors.Passive Picture Viewing Task

[0145] The picture viewing task comprised 30 pleasant images (e.g. erotic and affiliative images) and 30 neutral images (e.g. objects, humans with neutral facial expression) from the International Affective Picture System (ZAPS; Lang et al., 2008); these were presented for 1,500 ms in random order across three blocks of 20 trials. The LPP was quantified at a parietal electrode-pool (Pz, Cz, CPI, and CP2) as the mean amplitude from 400-1000 ms after picture-onset following pleasant and neutral pictures (e.g., Weinberg et al., 2016).ERP Psychometrics

[0146] Internal consistencies were computed by correlating the odd- and even-numbered trials and then correcting this using the Spearman-Brown prophecy formula (Nunnally, Bernstein, & Berge, 1967). The ERPs demonstrated the following internal consistencies ranging from acceptable-to-excellent reliability in each group: flanker P300 (DEP: .98; HC .98), Pe (DEP: .85, HC: .84), doors-locked P300 (DEP: .92 HC: .89), RewP (gain trials: DEP:.92, HC: .93; loss trials: DEP: .90, HC: .93), positive LPP (DEP: .78, HC: .84), and neutral LPP (DEP: .64, HC: .67).Statistical Analysis

[0147] Bivariate correlations were conducted between all ERP components of interest (Table 1). A principal axis factor analysis was conducted using the correlation matrix and promax rotation to extract factors with Eigenvalues greater than one. This resulted in three factors; the pattern matrix is presented in Table 2. Factor scores for each factor (i.e., neurotype) were saved, and between-group (i.e., MDD vs HC) differences were examined on each neurotype using independent-samples t-tests (Table 2, bottom). Among those with MDD, it was then examined whether factor-analytically derived neurotypes could differentiate those MDD who would remit in the upcoming 9 months (N=19) and those who would not achieve full remission in the upcoming 9 months (N=27). In addition, a range of demographic (i.e., age, sex, race, income, education level) and clinical (i.e., MADRS, BDI, medication status, comorbid disorders) measures were analyzed in relation to future remission. Given that earlier onset and chronicity are associated with worse course of depression (Klein & Hajcak, 2015), it was also examined whether the diagnosis of PDD and age of first onset of depression predicted remission. All independent sample t-tests were one-sided, unless otherwise noted. A logistic regression was then run to determine whether neurotypic data could predict remission even when accounting for demographic and clinical variables that also predicted remission. Finally, ROC analysis was utilized and the area under the curve (AUC) was examined for each factor score to differentiate those MDD who would achieve full remission at follow-up from those who would not.RewP FRN Doors Pleasant Neutral Flankers(Gain) (Loss) P300 LPP LPP P300RewP (Gain)FRN (Loss) 90**Doors P300 -.09 -.12Pleasant LPP .08 .05 .16Neutral LPP .21* .14 -.07 .48**Flankers P300 .26* .23* .25* .26** .14Error Positivity .25* .23* .23* .11 .03 .51**(Pe)Table 1: Zero-order correlations between all ERP components (n = 98). Note. *indicates p < .05. ** indicates p < .01.Factor 2: Decision-making and Monitoring .48*Factor 3: Affective Processing .33* .38*HC (N=34; mean [st dev]) .16 (1.04) .31 (.86)* .28 (.83)*MDD (N=46; mean [st dev]) -.06 (.95) -.20 (.87)* -.12 (.68)*Table 2: Pattern matrix for factor analysis (top) and intercorrelations between Factors (bottom Note. * indicates variables highest loading (top), significant correlation at p< 05 (middle), and significant group differences at p< 05 (bottom).Results

[0148] As evident from Table 1, various ERP measures ranged from being uncorrelated to highly correlated with one another. Across tasks, the general trend was for low correlations; the highest correlations were evident between within-task (e.g., although the flankers P300 was most correlated with the error positivity, it was modestly correlated with most other ERP measures).

[0149] The pattern matrix suggested the existence of three factors — which roughly group by task, with one notable exception: the doors-locked P300 loaded on Factor 2 with the P300 and Pe elicited within the Flankers task. These three neurotypes are referred to as Outcome Evaluation (i.e., ERPs elicited by gain and loss feedback), Decision-making and Monitoring (i.e., P300 elicited by stimuli requiring a response and response error), and Affective Processing (i.e., LPP elicited by neutral and emotional pictures). It is important to note that the cross-loadings were all low. Table 2 (middle) also presents correlations between the three factor scores, which suggest relatively modest overlap between factor scores.

[0150] Table 2 (bottom) presents the means and standard deviation for factor scores for healthy control participants (HC), as well as the full group of MDD. Factor scores on both the Decision-making and Monitoring (t(78)=2.63; / ?<.01) and Affective Processing (t(78)=2.30, / ?<.05) factor scores differentiated MDD from HC, whereas the Outcome Evaluation Factor scores (t(78)< 1 ) did not. These data suggest that MDD were characterized by reduced P300 and P300-like ERP responses, as well as reduced LPP to neutral and pleasant pictures. However, the shared variance between gain and loss trials in the timerange of the RewP did not differentiate MDD from HC.

[0151] To examine whether these factor scores also differentiated later remission among the MDD, those MDD who would go on to achieve full remission versus those who would not go on to remit in the upcoming 9 months (Table 3, top) were subsequently compared. Later remission was not predicted by the Outcome Evaluation (t(44)<l) or Decision-making and Monitoring Factors (Z(44)< 1 ); however, MDD who would go on to remit had more positive Affective Processing Factor scores than MDD who would not go on to remit in the following 9 months ( / (44)=2. l 5, <.05). The same three factor structure emerged when conducting the factor analysis in individuals with MDD only.1stMDE (age) 19.1 (9.9)* 25.3 (14.6)*Table 3: Demographic, clinical, and neurotyping factor scores (means and standard deviations) comparing individuals with depression (MDD) - as a function of later remission. Note. *p< 10; ** p< 05

[0152] To visualize these effects, FIG. 15 presents the ERP from the picture-viewing paradigm, averaged across neutral and pleasant pictures, for HC, MDD who would remit, and MDD who would not remit. Consistent with the factor analytic results in Table 3, the averaged LPP is smallest among MDD who would not later remit, and did not appear different when comparing HC and MDD who would later remit. FIG. 15 shows averaged neural activity collapsed across positive and neutral images in individuals with depression that did not remit (MDD; n =27), individuals with depression that did remit (REM; n = 19), and controls with no history of psychopathology (HC; n = 34).

[0153] Table 3 (middle, bottom) also presents demographic (i.e., age, sex, race, income, education) and clinical (i.e., MADRS score, BDI score, medication status, presence of comorbid disorders, whether or not criteria for PDD was met, and age of first MDE) characteristics among MDD who would go on to remit versus those who would not later remit. Diagnosis of PDD predicted less likelihood of remission (t(44)=1.96, / ?< 05, onesided); in addition, there were trend level associations indicating that individuals with MDD who would later achieve full remission were characterized by higher income (t(40)=1.34, < 10, one-sided) and earlier age of first MDE ( / (43)=1.57, p<.10, one-sided).

[0154] When entered into a logistic regression, Affective Processing neurotype remained a significant predictor (P=l.l 1, p<.05), and PDD continued to predict remission at a trend level (P=-l .29, p< 06). This regression analysis was consistent with the fact that Affective Processing neurotype and PDD diagnosis were unrelated to one another (r = -,03, / ?>.80). Affective Processing neurotype remained a significant predictor of remission (P=l .23, p< .05) even when including PDD diagnosis, income, and age of first MDE in a logistic regression; indeed, none of these other predictors of remission approached significance when all were included in the same logistic regression.

[0155] Finally, the AUC for differentiating individuals experiencing depression at baseline who would go on to achieve full remission from those who would not using Receiver Operating Characteristic (ROC) analyses and the Affective Processing neurotype exhibited excellent discrimination (AUC =.72).Discussion

[0156] Whereas previous studies have focused on individual differences in condition-related effects within a task to understand depression, the current approach leverages individual differences in ERPs recorded across multiple experimental paradigms to delineate neural functions and their relationships with depression. This approach exploits the large proportion of variance in neural activity that is shared between conditions within a task, as well as the relatively low correlation between brain activity across tasks. Seven ERPs (i.e., RewP, LPP, P300, and P300-like components) obtained across three different tasks were focused on, all from the same participants.

[0157] The resulting factor analysis suggested three neural factors which appeared to reflect the functional context in which the ERPs were elicited: an Outcome Evaluation factor emerged from ERP activity on gain and loss trials in the time-range of the RewP (i.e., time- locked to the presentation of gain and loss feedback within the doors gambling task); a Decision-making and Monitoring factor explained the shared variance between the flankers- locked P300, the error-positivity (Pe), and the doors-locked P300; additionally, an Affective Processing factor captured variance shared between the LPP elicited while viewing neutral and pleasant pictures. Importantly, the Decision-making and Monitoring P300 factor captured variance shared between the flankers and doors tasks (i.e., the resulting factors did not simply reflect variance shared within tasks). Insofar as the Decision-making and Monitoring factor included all ERP components that were elicited by stimuli that required responses (i.e., both the imperative doors and flankers stimuli) or response-related processing (i.e., the response- locked error positivity), these data suggest that neurotypes defined through factor analysis could reflect shared functions of the ERP components.

[0158] In comparing depressed to non-depressed individuals using these factor-analytically defined neurotypes, it was found that depressed individuals were characterized by reduced neural scores on the Decision-making and Monitoring factor as well as the Affective Processing factor — but not the Outcome Evaluation factor. These data suggest that shared variability in ERPs can be used to delineate neurotypes that vary with depression - as well as those that do not. These data highlight the potential of using the functional context in the which ERPs are elicited for understanding depression. The proposed approach would benefit from much larger and diverse samples with many more ERPs obtained across more cognitive domains.

[0159] The potential utility of the current approach was further demonstrated when neurotypic variability was considered in terms of subsequent remission within individuals suffering from depression. Those MDD who would go on to remit were characterized byincreased Affective Processing factor scores - that is, those MDD who would remit had LPPs to neutral and pleasant pictures that were more like healthy controls than like MDD. Those individuals with MDD who were more likely to later remit were less likely to have a diagnosis of PDD — and were also characterized by trend-level increases in income and an earlier age of first MDE. The Affective Processing neurotype continued to predict remission even when accounting for these demographic and clinical characteristics. Indeed, the Affective Processing neurotype was the best predictor of remission among MDD.

[0160] When considering the Affective Processing neurotype within depression, the classification accuracy (i.e., AUC) in predicting remission was acceptable (i.e., .72) - and exceeded previous measures that predicted remission using condition-based differences within a single task (Klawohn, Brush, & Hajcak, 2021). The ability of the Affective Processing neurotype to predict remission was also specific: although the Decision-making and Monitoring neurotype differentiated depressed from non-depressed individuals, it did not predict remission. Collectively, these data suggest important differences in task-based neurotypes in terms of those that differentiate depressed from non-depressed adults versus those that can also predict changes in depression prospectively. Individuals experiencing depression who would remit (REM) did not differ from controls (HC; Z(51 )<1 ) on their Affective Processing scores, whereas individuals with depression who would not remit (MDD) did differ from HC (t(59) = 3.0, / ?< 01).

[0161] The current approach to neurotyping depression has the potential to advance precision psychiatry - it reflects a scalable approach to reducing the heterogeneity of depression to improve outcomes with more targeted interventions. In considering the extreme heterogeneity of depression, one of the main ways in which depression can differ is in terms of course (Klein & Kotov, 2016). In fact, one meta-analysis suggests that nearly one-third of untreated depressed adults in primary care settings will remit within three months, and nearly a half will remit within a year (Whiteford et al., 2013). The current results suggest that reductions in a specific neurotype (i.e., reduced neural activity in the context of affective processing) could be used to predict remission. This particular neurotype (i.e., reduced LPP in the context of affective processing) may be particularly useful in the development and evaluation of novel therapies for depression as it could be used to stratify depressed individuals who would be more likely to remit on their own or respond to placebo.

[0162] The specific factor analytic approach in the current study was intended as an example of one method of how neurotyping is possible by examining multiple ERPs from different experimental paradigms. This approach may work particularly well if there were additional ERPs and tasks employed. For instance, evidence has been found that multiple reward-relatedfunctions (i.e., RewP, FN, cue-P300) are abnormal in relation to depression during adolescence (Thompson et al., 2023). It could be interesting, for instance, to examine multiple measures of reward functioning and reward anticipation (i.e., stimulus preceding negativity; SPN) across multiple tasks using the same factor analytic approach conducted in the current study to determine whether other neurotypes could be derived by examining variability shared between ERPs within a task. Future studies might consider alternative statistical approaches for leveraging shared variance across ERPs as measures of individual differences (e.g., nest trial types within tasks using multi-level modeling).

[0163] It is also plausible to consider combining the current method with more traditional approaches that examine condition-related differences in relation to depression. Consider the Affective Processing neurotype in the current study — this neurotype leverages the shared variance between ERPs elicited by neutral and pleasant pictures. We have previously found that the unique variance in the ERP elicited by pleasant pictures (i.e., the degree to which pleasant is greater than neutral) relates to and predicts course in depression (Klawohn et al., 2021; Klawohn, Brush, & Hajcak, 2021). Both measures in affective processing may be useful neurotypes for better understanding depression.

[0164] In summary, there is a pressing need to better understand the heterogeneity of depression to deliver better outcomes to depressive patients - and one promising approach is to examine neurotypic differences within depression using ERP -based measures of brain function. ERP measures are scalable, reliable, and have successfully been used in the study of depression (for review Keren et al., 2018; Bruder, Kayser & Tenke, 2011). The promise of ERP -based measures of brain function in the context of precision psychiatry could be greatly enhanced by simultaneously considering multiple ERPs derived from experimental paradigms differing in context. Though the current approach requires replication and extension, it suggests that one promising avenue for neurotyping depression could be to leverage the high within-task and low between-task associations in ERPs to better understand brain function and depression. If successful, this approach could pave the way to further the goals of precision psychiatry — potentially enabling psychiatrists to tailor the treatment for depressed patients based on their neurotype; relatedly, having neurotypic predictors of depression course could be useful in clinical trials to better identify patients likely to remit or respond to placebo.

Claims

CLAIMSWHAT IS CLAIMED IS:

1. A method for analyzing a user’ s behavioral, self-reported, and neural data to derive said user's neurob ehavi oral traits, comprising: a. obtaining data associated with said user from a plurality of sources, wherein said data associated with said user comprises electroencephalography (EEG) data and said user's responses to one or more questionnaires; b. processing said data associated with said user to assess a range of functional brain measures or metrics by generating a model that is specific to said user's neurotype, wherein said neurotype comprises time-domain and time-frequency representations of said user's functional neural activity; and c. using said model to generate one or more predictive insights indicative of said user's neurob ehavi oral traits.

2. The method of claim 1, wherein said EEG data is collected using an EEG headset.

3. The method of claim 1, wherein said EEG data is collected while said user is completing or performing a series of tasks on a device.

4. The method of claim 1, wherein said EEG data comprises event-related potential (ERP) data.

5. The method of claim 3, wherein said series of tasks comprises memory-related tasks, attention-related tasks, speed response tasks, gambling-style tasks, or viewing pictures.

6. The method of claim 1, wherein said functional brain metrics span a range of domains.

7. The method of claim 6, wherein said domains comprise attention, emotional reactivity, working memory, memory, stimulus categorization, decision-making speed, task engagement, reward sensitivity, emotional processing, emotional salience, error processing, or anticipation.

8. The method of claim 1, further comprising: comparing said user’s neurob ehavi oral traits with neurobehavioral traits of one or more other users, based at least in part on measured differences between said user's functional brain metrics and said one or more other users’ functional brain metrics.

9. The method of claim 8, wherein said predictive insights further comprise a comparison of said user’s neurobehavioral traits with said neurobehavioral traits of said one or more other users.

10. The method of claim 1, wherein said questionnaires are designed to elicit responses to a plurality of demographic and individual-specific variables.

11. The method of claim 1, wherein said questionnaires comprise topics relating to anxiety, depression, stress exposure, sleep quality, or occupational activity.

12. The method of claim 1, wherein said data further comprises data collected using one or more wearable devices.

13. The method of claim 12, wherein said data collected using said one or more wearable devices comprises movement data, sleep data, heart rate data, blood oxygen data, or electrocardiogram (ECG) data.

14. The method of claim 12, wherein said one or more wearable devices comprise a smartphone, a sleep monitoring device, a smart watch, a fitness tracker, smart glasses, smart jewelry, or smart clothing.

15. The method of claim 1, wherein said questionnaires are dynamically generated on a predefined schedule.

16. The method of claim 1, wherein said questionnaires are dynamically generated on a randomized schedule.

17. The method of claim 1, wherein said EEG data is collected on a predefined schedule.

18. The method of claim 1, wherein said EEG data is collected on a randomized schedule.

19. The method of claim 1, wherein said EEG data is collected on a schedule that is based at least in part on a sleep quality of the user.

20. The method of claim 1, wherein said questionnaires are dynamically generated on a schedule that is based at least in part on time of year, menstrual cycle schedule, time change, following important world events, on holidays, or after holidays.

21. The method of claim 1, wherein said EEG data is collected on a schedule that is based at least in part on time of year, menstrual cycle schedule, time change, following important world events, on holidays, or after holidays.

22. The method of claim 1, wherein said user's neurotype comprises associations between said user's self-reported measures, behavior, brain function, and one or more biomarkers of interest.

23. The method of claim 1, further comprising generating a time-series graphical representation of said functional brain metrics.

24. The method of claim 1, wherein said data is analyzed using principal component analysis to reduce dimensionality within said questionnaire data, said EEG data, or said wearable data.

25. The method of claim 24, wherein latent profile analysis is performed on factor scores from said principal component analysis.

26. The method of claim 1, wherein said EEG data is filtered.

27. The method of claim 26, wherein said EEG data is corrected for blinks.

28. The method of claim 27 , wherein said EEG data is scored in a time domain.

29. The method of claim 27, wherein said EEG data is scored in a frequency domain.

30. The method of claim 28 or 29, wherein results from said time domain and said frequency domain are averaged together.

31. The method of claim 1, wherein said questionnaires are generated using generative artificial intelligence (Al).

32. The method of claim 1, further comprising: using said model to recommend clinical interventions for said user based on said user’s neurotype or said predictive insights.

33. The method of claim 1, further comprising: using said model to recommend suggestions for said user to improve said user’s well-being.

34. The method of claim 1, further comprising: using said model to predict treatment responses for a health condition or disorder that said user is having or suspected to have.

35. The method of claim 1, wherein said data is pre-processed and analyzed to create data quality metrics, wherein said data quality metrics are measured or based on signal and noise indices, internal reliability, or comparisons between odd and even halves of collected data.

36. The method of claim 35, wherein said data quality metrics are compared to previous data collected to create internal norms.

37. The method of claim 36, wherein said data quality metrics are used to determine when sufficient data has been collected, wherein a metric used to determine whether sufficient data has been collected is whether a difference between odd and even trials on a given neural metric falls within one standard deviation of said internal norms.

38. The method of claim 1, wherein said data comprises time series data, structured data, or unstructured data.

39. The method of claim 1, wherein said EEG data is collected in a single session.

40. The method of claim 1, wherein said EEG data is collected over multiple sessions.

41. The method of claim 4, wherein said ERP data is collected in a single session.

42. The method of claim 4, wherein said ERP data is collected over multiple sessions.

43. The method of claim 39, wherein said EEG data collected in a single session is used to create a neuropsychiatric profile for a subject.

44. The method of claim 41, wherein said ERP data collected in a single session is used to create a neuropsychiatric profile for a subject.

45. The method of claim 40, wherein said EEG data collected over multiple sessions is used to build a within-subject model for a subject.

46. The method of claim 42, wherein said ERP data collected over multiple sessions is used to build a within-subject model for a subject.

Images