Bio-signal local collection, speech aided interface cursor control based on bio-electric signals, and arousal detection based on bio-electric signals

Abstract

The present disclosure relates to local collection of biosignals, cursor control in speech-assisted interfaces based on bioelectric signals, and arousal detection based on bioelectric signals. The present disclosure provides a device with electrodes configured to record electrical activity constrained to a limited area, use the recorded bioelectric signals to control cursor position in a speech-assisted interface, and use the recorded bioelectric signals to detect arousal during sleep.

Description

[0001] This application is a divisional application of the original application with the filing date of October 14, 2014, application number 201480066976.7, and the title "Local Collection of Bio-signals, Cursor Control in a Speech Assistive Interface Based on Bio-electric Signals, and Wakefulness Detection Based on Bio-electric Signals."

[0002] Cross Reference to Related Applications

[0003] This application claims the benefit of priority under 35 U.S.C. § 119(e) to U.S. Patent Application Serial No. 61 / 890,859, filed October 14, 2013, the entire contents of which are incorporated herein by reference. TECHNICAL FIELD

[0004] The present disclosure relates generally to medical devices, and more particularly to devices with electrodes configured to record electrical activity constrained to a limited area, use recorded bio-electric signals to control cursor position in a speech assistive interface, and use recorded bio-signals to detect wakefulness during sleep. BACKGROUND

[0005] In humans, various neurons in the brain cooperate to generate a rich and continuous set of neural electrical signals. Such signals have a strong influence on the control of the rest of our body. For example, the signals cause the body to move and facilitate cognitive thinking. In addition, neural signals can cause a human to wake up during sleep. Despite decades of intensive research, a direct translation from the signals to various human actions remains unknown due to the complexity of the signals. However, the utility of understanding such mapping provides the possibility of greatly improving the lives of many individuals with life function impairments. Thereafter, the understanding would likely allow conditions to be diagnosed or particular signal- action biological pathways to be hacked and / or replicated by technology. SUMMARY

[0006] For many years, various devices have been used to record neural activity. One such device includes an electroencephalogram (EEG) device. Traditionally, tens of electrodes are placed all around a person's head. A large number of electrodes are precisely placed on the scalp locations in an attempt to improve the signal-to-noise ratio. Even though many electrodes are used, many people still claim that EEG has poor spatial resolution and low signal-to-noise ratio. Thus, the application of EEG data is limited for at least two reasons: the number of electrodes and placement precision generally limit EEG recordings to clinical situations, and previous efforts to extract meaningful neural underpinnings in the data have constrained the interpretation and use of the data.

[0007] Certain embodiments of the present application can utilize discovered techniques to identify neural signatures within EEG data that have previously been discarded as too noisy for important interpretation. For example, in some embodiments, a single small device can house multiple EEG electrodes, including active electrodes, reference electrodes, and (optionally) ground electrodes. The device can have a footprint having a length and width of less than 6 (or even 4) inches, and the separation distance between any pair of electrodes can be less than 3 inches. The close proximity of active and reference electrodes is traditionally avoided because it is believed to introduce distortion into the recording. Moreover, because EEG analysis typically differentially amplifies signals from two electrodes, placing a reference electrode in a location that will itself be recorded as neural activity is believed to suppress detection of the neural signal of interest (typically in the high frequency band). Thus, reference electrodes have traditionally been placed away from active electrodes and at neural locations having relatively low or no neural activity. However, as described herein, processing of data from clustered electrodes can still extract signals of physiological interest.

[0008] The signals recorded using the electrodes can be collectively analyzed (e.g., at the device) to generate a single channel of neural recording. This channel can then be analyzed, for example, to identify absolute or relative amounts of sleep time in various sleep stages, to assess the number and type of potential sleep disorders, and / or to identify sleep abnormalities.

[0009] In one example, the spectrogram of the recorded signal is normalized one or more times across time bins and / or across frequencies. For example, in one example, the spectrogram can be normalized across time bins. In another example, the spectrogram is normalized across time bins and then across frequencies. In yet another example, an alternating pattern of time bin and frequency normalization can continue for a given number of normalizations, or until the normalization factor falls below a threshold. Normalization across time bins can include using all power of each frequency in the spectrogram to calculate a z-score for that frequency in the spectrogram. The power of that frequency can be normalized by the z-score. Normalization across frequencies can include using all power of each time bin in the spectrogram to calculate a z-score for that time bin in the spectrogram. The power of that time bin can be normalized by the z-score.

[0010] In some examples, for each time bin in the normalized spectrogram, a "strong frequency" of the time bin can be defined as the frequency associated with the high (e.g., above an absolute or relative threshold) or highest normalized power of the time bin. Thus, a time series strong frequency function can be determined. The distribution of strong frequencies can vary across sleep stages, such that identifying strong frequencies can support estimation of the associated sleep stage.

[0011] Further, at each time point, a fragmentation value can be defined. The fragmentation value can include a temporal fragmentation value or a spectral fragmentation value. For the temporal fragmentation value, a temporal gradient of the spectrogram can be determined. The spectrogram can include the original spectrogram and / or a spectrogram that has been normalized 1, 2, or more times across time bins and / or across frequency (e.g., a spectrogram that is first normalized across time bins and then across frequency). Thus, each time bin can be associated with a vector of partial derivative power values (across a set of frequencies). For a given time block or epoch (comprising multiple time bins), a frequency-specific variable can be determined for each frequency using the gradient values within the time block and corresponding to the given frequency. For example, the frequency-specific variable can include a mean of the absolute values of the gradient values corresponding to the given frequency. The temporal fragmentation value can then be defined as the frequency or epoch corresponding to the high or highest frequency-specific variable. Thus, the temporal fragmentation value can identify a frequency with high modulation. The spectral fragmentation value can be similarly defined, but can be based on a spectral gradient of the spectrogram. A high fragmentation value can be indicative of sleep stage disruption or changes in wakeful activity.

[0012] Analysis of the channel data can occur (in whole or in part) at the device or at a remote device. For example, the channel data (or signals that generate the channel data) can be transmitted (e.g., wirelessly) to other resources for more in-depth processing and / or storage. It should be understood that the device can also collect, transmit, and / or analyze non-EEG data. The device can also include one or more other external sensors, such as an accelerometer to provide additional data indicative of the context of the recording (e.g., to allow differentiation between stationary and active states), or a thermometer to estimate the temperature of the user.

[0013] The device can be positioned on a person by adhering an adhesive material to the device and to the individual. For example, an adhesive material (e.g., a double-sided adhesive film or substance) can be applied to at least a portion of the underside of the device such that it can attach the device to a skin location. As another example, an adhesive film can be positioned over the device, and a portion of the film extending beyond the device can be attached to a skin location.

[0014] Accordingly, the devices and techniques described herein enable EEG to be readily collected. A single device can independently provide data for an entire channel, and both the required head placement and the number of necessary placements can be relatively low. Thus, a patient can self-apply the device and initiate an EEG recording. Wireless transmission from the device further reduces the complexity of starting data collection. It should be understood that while a multi-electrode device can independently support a channel, multiple devices (in some cases) can be used to further enrich a recording by collecting multiple channels.

[0015] The embodiments herein can extend beyond the collection, analysis, and application of neural signals: the devices can be used to collect any bioelectric signal. For example, the devices can be positioned on a muscle and can collect electromyography (EMG) data. The EMG data can be used, for example, for biofeedback training (e.g., by providing a patient with a cue indicating when a muscle is activated), to aid in the diagnosis of neurological or myopathic diseases, and / or to translate muscle movement into control of an external object (e.g., a cursor on a screen of an electronic device or control of a prosthesis). In one exemplary embodiment, one or more devices can be used to allow a person afflicted with amyotrophic lateral sclerosis (ALS) to communicate, even if vocalization and hand control are limited. Specifically, one or more devices can be positioned on a single or multiple muscles, such as the masseter muscle, that the patient can still control. At the same time, the patient can be provided with a screen having multiple text options (e.g., individual letters, combinations of letters, words, or phrases). Analysis of the recordings from the masseter muscle can cause the cursor to move to the desired text option. Repetition of such selections can allow a sentence to be formed, which can be used for written communication or can be read out by an automated reader.

[0016] EMG recordings can be mapped to cursor movement. In one instance, the mapping can be determined based on, for example, analysis of raw EMG data from one or more channels (for a training or non-training situation) using clustering and / or component analysis to determine which signal signatures are associated with a particular cursor movement. In one instance, a particular strong frequency is associated with a cursor movement, such that, for example, muscle data dominated by a strong frequency in a high frequency band can be determined to correspond to an upward cursor movement. In another instance, a particular fragmentation value can be the associated cursor movement. For example, a high fragmentation value associated with EMG of one muscle can be associated with a first cursor movement, while a high fragmentation value associated with EMG of another muscle can be associated with a second cursor movement.

[0017] The sensitivity and non-invasiveness of the devices and techniques can also be used to assess physiological events that can be difficult for a patient or medical professional to detect in other ways. For example, the devices can record signals during sleep, and arousal (which can include micro-arousal) can be detected. In a binary case, a basic arousal can be defined as a transition from a sleep state to a wake state. However, this binary characterization of the states oversimplifies the complexity of sleep. Sleep is characterized using sleep stages: stages 1-4 and rapid eye movement (REM) stage. Little is known about the manner and timing of transitions between sleep stages, although the time spent in each sleep stage can be physiologically important. For example, insufficient REM sleep can impair learning ability, and stage 4 sleep is important for growth and development.

[0018] Accordingly, if a patient reports poor sleep at night or other sleep-related symptoms, it can be useful to monitor the patient's sleep in various stages. According to some embodiments, neural recordings can be recorded from a compact electrode device and analyzed to extract, amplify high frequency neural signals. The signals can then be classified into sleep (or wake) stages within various brief time windows. Awakenings can then be detected by quantifying the variability and / or stage transitions present within a series of time windows. The ability to classify sleep into such short time windows enables the detection of awakenings that would otherwise go unrecognized. Such awakenings can be used to assess sleep quality.

[0019] The sleep analysis can also be used to detect whether a person is experiencing a potentially life-threatening event in their sleep. For example, tracheostomy mechanical ventilation can be performed in a selected patient population (e.g., ALS, facial trauma patients with cancer) who are deemed to benefit from the procedure. The procedure can include inserting a tracheostomy tube into an incision in the neck. Unfortunately, it is possible for the tube to slide out of the tracheostomy. This possibility can be especially dire for patients who are impaired in their ability to communicate. Should their tube slide out at night, the patient can have difficulty alerting anyone to the problem. However, the devices and methods disclosed herein can monitor the sleep stages of these patients and detect abnormalities and / or sleep stage patterns of interest. The size of the device can improve compliance with use and monitoring, and the analysis can aid in detecting rapid sleep stage patterns.

[0020] The following DETAILED DESCRIPTION, together with the drawings, will provide a better understanding of the nature and advantages of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 A user is shown wearing a multi-electrode compact device in wireless communication with another electronic device.

[0022] Figure 2 An example of a device according to an embodiment of the application is shown connected on a network to facilitate collaborative assessment and use of bioelectric recordings.

[0023] Figure 3 A multi-electrode device in wireless communication with another electronic device according to an embodiment of the application is shown.

[0024] Figure 4 Is a simplified block diagram of a multi-electrode device according to an embodiment of the application.

[0025] Figure 5 Is a simplified block diagram of an electronic device in communication with a multi-electrode device according to an embodiment of the application.

[0026] Figure 6is a flowchart of a process according to embodiments of the application for analyzing channel bio data to identify frequency signatures of individual bio stages.

[0027] Figure 7 is a flowchart of a process according to embodiments of the application for analyzing channel bio data to identify frequency signatures of individual bio stages.

[0028] Figure 8 is a flowchart of a process according to embodiments of the application for analyzing channel bio data to identify frequency signatures of individual bio stages.

[0029] Figure 9 is a flowchart of a process according to embodiments of the application for normalizing a spectrogram and using discriminative group-wise frequency signatures to classify bio data.

[0030] Figure 10 is a flowchart of a process according to embodiments of the application for analyzing channel bio data to identify wakefulness.

[0031] Figures 11-14 An example of automated wakefulness detection is shown.

[0032] Figure 15 is a flowchart of a process according to embodiments of the application for normalizing a spectrogram and identifying frequencies to classify bio data.

[0033] Figure 16 is a flowchart of a process according to embodiments of the application for normalizing a spectrogram and using gradients to identify frequencies to classify bio data.

[0034] Figure 17 is a flowchart of a process according to embodiments of the application for determining a mapping of EMG data using reference data.

[0035] Figure 18A and 18B An example of communication-assisted visualization is shown.

[0036] Figure 19 is a flowchart of a process according to embodiments of the application for generating written or spoken text based on EMG data.

[0037] Figure 20 Raw and normalized spectrograms of sleep EEG data are shown.

[0038] Figure 21 Time series preferred frequency plots determined using raw or normalized spectrograms are shown. DETAILED DESCRIPTION

[0039] Certain embodiments of the present application can facilitate the use of a compact multi-electrode device to conveniently record biological signals (e.g., electroencephalogram (EEG) or electromyogram (EMG) data). Spectrograms can be generated based on differences in the recorded data and normalized in one or both directions of the spectrogram (e.g., such that each power value is normalized based on power values having the same frequency but for different time bins and / or based on power values having the same time bin but for different frequencies). The spectrograms can be divided into time bins or epochs (e.g., having a defined duration, such as 30 seconds), and each spectrogram portion can be normalized one or more times (e.g., across frequencies or across time bins in time). For a given time bin, a z-score can be determined using the normalized power values (e.g., such that the z-score is higher for frequencies corresponding to normalized power values in time bins across the time bin that are large differences). Strong frequencies can then be identified for the time bin as frequencies corresponding to high or the highest normalized power. The strong frequencies can be indicative of a sleep stage.

[0040] In addition, for each time bin, a fragmentation value can be defined. For example, a gradient (e.g., a temporal gradient) of the spectrogram (un-normalized, normalized, twice normalized, etc.) can be determined. For a given time bin, a fragmentation value can be defined to identify frequencies corresponding to high modulation across the associated power. For example, the fragmentation value can include a mean absolute or a relatively higher frequency of absolute values of the gradient values (across time bins in the time bin). When the device records neural data during sleep, a high fragmentation value can be indicative of inconsistent sleep characteristics, which can suggest sleep disorders and / or wakefulness.

[0041] The technique can be effectively applied to data with short time binning. Thus, it can identify even very short wakefulness. Wakefulness can be indicative of poor sleep quality and / or health factors of interest. Thus, the technique has the potential to be used to detect potentially relevant data that would otherwise be ignored due to larger time binning or an inability to collect a significant amount of data.

[0042] The multi-electrode device can also be used to collect EMG data from one or more muscles. Clustering and / or component techniques can be used to map features of the data to specific subject actions. Thus, for example, a contraction of one muscle can be indicative of the cursor being moved up, while a contraction of another muscle can be indicative of the cursor being moved down. Thus, a screen can be provided for a patient with limited vocalization capabilities that allows the patient to move the cursor in various directions to select among letters, words, phrases, or requests to convey thoughts.

[0043] Figure 1A user 105 is shown using a multi-electrode device 110. The device is shown adhered to the user's forehead 115 (e.g., via an adhesive positioned between the device and the user). The device can include multiple electrodes to detect and record neural signals. After signal recording, the device can transmit (e.g., wirelessly) data (or a processed form thereof) to another electronic device 120, such as a smartphone. The other electronic device 120 can then further process the data and / or respond to the data, as further described herein. Thus, Figure 1 The multi-electrode device 105 is illustrated as being small and simple to position. While only one device is shown in this example, it should be understood that multiple devices are used in some embodiments.

[0044] Further, while Figure 1 An adhesive is shown attaching the device 110 to the user 105, but other attachment means can be used. For example, a head harness or headband can be positioned around the user and the device. Additionally, while it is often advantageous for ease of use to house all electrodes of a channel in a single compact unit, it should be understood that in other examples, electrodes can be external to the main device housing and can be positioned far from one another. In one example, a device as described in PCT Application PCT / US2010 / 054346 is used. PCT / US2010 / 054346 is incorporated herein by reference for all purposes.

[0045] The devices 115a and 115b can communicate directly (e.g., through a Bluetooth connection or a BTLE connection) or indirectly. For example, each device can communicate with a server 120 (e.g., through a Bluetooth connection or a BTLE connection), which can be located near the tennis court 110.

[0046] Figure 2 An example of devices connected on a network to facilitate the coordinated assessment and use of bioelectric recordings is shown. One or more multi-electrode devices 205 can collect channel data derived from recorded bio data from a user. The data can then be provided to one or more other electronic devices, such as a mobile device 210a (e.g., a smartphone), a tablet 210b, or a laptop or desktop computer 201c. Inter-device communication can be through a connection such as a short-range connection 215 (e.g., a Bluetooth, BTLE, or ultra-wideband connection) or through a WiFi network 220 such as the Internet.

[0047] One or more of the devices 205 and / or 210 can also access a data management system 225, which can, for example, receive and evaluate data from a range of multi- electrode devices. For example, a health care provider or pharmaceutical company (e.g., conducting a clinical trial) can use data from a multi-electrode device to gauge a patient's health. Thus, for example, the data management system 225 can store data associated with a particular user and / or can generate population statistics.

[0048] Figure 3 A multi-electrode device 300 is shown in communication (e.g., wirelessly or via a cable) with another electronic device 302. This communication can be performed to enhance the functionality of the multi-electrode device by utilizing the resources of the other electronic device (e.g., faster processing speed, greater memory, display screen, input receiving capability). In one example, the electronic device 302 includes interface capabilities that allow a user (which can or can not be the same person whose signals are being recorded) to view information (e.g., a summary of recorded data and / or operational options) and / or control operations (e.g., control the functionality of the multi-electrode device 300 or control another operation, such as speech construction). The communication between the devices 300 and 302 can occur intermittently as the device 300 collects and / or processes data or after a data collection period. Data can be pushed from the device 300 to the other device 302 and / or pulled from the other device 302.

[0049] Figure 4 is a simplified block diagram of a multi-electrode device 400 (e.g., implementing the multi-electrode device 300) according to an embodiment of the present application, which can include a processing subsystem 402, a storage subsystem 404, an RF interface 408, a connector interface 410, a power subsystem 412, an environmental sensor 414, and electrodes 416. The multi-electrode device 400 need not necessarily include every illustrated component, and / or can also include other components (not explicitly shown).

[0050] The storage subsystem 404 can be implemented, for example, using magnetic storage media, flash memory, other semiconductor memory (e.g., DRAM, SRAM), or any other non-transitory storage medium, or combination of media, and can include volatile and / or non-volatile media. In some embodiments, the storage subsystem 404 can store biological data, information about a user (e.g., identifying information and / or medical history information), and / or analysis variables (e.g., strong frequencies determined previously or frequencies used to distinguish between signal groups). In some embodiments, the storage subsystem 404 can also store one or more applications (or apps) 434 to be executed by the processing subsystem 410 (e.g., to initiate and / or control data collection, data analysis, and / or transmission).

[0051] The processing subsystem 402 can be implemented as one or more integrated circuits, e.g., one or more single-core or multi-core microprocessors or microcontrollers, examples of which are known in the art. In operation, the processing system 402 can control the operation of the multi-electrode device 400. In various embodiments, the processing subsystem 404 can execute a variety of programs in response to program code and can maintain multiple simultaneously executing programs or processes. At any given time, some or all of the program code to be executed can be resident in the processing subsystem 404 and / or in storage such as the storage subsystem 404.

[0052] With appropriate programming, the processing subsystem 402 can provide a variety of functionality for the multi-electrode device 400. For example, in some embodiments, the processing subsystem 402 can execute code that can control the collection, analysis, application, and / or transmission of biological data. In some embodiments, some or all of the code can interact with other devices 302 in the system 300, e.g., by generating messages to be sent to the interface device and / or by receiving and interpreting messages from the interface device. In some embodiments, some or all of the code can operate locally to the multi-electrode device 400. Figure 3

[0053] The processing subsystem 402 can also execute data collection code 436 that can cause data detected by the electrodes 416 to be recorded and saved. In some instances, the signals are differentially amplified, and filtering can be applied. The signals can be stored in the biological data datastore 437 along with recording details, e.g., recording time and / or user identifier. The data can be further analyzed to detect physiological correspondences. As one example, processing of a spectrogram of the recorded signals can reflect frequency properties that correspond to particular sleep stages. As another example, wakefulness detection code 438 can analyze the gradient of the spectrogram to identify and evaluate sleep disorder indicators and detect wakefulness. As yet another example, signal actuator code 439 can translate particular biological signal features into motion of an external object, e.g., a cursor. Such techniques and code are further described herein.

[0054] The RF (radio frequency) interface 408 can allow the multi-electrode device 400 to communicate wirelessly with various interface devices. The RF interface 408 can include RF transceiver components such as antennas and supporting circuitry to enable, e.g., Wi-Fi (IEEE 802.11 family of standards), Bluetooth®, ZigBee®, and / or other wireless communication protocols. ​Bluetooth® (a set of standards promulgated by Bluetooth SIG, Inc.) or other protocols for wireless data communication to enable data communication over a wireless medium. In some embodiments, the RF interface 408 can enable short-range sensors (e.g., Bluetooth, BLTE, or ultra-wideband), proximity sensors 409 that support detection of proximity by estimation of signal strength and / or other protocols for determining proximity to another electronic device. In some embodiments, the RF interface 408 can provide near-field communication (“NFC”) capabilities, e.g., implementing the ISO / IEC 18092 standard; NFC can support wireless exchange of data between devices within a very short range (e.g., 20 centimeters or less). The RF interface 408 can be implemented using a combination of hardware (e.g., driver circuitry, antennas, modulators / demodulators, encoders / decoders, and other analog and / or digital signal processing circuitry) and software components. Multiple different wireless communication protocols and associated hardware can be incorporated into the RF interface 408.

[0055] The connector interface 410 can allow the multi-electrode device 400 to communicate with various interface devices via a wired communication path, e.g., using Universal Serial Bus (USB), Universal Asynchronous Receiver / Transmitter (UART), or other protocols for wired data communication. In some embodiments, the connector interface 410 can provide a power port to allow the multi-electrode device 400 to receive power, e.g., to charge an internal battery. For example, the connector interface 410 can include a connector such as a mini-USB connector or a custom connector and supporting circuitry. In some embodiments, the connector can be a custom connector that provides dedicated power and ground contacts, as well as digital data contacts (which can be used to implement different communication technologies in parallel); for example, two pins can be assigned as USB data pins (D+ and D-), and two other pins can be assigned as serial transmit / receive pins (e.g., implementing a UART interface). The assignment of pins to a particular communication technology can be hardwired or negotiated while the connection is established. In some embodiments, the connector can also provide connections to transmit and / or receive bioelectric signals, which can be transmitted to or from another device (e.g., the device 302 or another multi-electrode device) in analog and / or digital format.

[0056] The environmental sensors 414 can include various electronic, mechanical, electromechanical, optical, or other devices that provide information related to external conditions surrounding the multi-electrode device 400. In some embodiments, the sensors 414 can provide digital signals to the processing subsystem 402, e.g., on a streaming basis as needed or in response to polling by the processing subsystem 402. Any type and combination of environmental sensors can be used; shown by way of example as an accelerometer 442. Acceleration sensed by the accelerometer 442 can be used to estimate whether a user is sleeping or attempting to sleep and / or to estimate an activity state.

[0057] The electrodes 416 can include, e.g., round surface electrodes, and can include gold, tin, silver, and / or silver / silver chloride. The electrodes 416 can have a diameter greater than 1 / 8" and less than 1". The electrodes 416 can include active electrodes 450, reference electrodes 452, and (optionally) ground electrodes 454. The electrodes can or can not be distinguishable from one another. Electrode locations can be fixed within the device and / or movable (e.g., tethered to the device).

[0058] The power subsystem 412 can provide power and power management capabilities for the multi-electrode device 400. For example, the power subsystem 414 can include a battery 440 (e.g., a rechargeable battery) and associated circuitry to distribute power from the battery 440 to other components of the multi-electrode device 400 that require electrical power. In some embodiments, the power subsystem 412 can also include circuitry operable to charge the battery 440, e.g., when the connector interface 410 is connected to a power source. In some embodiments, the power subsystem 412 can include a "wireless" charger, e.g., an inductive charger, to charge the battery 440 without relying on the connector interface 410. In some embodiments, the power subsystem 412 can include other power sources, e.g., solar cells, in addition to or instead of the battery 440.

[0059] It will be appreciated that the multi-electrode device 400 is exemplary, and variations and modifications are possible. For example, the multi-electrode device 400 can include a user interface to enable a user to directly interact with the device. As another example, the multi-electrode device can have an attachment indicator that indicates (e.g., via light, color, or sound) whether contact between the device and the user's skin is adequate and / or whether recorded signals have an acceptable quality.

[0060] Furthermore, although multi-electrode devices are described with reference to specific blocks, it should be understood that these blocks are defined for ease of description and are not intended to suggest a specific physical arrangement of component parts. Moreover, blocks do not necessarily correspond to physically different components. Blocks can be configured to perform various operations, for example, by programming a processor or providing suitable control circuitry, and depending on how the initial configuration is obtained, various blocks may or may not be reconfigurable. Embodiments of the invention can be implemented in a variety of devices, including electronic devices implemented using any combination of circuitry and software. It is also not necessary to... Figure 4 Each box in the diagram is implemented in a given implementation of the multi-electrode device.

[0061] Such as Figure 3 The interface device of device 302 can be implemented as an electronic device using a frame (e.g., processor, storage medium, RF interface, etc.) and / or other frames or components similar to those described above. Figure 5 It is an interface device 500 according to an embodiment of the present invention (e.g., implementing...). Figure 3 A simplified block diagram of the device 302. The interface device 500 may include a processing subsystem 502, a storage subsystem 504, a user interface 506, an RF interface 508, a connector interface 510, and a power subsystem 512. The interface device 500 may also include other components (not explicitly shown). Many components of the interface device 500 may be similar to or equivalent to... Figure 3 Components of the multi-electrode device 300.

[0062] For example, storage subsystem 504 may be substantially similar to storage subsystem 404 and may include, for example, magnetic storage media, flash memory, other semiconductor memories (e.g., DRAM, SRAM), or any other non-transitory storage media, or combinations thereof, and may include volatile and / or non-volatile media. Similar to storage subsystem 504, storage subsystem 504 may be used to store data and / or program code to be executed by processing subsystem 502.

[0063] User interface 506 may include any combination of input and output devices. A user can operate the input devices of user interface 506 to invoke functions of interface device 500, and can view, hear, and / or otherwise experience output from interface device 500 via the output devices of user interface 506. Examples of output devices include display 520 and speaker 522. Examples of input devices include microphone 526 and touch sensor 528.

[0064] Display 520 can be implemented using compact display technologies such as LCD (liquid crystal display), LED (light emitting diode), OLED (organic light emitting diode), etc. In some embodiments, display 520 can incorporate a flexible display element or a curved glass display element to allow interface device 500 to conform to a desired shape. One or more speakers 522 can be provided using small form factor speaker technologies, including any technology capable of converting an electronic signal into an audible sound wave. Speakers 522 can be used to produce tones (e.g., a beep or chime) and / or speech.

[0065] Examples of input devices include microphone 526 and touch sensor 528. Microphone 526 can include any device that converts sound waves into electronic signals. In some embodiments, microphone 526 can be sensitive enough to provide a representation of specific words spoken by a user; in other embodiments, microphone 426 can be used to provide an indication of the general level of ambient sound without necessarily providing a high-quality electronic representation of specific sounds.

[0066] Touch sensor 528 can include, for example, a capacitive sensor array capable of locating a contact to a specific point or area on the surface of the sensor, and in some instances capable of distinguishing multiple simultaneous contacts. In some embodiments, touch sensor 428 can be overlaid on display 520 to provide a touch screen interface, and processing subsystem 504 can translate touch events into specific user input depending on what is currently displayed on display 520.

[0067] Processing subsystem 502 can be implemented as one or more integrated circuits, such as one or more single-core or multi-core microprocessors or microcontrollers, examples of which are known in the art. In operation, processing subsystem 502 can control the operation of interface device 500. In various embodiments, processing subsystem 502 can execute a variety of programs in response to program code and can maintain multiple simultaneously executing programs or processes. At any given time, some or all of the program code to be executed can be resident in processing subsystem 502 and / or in storage such as storage subsystem 504.

[0068] With appropriate programming, processing subsystem 502 can provide various functionality for interface device 500. For example, in some embodiments, processing subsystem 502 can execute an operating system (OS) 532 and various applications 534. In some embodiments, some or all of these applications can interact with a multi-electrode device, for example by generating messages to be sent to a multi-electrode device and / or by receiving and interpreting messages from a multi-electrode device. In some embodiments, some or all of the applications can operate locally at interface device 500.

[0069] The processing subsystem 502 can also execute data collection code 536 (which, as desired, can be part of the OS 532, part of an app, or separate). The data collection code 536 can be at least partially complementary to the data collection code 436 in the multi-electrode device 300. In some instances, the data collection code 536 is configured such that execution of the code causes the device 500 to receive raw or processed bioelectrical signals (e.g., EEG or EMG signals) from a multi-electrode device (e.g., the multi-electrode device 300). Figure 4 Figure 3 In some instances, execution of the data collection code 536 also causes the device 500 to collect data, which can include bio data (e.g., temperature or pulse of a patient) or external data (e.g., light level or geographic location). This information can be stored with the bioelectrical data (e.g., such that metadata for an EEG or EMG recording includes the temperature and / or location of the patient), and / or can be stored separately (e.g., with a timestamp to allow future time-synchronous matching). It will be appreciated that in these instances, the interface device 500 either can include suitable sensors to collect this additional data (e.g., a video camera, a thermometer, a GPS receiver), or can communicate with another device that has such sensors (e.g., via the RF interface 508).

[0070] In some instances, execution of the data collection code 536 also causes the device 500 to collect data, which can include bio data (e.g., temperature or pulse of a patient) or external data (e.g., light level or geographic location). This information can be stored with the bioelectrical data (e.g., such that metadata for an EEG or EMG recording includes the temperature and / or location of the patient), and / or can be stored separately (e.g., with a timestamp to allow future time-synchronous matching). It will be appreciated that in these instances, the interface device 500 either can include suitable sensors to collect this additional data (e.g., a video camera, a thermometer, a GPS receiver), or can communicate with another device that has such sensors (e.g., via the RF interface 508).

[0071] The processing subsystem 502 can also execute one or more codes that analyze raw or processed bioelectrical signals in real-time or retrospectively to detect events of interest. For example, execution of wakefulness detection code 538 can evaluate changes in a spectrogram (constructed using EEG data) corresponding to a sleep session of a patient to determine whether and / or when a wakefulness occurred. In one instance, the evaluation can include determining, for each time increment, a change variable that corresponds to a change in power (e.g., normalized power) at one or more frequencies for that time increment relative to one or more other time increments. In one instance, the evaluation can include assigning each time increment to a sleep stage and detecting points in time at which assignments change. The sleep stage classification can, in some instances, further detail any wakefulness that is occurring (e.g., by indicating a stage in which the wakefulness occurred and / or by identifying how many sleep stages the wakefulness traversed).

[0072] ​As another example, execution of signal actuator code 539 can evaluate and translate EMG data. Initially, a mapping can be constructed to associate particular EMG signatures with particular actions. The actions can be external actions, e.g., actions of a cursor on a screen. The mapping can be performed using cluster analysis and / or component analysis, and can utilize raw or processed signals recorded from one or more active electrodes, e.g., from one or more multi-electrode devices, each device positioned on a different muscle.

[0073] In one example, execution of signal actuator code 539 causes an interactive visualization to be presented on display 520. The position of a cursor on the screen can be controlled using the mapping based on real-time analysis of EMG data. A person from whom the recordings are collected can thus interact with the interface without using his hands. In one illustrative example, the visualization can include a speech-assisted visualization that allows the person to select letters, sequences of letters, words, or phrases. Sequential selection can allow the person to construct sentences, paragraphs, or conversations. The text can be used electronically, e.g., to generate an email or letter, or can be vocalized, e.g., using a speech component of signal actuator 539 to send audio output to speaker 522, to communicate with other people nearby.

[0074] RF (radio frequency) interface 508 and / or connector interface 510 can allow interface device 500 to wirelessly communicate with various other devices, e.g., multi-electrode devices 400, and networks. RF interface 508 can correspond to RF interface 408 (e.g., including the features described thereof), and / or connector interface 510 can correspond to connector interface 410 (e.g., including the features described thereof). Power subsystem 512 can provide power and power management capabilities for interface device 512. Power subsystem 512 can correspond to power subsystem 41 (e.g., including the features described thereof). Figure 4 Figure 4 It should be appreciated that interface device 500 is exemplary, and variations and modifications are possible. In various embodiments, other controllers or components can be provided in addition to or instead of those described above. Any device that is capable of interacting with another device, e.g., a multi-electrode device, to store, process, and / or use recorded bioelectrical signals can be an interface device.

[0075] It should be appreciated that interface device 500 is exemplary, and variations and modifications are possible. In various embodiments, other controllers or components can be provided in addition to or instead of those described above. Any device that is capable of interacting with another device, e.g., a multi-electrode device, to store, process, and / or use recorded bioelectrical signals can be an interface device.

[0076] ​Furthermore, although the interface device is described with reference to particular blocks, it is to be understood that such blocks are defined for convenience to aid in understanding the present disclosure and are not intended to confine the component parts of the application to particular physical arrangements. Further, the blocks are not necessarily to be understood as referring to physical arrangements of components. The blocks can be configured to perform various operations, e.g., by programming processors or providing suitable control circuitry, and, depending on the implementation, various blocks might or might not be reconfigurable. Embodiments of the present application can be implemented in a variety of devices including electronic devices using any combination of hardware, firmware, and software. All such permutations and combinations are intended to fall within the scope of the present disclosure. Figure 5 Each of the blocks in process 600 can be implemented in a given embodiment of the interface device.

[0077] Communication between the one or more multi-electrode devices, the one or more mobile devices, and the interface device can be implemented according to any communication protocol (or combination of protocols) that the two devices are programmed or otherwise configured to use. In some examples, a standard protocol such as a Bluetooth scheme or an Ultra-Wideband protocol can be used. In some examples, custom message formats and syntax (including, e.g., a set of rules for interpreting a particular byte or sequence of bytes in a digital data transmission) can be defined, and messages can be transmitted using a standard serial protocol such as a virtual serial port defined in certain Bluetooth standards. Embodiments of the present application are not limited to a particular protocol, and one skilled in the art with the benefit of this teaching will recognize that many protocols can be used.

[0078] According to certain embodiments of the present application, the one or more multi-electrode devices can be used to advantage to collect electrophysiological data from a patient. This data can be processed to identify signals of physiological interest. Detection itself can be useful because it can let a user or third party know about the health of the patient and / or the efficacy of a current treatment. In some examples, the signals can be used to automatically control another object, such as a computer cursor. This capability can extend the physical capabilities of a user (e.g., who can be disabled due to illness) and / or improve operational convenience.

[0079] Figure 6 is a flowchart of a process 600 according to an embodiment of the present application for collecting channels of bioelectrical data using a multi-electrode device. Part or all of process 600 can be implemented in a multi-electrode device (e.g., multi-electrode device 400). In some examples, part of process 600 (e.g., one or more of blocks 610-635) can be implemented in an electronic device remote from the multi-electrode device, where the blocks can be performed immediately after receiving signals from the multi-electrode device (e.g., immediately after collection), before storing data about the recording, in response to a request that relies on the collected data, and / or before using the collected data.

[0080] At block 605, active and reference signals can be collected using respective electrodes. In some examples, a ground signal is further collected from a ground electrode. The active and reference electrodes and / or the active and ground electrodes can be attached to a single device (e.g., a multi-electrode device), fixed distances from each other and / or close to each other (e.g., such that centers of the electrodes are positioned less than 12, 6, or 4 inches apart from each other and / or such that the electrodes are positioned to likely record signals from the same muscle or the same brain region).

[0081] In some examples, the reference electrode is positioned near the active electrode such that both electrodes will likely sense electrical activity from the same brain region or from the same muscle. In other examples, the reference electrode is positioned further away from the active electrode (e.g., in a region that is relatively electrically neutral, which can include a region not on the brain or a protruding muscle) to reduce overlap of signals of interest.

[0082] Prior to collection, the electrodes can be attached to the skin of a person. This can include, for example, attaching a single device that fully houses one or more electrodes and / or attaching one or more separate electrodes (e.g., flexibly extending beyond between device housings). In one example, such attachment is performed using an adhesive (e.g., applying an adhesive substance to at least a portion of an underside of the device, applying an adhesive patch over and around the device, and / or applying a double-sided adhesive patch under at least a portion of the device) to attach a multi-electrode device including an active electrode and a reference electrode to a person. For EEG recording, the device can be attached near the frontal lobe of the person (e.g., on her forehead), for example. For EMG recording, the device can be attached on a muscle (e.g., on a masseter muscle or a neck muscle).

[0083] In some examples, only one active signal is recorded at a time. In other examples, each of a set of active electrodes records an active signal. In such cases, the active electrodes can be positioned at different body locations (e.g., on different sides of the body, on different muscle types, or on different brain regions). Each active electrode can be associated with a reference electrode, or fewer reference values relative to the number of active signals collected can be collected. Each active electrode can exist in a separate multi-electrode device.

[0084] At block 610, the reference signal can be subtracted from the active electrode. This can reduce noise in the active signal, e.g., recording noise or noise caused by patient respiration or movement. Although close positioning of the reference and active electrodes has traditionally been eschewed, such a position can increase the portion of the active electrode noise (e.g., patient movement noise) that will be shared at the reference electrode noise. For example, if a patient is rolling over, the movement experienced by an active electrode positioned on the central F7 of the brain will be very different from the movement experienced by a reference electrode positioned on the contralateral ear. Meanwhile, if both electrodes are positioned on the same F7 region, they can experience similar movement artifacts. Although the signal difference can not represent some cellular electrical activity from underlying physiological structures, a large portion of the remaining signal can be attributable to such activity of interest (due to the removal of noise).

[0085] At block 615, the signal difference can be amplified. The amplification gain is between, e.g., 100 and 100,000. At block 620, the amplified signal difference can be filtered. The filter applied can include, e.g., an analog high-pass or band-pass filter. The filtering can reduce signal contributions from flowing potentials such as respiration. The filter can include a lower cutoff frequency of about 0.1-1 Hz. In some examples, the filter can also include a high cutoff frequency, which can be set to a frequency less than the Nyquist frequency determined based on the sampling rate.

[0086] At block 625, the filtered analog signal can be converted to a digital signal. At block 630, a digital filter can be applied to the digital signal. The digital filter can reduce DC signal components. The digital filtering can be performed using linear or non-linear filters. The filter can include, e.g., a finite or infinite impulse response filter or a window function (e.g., Hanning, Hamming, Blackman, or rectangular function). The filter characteristics can be defined to reduce DC signal contributions while preserving high frequency signal components.

[0087] At block 635, the filtered signal can be analyzed. As described in greater detail herein, the analysis can include micro-analysis, e.g., classifying individual segments of the signal (e.g., into sleep stages, wakefulness or non-wakefulness, and / or anticipated movement). The analysis can alternatively or additionally include macro-analysis, e.g., characterizing overall sleep quality or muscle activity.

[0088] As noted above, in some examples, multiple devices cooperate to perform the process 600. For example, Figure 4The multi-electrode device 400 can perform blocks 605-625, and a remote device (e.g., a server, a computer, a smartphone, or the interface device 405) can perform blocks 630-635. It should be appreciated that, to facilitate such shared process operations, the devices can communicate to share appropriate information. For example, after block 625, the multi-electrode device 400 can transmit the digital signal (e.g., using a short-range network or a WiFi network) to another electronic device, such as the interface device 500. Figure 5 The other electronic device can receive the signal and then perform blocks 630-635.

[0089] Although not explicitly shown in the process 600, raw and / or processed data can be stored. The data can be stored on the multi-electrode device, the remote device, and / or in the cloud. In some instances, raw data and processed data (e.g., identifying classifications associated with portions of the data) can be stored.

[0090] It should also be appreciated that the process 600 can be a continuous process. For example, the active signals and the reference signals can be collected continuously or periodically over an extended period of time (e.g., overnight). Part or all of the process 600 can be performed in real-time as the signals are collected, and / or the data can be processed entirely or partially in batches. For example, blocks 605-625 can be performed in real-time during a recording session, and the digital signal can then be performed. Blocks 630-635 can be performed periodically (e.g., every hour or after a threshold of unanalyzed data is reached) or at the end of the recording session.

[0091] Figure 7 is a flowchart of a process 700 for analyzing channel bio data to identify frequency signatures for individual bio stages, according to embodiments of the present application. Part or all of the process 700 can be implemented in a multi-electrode device (e.g., the multi-electrode device 400) Figure 4 and / or in an electronic device remote from the multi-electrode device (e.g., the interface device 500). Figure 5

[0092] At block 705, the signals can be transformed into a spectrogram. The signals can include recorded signals based from electrodes positioned on a person’s body, e.g., differentially amplified and filtered signals. By resolving the signals into time bins and computing a spectrogram for each time bin (e.g., using a Fourier transform), a spectrogram can be generated. Thus, the spectrogram can include a multi-dimensional power matrix, where the dimensions correspond to time and frequency.

[0093] ​At block 710, selected portions of the spectrogram can optionally be removed. These portions can include portions associated with a particular time bin for which poor signal quality can be determined and / or for which no reference data or insufficient reference data exists. For example, to develop a translation or mapping from signals to physiological events, reference data (e.g., human ratings corresponding to the data) can be used to determine signatures for various physiological events. Portions of data for which no reference data is available can thus be ignored in determining the signatures.

[0094] At block 715, the spectrogram can be segmented into a set of time bins or epochs. Each time bin can have the same duration (e.g., 30 seconds) and can (in some instances) include a number (e.g., and a fixed number) of time increments, where a time increment corresponds to each recorded time. In some instances, a time bin is defined as a single time increment in the spectrogram. In some instances, a time bin is defined as a plurality of time increments. The duration of a time bin can be determined based on, for example, the time scale of the physiological event of interest; the temporal precision or duration of the corresponding reference data; and / or the required precision, accuracy, and / or speed of signal classification.

[0095] At block 720, each time bin can be assigned to a cluster based on the reference data. For example, human ratings of EEG data can identify a sleep (or wake) stage for each time bin. The time bins in a given time bin can then be associated with the corresponding identified stage. As another example, the same reference data can be used to detect awakenings, which can be defined as occurring within any time bin associated with a sleep stage closer to "wake" relative to the stage of the preceding time bin. The time bins in a time bin can then be assigned to an "awakening" cluster (if an awakening occurred during the bin) or a "non-awakening" cluster. As yet another example, for a given EMG recording, a patient can indicate (e.g., verbally, with a mouse click, or with a blink) an intended control. For example, after contracting a right masseter muscle, a patient can indicate that he wants the cursor to move down. The time bins associated with the masseter contraction can then be assigned to a "down" cluster.

[0096] At block 725, the spectral features can be compared across the clusters. In one example, one or more spectral features can be determined for each time bin first, and the features of these groups can be compared at block 725. For example, a strong frequency or fragmentation value can be determined, as described in more detail herein. As another example, the power (or normalized power) at each of one or more frequencies can be compared for each time bin. In another example, a collective spectrum can be determined based on the spectra associated with the time bins assigned to a given cluster, and features can then be determined based on the collective spectrum. For example, the collective spectrum can include an average or median spectrum, and the features can include a strong frequency, a fragmentation value, or a power (at one or more frequencies). As another example, the collective spectrum can include, for each time bin, a feature that can include an nl% power (the power when nl% of the power at that frequency is below the power) and an n2% power (the power when n2% of the power at that frequency is below the power).

[0097] With these features, at block 730, one or more cluster-distinguishing frequency signatures can be identified. A frequency signature can include an identification of a variable that is identified or determined based on a given spectrum for use in cluster assignment. This variable can then be used, for example, in a clustering algorithm or data model, or compared to an absolute or relative threshold, in order to determine to which state a time bin associated with the spectrum will be assigned. For example, a cluster-distinguishing frequency signature can include a particular frequency, such that the power at that frequency will be used for cluster assignment. As another example, a cluster-distinguishing frequency can include a weight associated with each of one or more frequencies, such that a weighted sum of the powers of the frequencies will be used for cluster assignment.

[0098] A frequency signature can include a subset of frequencies and / or weights for one or more frequencies. For example, an overlap between the power distributions of two or more clusters can be determined, and a cluster-distinguishing frequency can be identified as a frequency with a low threshold overlap or a frequency with a relatively small (or smallest) overlap. In one example, a model can be used to determine which frequencies (or which frequency) features can be reliably used to distinguish between clusters. In one example, a cluster-distinguishing signature can be identified as a frequency associated with an information value (e.g., based on an entropy difference) that is higher than an absolute or relative (e.g., relative to the values of other frequencies) value.

[0099] In one example, block 730 can include assigning a weight to each of two or more frequencies. Then, in order to subsequently determine a cluster to which a spectrum will be assigned, a variable can be computed that is a weighted sum of the (normalized or unnormalized) powers. For example, block 725 can include using a component analysis (e.g., principal component analysis or independent component analysis), and block 730 can include identifying one or more components.

[0100] Figure 8is a flowchart of a process 800 according to an embodiment of the application for analyzing channel bio data to identify frequency signatures for individual bio stages. Part or all of the process 800 can be implemented in a multi-electrode device (e.g., the multi-electrode device 400 of Figure 4 FIG. 1) and / or in an electronic device (e.g., the interface device 500 of Figure 5 FIG. 2) that is remote from the multi-electrode device.

[0101] At block 805, spectrograms for samples corresponding to various physiological states can be collected. In some examples, at least some of the states correspond to sleep stages or sleep periods having particular properties. For example, samples can be collected from both sleep periods and wake periods, such that the samples can include data from one or more of the sleep and wake stages. As another example, with human sleep stage scoring, samples can be collected to ensure (e.g., equal or approximately equal) representation from each sleep stage or wake stage. As another example, samples can be collected from sleep periods that include (e.g., based on patient report or human scoring) frequent awakenings and from sleep periods that include infrequent awakenings. In some examples, the samples collected are based on records from a single individual. In another example, the samples are based on records from multiple individuals.

[0102] In some examples, at least some of the states correspond to states of consciousness. For example, samples (e.g., based on EMG data) can be collected such that some of the data corresponds to intent to initiate a particular action (e.g., move a cursor up or down) and other data corresponds to no such action.

[0103] The spectrogram data can include a spectrogram of raw data, a spectrogram of filtered data, a spectrogram that is normalized once (e.g., normalizing the power at each frequency based on the power across time bins for the same frequency or based on the power across frequencies for the same time bin), or a spectrogram that is normalized multiple times (e.g., normalizing the power at each frequency based on the normalized or unnormalized power across time bins for the same frequency or based on the normalized or unnormalized power across frequencies for the same time bin at least once).

[0104] At block 810, spectrogram data from a base state (e.g., a wake stage, a low arousal sleep state, or intent to not move a cursor) can be compared to spectrogram data from one or more non-base states (e.g., a sleep stage, a frequent arousal sleep state, or intent to move a cursor in a particular direction) to identify a significance value. In one example, for comparisons between a base state and a single non-base state, the frequency-specific significance value can include a p-value and can be determined for each frequency based on a statistical test of the distribution of power in the two states.

[0105] Then, blocks 815-820 are executed for each pairwise comparison between a non- fundamental state (e.g., a sleep stage) and a fundamental state (e.g., wakefulness). At block 815, a threshold significance number can be set. The threshold can be determined based on a distribution of the set of frequency-specific significance values and a defined percentage (n%) of the threshold. For example, the threshold significance number can be defined as the n% (e.g., 60%) of the frequency-specific significance values that are below the threshold significance number.

[0106] At block 820, a set of frequencies for which the frequency-specific significance values are below the threshold can be identified. Thus, these frequencies can include (based on the threshold significance number) frequencies that sufficiently distinguish the fundamental state from the non-fundamental state.

[0107] Then, blocks 815 and 820 are repeated for each additional comparison between the fundamental state and another non-fundamental state. The results thus include a set of n% most significant frequencies associated with each non-fundamental state.

[0108] At block 825, frequencies that are present in all of the sets (or a threshold number of the sets) are identified. Thus, the identified overlapping frequencies can include those of the n% most significant frequencies that are in the process of distinguishing each of the plurality of non-fundamental states from the fundamental state.

[0109] At block 830, a determination can be made as to whether the overlap percentage is greater than an overlap threshold. When not, process 800 can return to block 815, where a new (e.g., higher) threshold significance number can be set. For example, the threshold percentage (n%) used to define the threshold significance number at block 820 can be incremented (e.g., by 1%) so as to include more frequencies in the identified sets.

[0110] When it is determined that the overlap is greater than the overlap threshold, process 800 can continue to block 835, where one or more discriminant cluster frequency signatures can be defined using the frequencies in the overlap between the sets. The signature can include an identification of a subset of frequencies in the frequency spectrum and / or a weight for each of the one or more frequencies. The weight can be based on, for example, the frequency-specific significance value for the frequency for each of the one or more fundamental state versus non-fundamental state comparisons or (in cases where the overlap evaluation does not require the identified frequencies to be present in all of the frequency sets) a number of sets that include a given frequency. In some instances, the signature includes one or more components defined by assigning weighted frequencies in the overlap. For example, component analysis can be performed using the state assignments and power at the frequencies in the overlap to identify one or more components.

[0111] Subsequent analysis (e.g., of different data) can focus on the discriminative group(s)'s frequency signature. In some examples, a spectrogram (e.g., a normalized or unnormalized spectrogram) can be cropped to exclude frequencies that are not confined to the discriminative group's frequencies. For example, process 800 can be performed first to identify the discriminative group's frequencies, and process 700 (e.g., subsequent analysis of different data) can use the discriminative group's frequencies to crop a spectrogram of a signal before comparison.

[0112] Figure 9 is a flowchart of a process 900 according to embodiments of the application for normalizing a spectrogram and using a discriminative group's frequency signature to classify biological data. Part or all of process 900 can be implemented in a multi-electrode device (e.g., multi-electrode device 400 of Figure 4 ), and / or in an electronic device (e.g., interface device 500 of Figure 5 ) remote from the multi-electrode device.

[0113] At blocks 905 and 910, a spectrogram constructed from recorded bioelectrical signals (e.g., EEG or EMG data) is normalized (e.g., once, multiple times, or iteratively). In some embodiments, the spectrogram is constructed from channel data of one or more channels, each data generated based on a signal recorded using a device that fixes or tethers multiple electrodes relative to each other.

[0114] A first normalization performed at block 905 can be performed by first determining, for each frequency in the spectrogram, a z-score of the power associated with that frequency (i.e., across all time bins). This z-score value can then be used to normalize the power at that frequency.

[0115] An (optional) second normalization performed at block 910 can be performed by first determining, for each time bin in the spectrogram, a z-score based on the power associated with that time bin (i.e., across all frequencies). This z-score value can then be used to normalize the power at that time bin.

[0116] These normalizations can be performed (in an alternating fashion) a set number of times, or until the normalization factor (or a change in the normalization factor) is below a threshold. In some examples, normalization is performed only once, such that block 905 or block 910 is omitted from process 900. In some examples, the spectrogram is not normalized.

[0117] For each time bin in the spectrogram, a corresponding spectrum can be collected at block 915. At block 920, one or more variables can be determined for the time bin based on the spectrum and the frequency signature of one or more discriminating groups. For example, a variable can include the power at a selected frequency identified in the signature. As another example, a variable can include a value of a component defined in the signature (e.g., determined by computing a weighted sum of power values in the spectrum). Thus, in some instances, block 920 includes projecting the spectrum onto a new basis. Blocks 915 and 920 can be performed for each time bin.

[0118] At block 925, group assignments are made based on the associated variables. In some instances, individual time bins are assigned. In some instances, a set of time bins (e.g., individual epochs) are assigned to a group. The assignments can be performed, for example, by comparing the variables to a threshold (e.g., such that a variable is assigned to one group when it is below the threshold and otherwise to another group) or by using clustering or modeling techniques (e.g., Gaussian Naive Bayes classifier). In some instances, the assignments are constrained such that a given feature (e.g., time bin or time epoch) cannot be assigned to more than a specified number of groups. This number can or can not be the same as the number of groups or states (both basic and non-basic states) used to determine the frequency signature of one or more discriminating groups, depending on the implementation. The assignments can be general (e.g., such that the clustering analysis yields an assignment to one of five groups without binding any group to a particular physiological significance) or state-specific.

[0119] Further, at each time point, a fragmentation value can be defined. The fragmentation value can include a time fragmentation value or a spectral fragmentation value. For the time fragmentation value, a time gradient of the spectrogram can be determined and divided into fragments. The spectrogram can include the original spectrogram and / or a spectrogram that has been normalized 1, 2, or more times across time bins and / or across frequencies (e.g., a spectrogram that is first normalized across time bins and then across frequencies). A given fragment can include a set of time bins, each of which can be associated with a vector of partial derivative power values (across a set of frequencies). For each frequency, a gradient frequency-specific variable can be defined based on the partial derivative power values defined for any of the time bins in the time block and for the frequency. For example, the variable can be defined as the mean of the absolute values of the partial derivative power values for that frequency. The fragmentation value can be defined as the frequency with the highest or highest frequency-specific variable. The spectral fragmentation value can be similarly defined, but can be based on a spectral gradient of the spectrogram. A high fragmentation value can indicate sleep stage disruption.

[0120] Figure 10 is a flowchart of a process 1000 according to an embodiment of the present application for analyzing channel bio data to identify arousals. Part or all of the process 1000 can be performed by a multi-electrode device (e.g., a polysomnography device), a computer system, or a combination thereof. Figure 4electronic devices (e.g., a computer, a tablet, a smartphone, a smart watch, a smart television, a smart appliance, a smart car, etc.) that are in communication with the multi-electrode device 400 and / or are remote from the multi-electrode device. Figure 5 implemented in the interface device 500.

[0121] Blocks 1005 and 1010 of the process 1000 can correspond to blocks 805 and 810, respectively, of the process 800 in Figure 8 However, in the process 1000, the base state is defined as the wake state, and each of the plurality of sleep stages (e.g., stages 1-3 and the REM stage) is defined as a non-base state.

[0122] At block 1015, the frequency signatures that distinguish the clusters can be identified using an overlap analysis (e.g., in conjunction with the analysis described in blocks 815-835 of the process 800 shown in FIG. 8). The signatures can include, for example, projections to a new basis. Figure 8

[0123] At block 1020, new EEG data can be received from the device described herein or another recording device. Spectrograms can be constructed and normalized as described herein. The normalization can include one or more normalizations, such as described, for example, with reference to blocks 905 and 910 of the process 900.

[0124] The spectrograms can be divided into time bins (e.g., 30 second time bins), and each bin can be classified as "wake" or "sleep" at block 1025. This classification can be performed using any of a variety of techniques, which can include analyzing a variable corresponding to the determined signatures, analyzing the power at a particular frequency or band, or analyzing which frequencies represent the normalized power.

[0125] One or more time bins that are classified into the sleep category can be further analyzed to detect any arousals that occurred within the bin. Thus, at block 1030, a variable can be determined for each time bin based on the frequency signatures that distinguish the clusters identified at block 1015 and the power in the spectrogram for that time bin.

[0126] At block 1035, the variable can be used to assign the bin or a set of bins (e.g., a time epoch) to an arousal cluster or a non-arousal cluster. In some instances, the assignment is made by determining whether that particular variable aligns more closely with a similar variable based on wake data as compared to a variable based on stable sleep data. Thus, even brief arousals can be detected.

[0127] The following examples provide further illustration of embodiments of the application and are not intended to limit the scope of the present application. While they are exemplary of those that can be used, other procedures, methods, or techniques known to those skilled in the art can alternatively be used.

[0128] Wakefulness detection embodiments ​

[0129] Figures 11-14 An example of automated wakefulness detection performed using Process 1000 is illustrated. For each graph, a single-channel EEG recording of one night's sleep was analyzed, and an analysis of a portion of the data is shown. The top graph shows automated detection of wakefulness using Process 1000 (each detection is indicated by a vertical bar at the top) and manual detection of wakefulness (each detection is indicated by a vertical bar at the bottom). The bottom graph shows a sleep structure diagram, which indicates the manual assessment of whether the signal corresponds to a wakefulness state or a sleep stage (and which sleep stage). Sleep / wake states are assigned for each 30-second period (as shown in the bottom sleep structure diagram). Wakefulness detection occurs at a finer time scale. Therefore, wakefulness can be detected even within periods corresponding to a flat sleep structure diagram. For each dataset, sensitivity, specificity, and accuracy variables are calculated by comparing automated and manual wakefulness detection.

[0130] Example 1

[0131] General Awakening Detection

[0132] like Figure 11 As shown, manual and automated detection largely follow each other. The automated detection has a sensitivity of 72.7%, a specificity of 99.0%, and an accuracy of 98.4%. This automated wakefulness detection can also be combined with manual or automated sleep stage detection to determine the percentage of sleep stages interrupted by wakefulness. In this case (using manual sleep stage detection) and / or the amount of sleep time between wakefulness. For this dataset, the average sleep time between wakefulness is only 2.4 minutes, and the maximum is only 19 minutes. Therefore, wakefulness detection can be used to quickly analyze sleep data and provide quantifiable indicators of sleep quality.

[0133] Example 2

[0134] Awakening-based therapeutic analysis

[0135] Arousal testing can also provide an assessment of treatment. Figure 12A and 12B This illustrates the situation where continuous positive airway pressure (CPAP) is not being performed. Figure 12A Treatment followed by CPAP therapy. Figure 12BThe analysis of the sleep data of the first patient is shown. Again, the automated arousal detection is followed by the manual detection. Furthermore, the difference corresponding to the presence of CPAP is highlighted using both types of detection. Overall, arousal was present in 2.1% of the bins of time in the no-CPAP dataset (4.0% in stage 1 sleep, 3.0% in stage 2 sleep, 0% in stage 3 sleep, and 2.3% in REM) and only 1.2% of the bins of time in the CPAP dataset (2.9% in stage 1 sleep, 1.2% in stage 2 sleep, 0% in stage 3 sleep, and 1.4% in REM). Thus, arousal was reduced by 43% in the CPAP dataset, which means that the treatment was effective.

[0136] Figure 13A and 13B Similar data but for the analysis of the second patient is shown. Overall, arousal was present in 1.1% of the bins of time in the no-CPAP dataset (2.0% in stage 1 sleep, 0.9% in stage 2 sleep, 0% in stage 3 sleep, and 1.2% in REM) and only 0.9% of the bins of time in the CPAP dataset (2.5% in stage 1 sleep, 0.6% in stage 2 sleep, 0% in stage 3 sleep, and 1.3% in REM). Interestingly, for this patient, arousal was thus reduced overall by 18%, although arousal in REM was increased by 8%.

[0137] Example 3

[0138] Arousal-based pharmacodynamic analysis

[0139] Arousal-based statistical values were used to compare four cohorts in a drug study. One of the cohorts included a placebo cohort. The remaining three corresponded to drugs, each cohort associated with a different dose of the drug. For each patient, the average time between arousals was determined. ANOVA was performed to determine if the average inter-arousal time was significantly different for any of the cohorts. The second cohort was associated with a short average inter-arousal time when compared to each of the other three cohorts with p-values of 0.004, 0.002, and 0.004. Thus, automated arousal detection can be used to check the efficacy and / or side-effect profile of a drug.

[0140] Example 4

[0141] Detection of hyperarousal

[0142] Figure 14 Arousal detection in a patient experiencing multiple arousals is shown. The average inter-arousal time is only 1.4 minutes. Such frequent arousals can imply or indicate insomnia and can be used for diagnostic, monitoring, and / or treatment-evaluation purposes.

[0143] As described herein, using group-differentiated frequency signatures can be useful for biological electrical signal classification. In one embodiment, the technique relies on using the power from normalized or non-normalized spectrograms from biological data to assign each time bin to a physiologically relevant group. In some embodiments, the classification can additionally or alternatively depend on the identification of frequencies (for time bins) that are associated with particular characteristics.

[0144] Figure 15 is a flowchart of a process 1500 according to an embodiment of the application for normalizing spectrograms and identifying frequencies to classify biological data. Part or all of the process 1500 can be implemented in a multi-electrode device (e.g., the multi-electrode device 400 of Figure 4 and / or in an electronic device (e.g., the interface device 500 of Figure 5 remote from the multi-electrode device.

[0145] Block 1505 of the process 1500 can correspond to block 905 of the process 900. Thus, as should be understood from the above disclosure, each value in a spectrogram generated from biological electrical data can be normalized based on other values at the same frequency, but in different time bins. In some instances, no spectrogram normalization is performed (although in some embodiments it is).

[0146] At block 1510, for each time bin, the frequency associated with the high or highest normalized power can be identified as a strong frequency for that time bin. At block 1515, the identified strong frequencies can be used to assign each time bin or each set of time bins (e.g., a time epoch) to a group. For example, a particular sleep stage can be associated with activity in a particular frequency band. Thus, for example, strong frequencies in a particular frequency band can tend to be assigned to a particular sleep stage. The assignment can be performed, for example, using cluster analysis, component analysis, data modeling, and / or comparison to one or more thresholds.

[0147] In some embodiments, the spectrogram can be processed to highlight temporal changes in power. Frequencies associated with large change values can then be used to classify portions of the recording. Figure 16 is a flowchart of a process 1600 according to an embodiment of the application for normalizing spectrograms and using gradients to identify frequencies to classify biological data. Part or all of the process 1600 can be implemented in a multi-electrode device (e.g., the multi-electrode device 400 of Figure 4 and / or in an electronic device (e.g., the interface device 500 of Figure 5 remote from the multi-electrode device.

[0148] Blocks 1605 and 1610 of process 1600 can correspond to blocks 905 and 910 of process 900. Thus, it should be appreciated from the above disclosure that the spectrogram generated from the bioelectrical data can be normalized one, two, or more times based on the power deviation (e.g., difference) across horizontal or vertical vectors in the spectrogram.

[0149] At block 1615, a temporal gradient can be determined based on the normalized spectrogram. It should be appreciated that block 1615 can be modified to include other processing that quantifies (for each frequency) the temporal power change.

[0150] The gradient can be divided into (e.g., fixed duration) time blocks or time epochs, and at block 1620, a portion of the gradient defined for a given time block is accessed.

[0151] At block 1625, a gradient frequency-specific variable can be determined for each frequency based on the gradient portion for the time block. For a given frequency, the variable can depend on each value in the gradient portion corresponding to the frequency. The variable can include a population statistic, such as a mean, median, or maximum. In some instances, the absolute value of the gradient is computed and used in the population analysis to determine the variable.

[0152] At block 1630, a fragmentation value can be defined (for a given time block) as the frequency of the time block associated with the high (or highest) gradient frequency-specific variable. Thus, the fragmentation value can include the frequency associated with a large power adjustment over time. Process 1600 can then return to block 1620 to determine the fragmentation value for another time block.

[0153] At block 1635, the identified fragmentation values can be used in the assignment of the time block. For example, wakefulness can be associated with strong power deviations in a particular frequency band. In some instances, in addition to or instead of analyzing the frequency associated with the high gradient value, the gradient value itself (e.g., at the fragmentation value frequency and / or other frequencies) can be used in the assignment. The assignment can be performed, for example, using cluster analysis or component analysis or a data model.

[0154] In some instances, the recorded bioelectrical data can be used to assist a user's communication efforts. Figure 17 is a flowchart of a process 1700 according to an embodiment of the application for determining a mapping of EMG data using reference data. Part or all of process 1700 can be implemented in a multi-electrode device (e.g., multi-electrode device 400 of Figure 4 ) and / or an electronic device remote from the multi-electrode device (e.g., interface device 500 of Figure 5 ).

[0155] At box 1705, one or more electrodes are positioned on one or more muscles. The electrodes may include, for example, one or more active electrodes, one or more reference electrodes, and (optionally) a ground electrode. In some instances, multiple active electrodes are used, and each active electrode is positioned on a different muscle. In some instances, a single device houses the active electrodes and reference electrodes (e.g., which may be fixedly positioned within the device or flexibly tethered to the device). However, it should be understood that any electrode device configured to facilitate EMG data collection by the electrodes can be used.

[0156] At box 1710, communication-aided visualizations can be presented (e.g., on the screen of the interface device). Figure 18A and 18B An example of communication-aided visualization is shown. A visualization may include a group of letters, letter combinations, words, or phrases. A cursor can be navigated to select within that group. Selection can continue, allowing the user to progressively construct sentences or paragraphs. An example of communication-aided visualization includes... Visualization provided. In some instances, visualization is not provided during the mapping determination process.

[0157] At box 1715, real-time EMG data is accessed from the positioned electrodes. Reference data is available at box 1720 as EMG data is received from the muscle. Reference data may include any data indicating expected or desired cursor movement specified by the user from which it is collected and recorded. For example, reference data may include mouse movements, speech, or blinks in response to a question.

[0158] At box 1730, a mapping can be established between the EMG data and the cursor space using EMG data and reference data. The mapping may include, for example, a projection definition or frequency specification (e.g., indicating the desired cursor movement at a given frequency where power can be identified). The mapping may include frequency signatures that distinguish groups, where different groups may represent different cursor movements (e.g., movement direction). In some instances, the mapping includes specifications on how to preprocess the data. Such preprocessing may include, for example, normalization to be performed on the spectrogram or subtraction of the data based on records from multiple active electrodes.

[0159] In some instances, the training performed via procedure 1700 can occur in anticipation of the possibility that the user may soon be unable to communicate the reference data. Therefore, the mapping can be established before the user's ability to communicate the expected cursor movement disappears.

[0160] Figure 19 This is a flowchart of process 1900 according to an embodiment of the present invention, a process for generating written or spoken text based on EMG data. Part or all of process 1900 may be available in a multi-electrode device (e.g., Figure 4electronic devices (e.g., in the multi-electrode device 400) and / or remote from the multi-electrode device. Figure 5 implemented in the interface device 500.

[0161] At block 1905, a communication-assisted visualization (e.g., such as shown in FIGS. 18A Figure 18A or 18B) can be presented (e.g., on a display of the interface device). At block 1910, a mapping between EMG space and cursor space (e.g., determined at block 1730 in process 1700) can be accessed.

[0162] At block 1915, real-time raw or processed EMG data can be accessed. For example, the data can be processed such that it is transformed to form a spectrogram, and / or such that it is normalized (e.g., one or more times). The data can include data (or processed form thereof) received from the electrode or multi-electrode device.

[0163] With the time block of data and the mapping, a cursor position can be determined at block 1920. For example, a component value corresponding to a spectrum generated using the EMG data can be determined and mapped to a direction for moving the cursor.

[0164] A representation of the cursor can then be presented at the determined position on the visualization. A determination can be made at block 1935 as to whether a letter (or combination of letters, word, or phrase) has been selected. For example, a selection can be inferred when the cursor has reached a representation of a letter (or combination of letters, word, or phrase). In one instance, another EMG signature can be used to indicate a selection.

[0165] When a letter is not determined to have been selected, process 1900 can return to 1915, where EMG data can be monitored and processed to identify further cursor movement and reevaluate letter selection. When a selection is determined to have been made, a determination can be made at block 1940 as to whether a word is complete. The determination can be based on the selection made at block 1935 (e.g., a selection corresponding to multiple letters that would complete a word can indicate work is complete), whether a next cursor movement corresponds to a space or punctuation, or whether a combination of the currently selected letters has formed a complete word and any sentence formed using the word is grammatically correct.

[0166] When a word is not determined to be complete, process 1900 can return to 1915, where EMG data can be monitored and processed to identify further cursor movement and reevaluate word completeness. When a word is determined to be complete, process 1900 can continue to block 1945, where the word can be written on a display (if not already), in an email or document, and / or vocalized (e.g., using a speaker). Thus, the collection and analysis of EMG data can facilitate the ability of a user to communicate, even without the use of traditional speech and / or manual controls.

[0167] In some embodiments, the technology disclosed herein can analyze spectral properties of recorded bioelectric signals. The analysis can include generating a spectrogram. Embodiments of the invention can include normalizing the spectrogram one or more times (e.g., as described with reference to blocks 905 and 910 in process 900). Such normalization can highlight high frequency signal components, which can be indicative of physiological states, such as various sleep states.

[0168] Figure 20 and 21 The impact that such normalization can have on spectrogram data is illustrated. In Figure 20 each column, the two plots were generated using the same bio-signal. Also, the spectrogram in the bottom row was generated by normalizing the values in the top spectrogram across time bins and across frequency. Each column corresponds to a different recording arrangement. The leftmost column uses two un-anchored electrodes (an active electrode and a reference electrode) and positions them close to each other for recording. The middle column separates the two electrodes. The rightmost electrode includes a multi-electrode device that anchors the electrodes close to each other.

[0169] As shown, the original spectrogram is dominated by low frequency activity and has essentially no visible activity at higher frequencies. In contrast, the normalized spectrogram includes prominent activity across the entire frequency range. These spectrograms also include patterns that change over time, which suggests that activity at particular frequencies can be indicative of sleep stages.

[0170] In Figure 21 the time series "preferred frequency" plot is shown, which was determined using either the original spectrogram (top) or the spectrogram normalized across time bins and frequency (bottom). At each time point, the preferred frequency is defined as the frequency within the spectrogram that is associated with the time point having the highest z-score. In the top plot, the preferred frequency is generally 60 Hz, sometimes a very low frequency, and occasionally another frequency. The preferred frequency during the wake state shows greater variability than for the other states, although the distinction between sleep stages is difficult to discern using this variable.

[0171] In contrast, the preferred frequency determined using the normalized spectrogram varies much more. Furthermore, state-specific patterns are apparent, and even between sleep stages are discernible. Thus, Figure 20 and 21 It is shown that iterative normalization of the spectrogram can highlight subtle characteristics that distinguish spectral states.

[0172] The embodiments described herein can also be extended or detailed by the disclosure in any of U.S. Patent Application No. 13 / 129,185, U.S. Patent Application No. 11 / 431,425, U.S. Patent Application No. 13 / 270,099, WO / 2010 / 057119, WO / 2013 / 112771, and WO / 2011 / 056679. Each of these applications is incorporated by reference in its entirety herein for all purposes. Additionally, Low, P. S. "A new way to look at sleep: separation & convergence," eScholarship (2007), available on the World Wide Web at escholarship.Org / uc / item / 6250v3wk#page-56, is also incorporated by reference in its entirety herein for all purposes.

[0173] While the application has been described in connection with specific embodiments thereof, a person skilled in the art will recognize that modifications are possible. For example, the disclosure regarding signals collected by a multi-electrode device can also apply to signals collected from multiple single-electrode devices or any other one or more devices that can collect bioelectrical signals. In addition, for the disclosure regarding signals or channels for which a recording device is not specified, any device disclosed herein or any other device that can collect one or more bioelectrical signals can be used. It will also be appreciated that the embodiments disclosed herein can be combined in various combinations. For example, blocks from various flowcharts can be combined and organized in ways not explicitly shown or described herein.

[0174] Embodiments of the application (e.g., in methods, devices, computer-readable media, etc.) can be implemented using any combination of special-purpose components and / or programmable processors and / or other programmable devices. The various processes outlined herein can be implemented on the same processor or different processors in any combination. Where a component is described as being configured to perform certain operations, such configuration can be accomplished, e.g., by designing electronic circuitry to perform the operation, by programming programmable electronic circuitry (e.g., a microprocessor) to perform the operation, or any combination thereof. Further, while the embodiments described above can make reference to specific hardware and software components, those skilled in the art will appreciate that different combinations of hardware and / or software components can also be used and that particular operations described as being implemented in hardware might also be implemented in software or vice versa.

[0175] Computer programs incorporating various features of the present application can be coded and stored on a variety of computer readable media; appropriate media include magnetic disk or tape, optical storage media such as high density magnetic disk (CD) or DVD (digital versatile disk), flash memory, and other non-transitory media. Computer readable media coded with program code can be packaged with a compatible electronic device, or the program code can be provided separately from the electronic device (e.g., via Internet download or as a separately packaged computer readable storage medium).

[0176] While the application has been described with reference to the above example, it will be understood that modifications and variations are encompassed within the spirit and scope of the application. Accordingly, the application is limited only by the appended claims.

Claims

1. A system for analyzing acquired physiological data, the system comprising: A physiological data acquisition component, the physiological data acquisition component including a housing having one or more sets of electrodes, wherein each of the one or more sets of electrodes includes at least an active electrode and a reference electrode; and A computing device, a remote server, or a remote network, wherein the physiological data acquisition component communicates with the computing device, the remote server, or the remote network, and wherein the physiological data acquisition component, the computing device, the remote server, or the remote network includes: One or more data processors; and A non-transitory computer-readable storage medium comprising instructions that, when executed on the one or more data processors, cause the one or more data processors to perform a set of actions, the actions including: A spectrogram is generated using a difference signal, which is generated by subtracting a reference signal collected from a reference electrode from an active signal collected from the active electrode. A normalized spectrum is generated by normalizing the spectrum at least twice, wherein the at least two normalizations include cross-frequency normalization and cross-time normalization. For each of the multiple time slots represented in the normalized spectrogram: Identify a portion of the normalized spectrogram corresponding to the time bin; Based on the portion of the normalized spectrogram, determine one or more features of the time chamber; and Based on one or more of the aforementioned features, the portion of the time warehouse and the normalized spectrogram is assigned to a group; and An output is generated based on the allocation, wherein the output: Identify sleep stages; or Including movement control signals; or Including synthetic speech; or This includes text presented at or transmitted by the speech aid interface.

2. The system according to claim 1, wherein the group action further includes: For each of the groups, identify a set of spectra corresponding to that group, wherein each of the groups includes one or more time bins assigned based on the corresponding one or more features; and For each group in a set of groups, the collective spectrum is defined based on the group spectrum.

3. The system according to claim 2, The one or more features mentioned therein are determined based on the strong frequencies of the collective spectrum.

4. The system according to claim 2, The one or more features mentioned therein are determined based on the fragmentation values ​​of the collective spectrum.

5. The system according to claim 2, The one or more features mentioned therein are determined based on the power at a specific frequency of the collective spectrum.

6. The system of claim 1, wherein the group action further includes accessing a plurality of signal signatures, each of the plurality of signal signatures corresponding to a different type of movement or a different type of sleep, and wherein for each of the plurality of time slots, the time slot is assigned to the group further based on the plurality of signal signatures.

7. The system of claim 1, wherein each of the one or more groups of electrodes comprises the active electrode, the reference electrode, and the ground electrode.

8. The system of claim 1, wherein the component further comprises one or more additional sensors selected from accelerometers, GPS sensors, head positioning sensors, nasal airflow velocity meters, body temperature sensors, and pulse oximeters.

9. The system according to claim 1, wherein the data acquired by the physiological data acquisition component includes electromyography data represented by blinking.

10. The system of claim 1, wherein the data acquired by the physiological data acquisition component includes electromyographic data representing jaw movements.

11. The system according to claim 1, wherein the data acquired by the physiological data acquisition component is electroencephalogram (EEG) data.

12. The system of claim 1, wherein generating the output comprises: Based on the allocation, determine one or more cursor positions on the graphics display; and Based on the one or more cursor positions, determine the desired text options on the graphics display.

13. The system of claim 1, wherein the output includes the motion control signal, and wherein the one or more actions further include moving the prosthesis in accordance with the motion control signal.

14. Methods for acquiring and analyzing physiological data of subjects, including: Physiological data from the subject is acquired from a physiological data acquisition component, the physiological data acquisition component including a housing having one or more sets of electrodes, wherein each of the one or more sets of electrodes includes at least an active electrode and a reference electrode; A difference signal is generated by subtracting the reference signal collected by the reference electrode from the active signal collected by the active electrode. A spectrogram is generated using the difference signal; A normalized spectrum is generated by normalizing the spectrum at least twice, wherein the at least two normalizations include cross-frequency normalization and cross-time normalization. For each of the multiple time slots represented in the normalized spectrogram: Identify a portion of the normalized spectrogram corresponding to the time bin; Based on the portion of the normalized spectrogram, one or more features of the time chamber are determined; Based on one or more of the aforementioned features, the portion of the time warehouse and the normalized spectrogram is assigned to a group; and An output is generated based on the allocation, wherein the output: Identify sleep stages; or Including movement control signals; or Including synthetic speech; or This includes text presented at or transmitted by the speech aid interface.

15. The method of claim 14, further comprising: For each of the groups, identify a set of spectra corresponding to that group, wherein each of the groups includes one or more time bins assigned based on the corresponding one or more features; and For each group in a set of groups, the collective spectrum is defined based on the group spectrum.

16. The method according to claim 15, The one or more features mentioned therein are determined based on the strong frequencies of the collective spectrum.

17. The method according to claim 15, The one or more features mentioned therein are determined based on the fragmentation values ​​of the collective spectrum.

18. The method according to claim 15, The one or more features mentioned therein are determined based on the power at a specific frequency of the collective spectrum.

19. The method of claim 14, wherein each of the one or more groups of electrodes comprises the active electrode, the reference electrode, and the ground electrode.

20. The method of claim 14, further comprising accessing a plurality of signal signatures, each of the plurality of signal signatures corresponding to a different type of movement or a different type of sleep, wherein for each of the plurality of time slots, the time slot is assigned to a group further based on the plurality of signal signatures.

21. The method of claim 14, further comprising using the output to control or manipulate the prosthesis.

22. The method of claim 14, comprising: Perform independent component analysis or principal component analysis; and Identify each group.

23. A system for predicting the presence of a disease, the system comprising: A physiological data acquisition component, the physiological data acquisition component including a housing having one or more sets of electrodes, wherein each of the one or more sets of electrodes includes at least an active electrode and a reference electrode; and A computing device, a remote server, or a remote network, wherein the physiological data acquisition component communicates with the computing device, the remote server, or the remote network, and wherein the physiological data acquisition component, the computing device, the remote server, or the remote network includes: One or more data processors; and A non-transitory computer-readable storage medium comprising instructions that, when executed on the one or more data processors, cause the one or more data processors to perform a set of actions, the actions including: A spectrogram is generated using a difference signal, which is generated by subtracting a reference signal collected from a reference electrode from an active signal collected from the active electrode. A normalized spectrum is generated by normalizing the spectrum at least twice, wherein the at least two normalizations include cross-frequency normalization and cross-time normalization. For each of the multiple time slots represented in the normalized spectrogram: Identify a portion of the normalized spectrogram corresponding to the time bin; Based on the portion of the normalized spectrogram, one or more features of the time chamber are determined; Based on one or more of the features, the portion of the time warehouse and the normalized spectrogram is assigned to a group; Based on the allocation, multiple awakenings occurring during the sleep state are identified; and Based on the identified multiple awakenings, the presence of the subject's disease is predicted.

24. The system of claim 23, wherein the group action further comprises: For each of the groups, identify a set of spectra corresponding to that group, wherein each of the groups includes one or more time bins assigned based on the corresponding one or more features; and For each group in a set of groups, the collective spectrum is defined based on the group spectrum.

25. The system according to claim 24, The one or more features mentioned therein are determined based on the strong frequencies of the collective spectrum.

26. The system according to claim 24, The one or more features mentioned therein are determined based on the fragmentation values ​​of the collective spectrum.

27. The system according to claim 24, The one or more features mentioned therein are determined based on the power at a specific frequency of the collective spectrum.

28. The system of claim 23, wherein the group action further includes accessing a plurality of signal signatures, each of the plurality of signal signatures corresponding to a different type of movement or a different type of sleep, and wherein for each of the plurality of time slots, the time slot is assigned to the group further based on the plurality of signal signatures.

29. The system of claim 23, wherein each of the one or more groups of electrodes comprises the active electrode, the reference electrode, and the ground electrode.

30. The system of claim 23, wherein the component further comprises one or more additional sensors selected from accelerometers, GPS sensors, head positioning sensors, nasal airflow velocity meters, body temperature sensors, and pulse oximeters.

31. The system of claim 23, wherein the data acquired by the physiological data acquisition component includes electromyography data represented by blinking.

32. The system of claim 23, wherein the data acquired by the physiological data acquisition component includes electromyographic data representing jaw movements.

33. The system according to claim 23, wherein the data acquired by the physiological data acquisition component is electroencephalogram (EEG) data.

34. Methods for acquiring and analyzing physiological data of subjects, including: Physiological data from the subject is acquired from a physiological data acquisition component, the physiological data acquisition component including a housing having one or more sets of electrodes, wherein each of the one or more sets of electrodes includes at least an active electrode and a reference electrode; A difference signal is generated by subtracting the reference signal collected by the reference electrode from the active signal collected by the active electrode. A spectrogram is generated using the difference signal; A normalized spectrum is generated by normalizing the spectrum at least twice, wherein the at least two normalizations include cross-frequency normalization and cross-time normalization. For each of the multiple time slots represented in the normalized spectrogram: Identify a portion of the normalized spectrogram corresponding to the time bin; Based on the portion of the normalized spectrogram, one or more features of the time chamber are determined; and Based on one or more of the features, the portion of the time warehouse and the normalized spectrogram is assigned to a group; and Based on the allocation, multiple awakenings occurring during sleep are identified.

35. The method of claim 34, further comprising: For each of the groups, identify a set of spectra corresponding to that group, wherein each of the groups includes one or more time bins assigned based on the corresponding one or more features; and For each group in a set of groups, the collective spectrum is defined based on the group spectrum.

36. The method according to claim 35, The one or more features mentioned therein are determined based on the strong frequencies of the collective spectrum.

37. The method according to claim 35, The one or more features mentioned therein are determined based on the fragmentation values ​​of the collective spectrum.

38. The method according to claim 35, The one or more features mentioned therein are determined based on the power at a specific frequency of the collective spectrum.

39. The method of claim 34, wherein each of the one or more groups of electrodes comprises the active electrode, the reference electrode, and the ground electrode.

40. The method of claim 34, further comprising accessing a plurality of signal signatures, each of the plurality of signal signatures corresponding to a different type of movement or a different type of sleep, wherein for each of the plurality of time slots, the time slot is assigned to a group further based on the plurality of signal signatures.

41. A system for predicting the efficacy of a treatment, the system comprising: A physiological data acquisition component, the physiological data acquisition component including a housing having one or more sets of electrodes, wherein each of the one or more sets of electrodes includes at least an active electrode and a reference electrode; and A computing device, a remote server, or a remote network, wherein the physiological data acquisition component communicates with the computing device, the remote server, or the remote network, and wherein the physiological data acquisition component, the computing device, the remote server, or the remote network includes: One or more data processors; and A non-transitory computer-readable storage medium comprising instructions that, when executed on the one or more data processors, cause the one or more data processors to perform a set of actions, the actions including: A spectrogram is generated using a difference signal, which is generated by subtracting a reference signal collected from a reference electrode from an active signal collected from the active electrode. A normalized spectrum is generated by normalizing the spectrum at least twice, wherein the at least two normalizations include cross-frequency normalization and cross-time normalization. For each of the multiple time slots represented in the normalized spectrogram: Identify a portion of the normalized spectrogram corresponding to the time bin; Based on the portion of the normalized spectrogram, determine one or more features of the time chamber; and Based on one or more of the aforementioned features, the portion of the time warehouse and the normalized spectrogram is assigned to a group; and Based on the allocation, multiple awakenings occurring during the sleep state are identified; and Based on the identified multiple awakenings, the efficacy of treatment on the subject can be predicted.

42. The system of claim 41, wherein the group action further comprises: For each of the groups, identify a set of spectra corresponding to that group, wherein each of the groups includes one or more time bins assigned based on the corresponding one or more features; and For each group in a set of groups, the collective spectrum is defined based on the group spectrum.

43. The system according to claim 42, The one or more features mentioned therein are determined based on the strong frequencies of the collective spectrum.

44. The system according to claim 42, The one or more features mentioned therein are determined based on the fragmentation values ​​of the collective spectrum.

45. The system according to claim 42, The one or more features mentioned therein are determined based on the power at a specific frequency of the collective spectrum.

46. ​​The system of claim 41, wherein the group action further includes accessing a plurality of signal signatures, each of the plurality of signal signatures corresponding to a different type of movement or a different type of sleep, and wherein for each of the plurality of time slots, the time slot is assigned to the group further based on the plurality of signal signatures.

47. The system of claim 41, wherein each of the one or more groups of electrodes comprises the active electrode, the reference electrode, and the ground electrode.

48. The system of claim 41, wherein the component further comprises one or more additional sensors selected from accelerometers, GPS sensors, head positioning sensors, nasal airflow velocity meters, body temperature sensors, and pulse oximeters.

49. The system of claim 41, wherein the data acquired by the physiological data acquisition component includes electromyography data represented by blinking.

50. The system of claim 41, wherein the data acquired by the physiological data acquisition component includes electromyographic data representing jaw movements.

51. The system according to claim 41, wherein the data acquired by the physiological data acquisition component is electroencephalogram (EEG) data.

52. The system of claim 41, wherein the treatment is continuous positive airway pressure (CPAP) therapy.

53. A system for predicting the presence of one or more effects related to a specific drug, the system comprising: A physiological data acquisition component, the physiological data acquisition component including a housing having one or more sets of electrodes, wherein each of the one or more sets of electrodes includes at least an active electrode and a reference electrode; and A computing device, a remote server, or a remote network, wherein the physiological data acquisition component communicates with the computing device, the remote server, or the remote network, and wherein the physiological data acquisition component, the computing device, the remote server, or the remote network includes: One or more data processors; and A non-transitory computer-readable storage medium comprising instructions that, when executed on the one or more data processors, cause the one or more data processors to perform a set of actions, the actions including: A spectrogram is generated using a difference signal, which is generated by subtracting a reference signal collected from a reference electrode from an active signal collected from the active electrode. A normalized spectrum is generated by normalizing the spectrum at least twice, wherein the at least two normalizations include cross-frequency normalization and cross-time normalization. For each of the multiple time slots represented in the normalized spectrogram: Identify a portion of the normalized spectrogram corresponding to the time bin; Based on the portion of the normalized spectrogram, determine one or more features of the time chamber; and Based on one or more of the features, the portion of the time warehouse and the normalized spectrogram is assigned to a group; Based on the assignment, multiple pairs of awakenings occurring during sleep stages are identified; For each of the multiple pairs of awakenings: Determine the time interval between paired awakenings; Based on the inter-awake time of the aforementioned multiple pairs of awakenings, statistical values ​​were determined; and Based on the statistical values, the presence of one or more effects related to the specific drug injected or ingested by the subject is predicted.

54. The system of claim 53, wherein the group action further comprises: For each of the groups, identify a set of spectra corresponding to that group, wherein each of the groups includes one or more time bins assigned based on the corresponding one or more features; and For each group in a set of groups, the collective spectrum is defined based on the group spectrum.

55. The system according to claim 54, The one or more features mentioned therein are determined based on the strong frequencies of the collective spectrum.

56. The system according to claim 54, The one or more features mentioned therein are determined based on the fragmentation values ​​of the collective spectrum.

57. The system according to claim 54, The one or more features mentioned therein are determined based on the power at a specific frequency of the collective spectrum.

58. The system of claim 53, wherein the group action further includes accessing a plurality of signal signatures, each of the plurality of signal signatures corresponding to a different type of movement or a different type of sleep, and wherein for each of the plurality of time slots, the time slot is assigned to the group further based on the plurality of signal signatures.

59. The system of claim 53, wherein each of the one or more groups of electrodes comprises the active electrode, the reference electrode, and the ground electrode.

60. The system of claim 53, wherein the component further comprises one or more additional sensors selected from accelerometers, GPS sensors, head positioning sensors, nasal airflow velocity meters, body temperature sensors, and pulse oximeters.

61. The system of claim 53, wherein the data acquired by the physiological data acquisition component includes electromyography data represented by blinking.

62. The system of claim 53, wherein the data acquired by the physiological data acquisition component includes electromyographic data representing jaw movements.

63. The system according to claim 53, wherein the data acquired by the physiological data acquisition component is electroencephalogram (EEG) data.

64. The system according to claim 53, wherein, The statistical value is the average time between awakenings of the multiple pairs of awakenings.

65. Methods for acquiring and analyzing physiological data of subjects, including: Physiological data from the subject is acquired from a physiological data acquisition component, the physiological data acquisition component including a housing having one or more sets of electrodes, wherein each of the one or more sets of electrodes includes at least an active electrode and a reference electrode; A difference signal is generated by subtracting the reference signal collected by the reference electrode from the active signal collected by the active electrode. A spectrogram is generated using the difference signal; A normalized spectrum is generated by normalizing the spectrum at least twice, wherein the at least two normalizations include cross-frequency normalization and cross-time normalization. For each of the multiple time slots represented in the normalized spectrogram: Identify a portion of the normalized spectrogram corresponding to the time bin; Based on the portion of the normalized spectrogram, one or more features of the time chamber are determined; Based on one or more of the features, the portion of the time warehouse and the normalized spectrogram is assigned to a group; Based on the assignment, multiple pairs of awakenings occurring during sleep stages are identified; For each of the multiple pairs of awakenings: Determine the time interval between paired awakenings; Based on the inter-awake time of the aforementioned multiple pairs of awakenings, statistical values ​​were determined; and Based on the statistical values, the presence of one or more effects related to a specific drug injected or ingested by the subject is predicted.

66. The method of claim 65, further comprising: For each of the groups, identify a set of spectra corresponding to that group, wherein each of the groups includes one or more time bins assigned based on the corresponding one or more features; and For each group in a set of groups, the collective spectrum is defined based on the group spectrum.

67. The method according to claim 66, The one or more features mentioned therein are determined based on the strong frequencies of the collective spectrum.

68. The method according to claim 66, The one or more features mentioned therein are determined based on the fragmentation values ​​of the collective spectrum.

69. The method according to claim 66, The one or more features mentioned therein are determined based on the power at a specific frequency of the collective spectrum.

70. The method of claim 65, wherein each of the one or more groups of electrodes comprises the active electrode, the reference electrode, and the ground electrode.

71. The method of claim 65, further comprising accessing a plurality of signal signatures, each of the plurality of signal signatures corresponding to a different type of movement or a different type of sleep, wherein for each of the plurality of time slots, the time slot is assigned to a group further based on the plurality of signal signatures.

72. The method of claim 65, wherein the statistical value is the average of the inter-awake time of the plurality of pairs of awakenings.

73. A system for estimating sleep quality, the system comprising: A physiological data acquisition component, the physiological data acquisition component including a housing having one or more sets of electrodes, wherein each of the one or more sets of electrodes includes at least an active electrode and a reference electrode; and A computing device, a remote server, or a remote network, wherein the physiological data acquisition component communicates with the computing device, the remote server, or the remote network, and wherein the physiological data acquisition component, the computing device, the remote server, or the remote network includes: One or more data processors; and A non-transitory computer-readable storage medium comprising instructions that, when executed on the one or more data processors, cause the one or more data processors to perform a set of actions, the actions including: A spectrogram is generated using a difference signal, which is generated by subtracting a reference signal collected from a reference electrode from an active signal collected from the active electrode. A normalized spectrum is generated by normalizing the spectrum at least twice, wherein the at least two normalizations include cross-frequency normalization and cross-time normalization. For each of the multiple time slots represented in the normalized spectrogram: Identify a portion of the normalized spectrogram corresponding to the time bin; Based on the portion of the normalized spectrogram, determine one or more features of the time chamber; and Based on one or more of the features, the portion of the time warehouse and the normalized spectrogram is assigned to a group; Based on the assignment, multiple pairs of awakenings occurring during sleep stages are identified; For each of the multiple pairs of awakenings: Determine the time interval between paired awakenings; and Based on the wake-up time, an output characterizing the sleep quality of the subject is generated.

74. The system of claim 73, wherein the group action further comprises: For each of the groups, identify a set of spectra corresponding to that group, wherein each of the groups includes one or more time bins assigned based on the corresponding one or more features; and For each group in a set of groups, the collective spectrum is defined based on the group spectrum.

75. The system according to claim 74, The one or more features mentioned therein are determined based on the strong frequencies of the collective spectrum.

76. The system according to claim 74, The one or more features mentioned therein are determined based on the fragmentation values ​​of the collective spectrum.

77. The system according to claim 74, The one or more features mentioned therein are determined based on the power at a specific frequency of the collective spectrum.

78. The system of claim 73, wherein the group action further includes accessing a plurality of signal signatures, each of the plurality of signal signatures corresponding to a different type of movement or a different type of sleep, and wherein for each of the plurality of time slots, the time slot is assigned to the group further based on the plurality of signal signatures.

79. The system of claim 73, wherein each of the one or more groups of electrodes comprises the active electrode, the reference electrode, and the ground electrode.

80. The system of claim 73, wherein the component further comprises one or more additional sensors selected from accelerometers, GPS sensors, head positioning sensors, nasal airflow velocity meters, body temperature sensors, and pulse oximeters.

81. The system of claim 73, wherein the data acquired by the physiological data acquisition component includes electromyography data represented by blinking.

82. The system of claim 73, wherein the data acquired by the physiological data acquisition component includes electromyographic data representing jaw movements.

83. The system according to claim 73, wherein the data acquired by the physiological data acquisition component is electroencephalogram (EEG) data.

84. Methods for acquiring and analyzing physiological data of subjects, including: Physiological data from the subject is acquired from a physiological data acquisition component, the physiological data acquisition component including a housing having one or more sets of electrodes, wherein each of the one or more sets of electrodes includes at least an active electrode and a reference electrode; A difference signal is generated by subtracting the reference signal collected by the reference electrode from the active signal collected by the active electrode. A spectrogram is generated using the difference signal; A normalized spectrum is generated by normalizing the spectrum at least twice, wherein the at least two normalizations include cross-frequency normalization and cross-time normalization. For each of the multiple time slots represented in the normalized spectrogram: Identify a portion of the normalized spectrogram corresponding to the time bin; Based on the portion of the normalized spectrogram, one or more features of the time chamber are determined; Based on one or more of the features, the portion of the time warehouse and the normalized spectrogram is assigned to a group; Based on the assignment, multiple pairs of awakenings occurring during sleep stages are identified; For each of the multiple pairs of awakenings: Determine the time interval between paired awakenings; and Based on the wake-up time, an output characterizing the sleep quality of the subject is generated.

85. The method of claim 84, further comprising: For each of the groups, identify a set of spectra corresponding to that group, wherein each of the groups includes one or more time bins assigned based on the corresponding one or more features; and For each group in a set of groups, the collective spectrum is defined based on the group spectrum.

86. The method according to claim 85, The one or more features mentioned therein are determined based on the strong frequencies of the collective spectrum.

87. The method according to claim 85, The one or more features mentioned therein are determined based on the fragmentation values ​​of the collective spectrum.

88. The method according to claim 85, The one or more features mentioned therein are determined based on the power at a specific frequency of the collective spectrum.

89. The method of claim 84, wherein each of the one or more groups of electrodes comprises the active electrode, the reference electrode, and the ground electrode.

90. The method of claim 84, further comprising accessing a plurality of signal signatures, each of the plurality of signal signatures corresponding to a different type of movement or a different type of sleep, wherein for each of the plurality of time slots, the time slot is assigned to a group further based on the plurality of signal signatures.

Images