Brain wave classification with beneficial action recommendations
Systems and methods for interpreting a user's brain wave data from EEGs using machine learning algorithms are described. In one example, a brain activity interpretation system classifies the EEG recording into five separate component signals, representing the five categories of brain waves (alpha, beta, theta, delta, gamma). In one example, the most dominant component signal is analyzed to determine whether the amplitude is higher or lower than an optimal range within the bandwidth for that brain wave. The system can then provide intelligent recommendations to the user for beneficial action(s) to help improve their everyday functioning and promote better regulation of their brain states. The system can help the user become more aware of their mental state and how to improve their mental state. EEG data is highly complex and unsuitable for immediate human comprehension, thus, the disclosed systems and methods improve the speed and accuracy of brain wave analysis.
1 . A method for interpreting brain wave data for medical research, diagnosis, or behavioral therapy, the method comprising:
receiving, at a brain activity interpretation system, a first electroencephalogram (EEG) recording for a first person;
passing the first EEG recording as input to a machine learning model of the brain activity system, the machine learning model trained to identify a user's brain states based on distinct bandwidths detected in EEG data;
separating, via the machine learning model, the first EEG recording into at least a first component signal falling within a first bandwidth and a second component signal falling within a second bandwidth that differs from the first bandwidth;
determining, at the brain activity interpretation system, the first component signal is outside of a predefined sub-range, the predefined sub-range lies within the first bandwidth, and wherein the first component signal is associated with a first type of brain state dysregulation;
selecting, at the brain activity interpretation system and based on the first type of brain state dysregulation, a first action from an action repository,
wherein the first action includes adjusting a nutritional intake at lunch, meditation or taking a walk after lunch, or refraining from writing emails or engaging in communications that require an emotional investment from the user;
displaying, via a dashboard for the brain activity interpretation system, the first action;
receiving, at the brain activity interpretation system, a second EEG recording for the first person after display of the first action;
determining, at the brain activity interpretation system, the first component signal is outside of a predefined sub-range, the predefined sub-range lies within the first bandwidth, and wherein the first component signal is associated with a first type of brain state dysregulation;
separating, using the machine learning model of the brain activity interpretation system, the second EEG recording into at least a third component signal, and wherein the third component signal is falling within the first bandwidth;
determining, at the brain activity interpretation system, the third component signal is within a predefined sub-range, the predefined sub-range lies within the first bandwidth,
wherein the third component signal is associated with a normal brain activity, and
wherein in the normal brain activity wave shapes are measured from 0.5 μV to 100 μV in amplitude peak to peak; and
displaying, via the dashboard, brain wave data indicates a numerical comparison between the first component signal and third component signal, wherein the brain wave data provides instructions to the user to improve the user's mental health.
2 . The method of claim 1 , wherein the first bandwidth corresponds to a frequency range associated with one of alpha brainwaves, beta brainwaves, delta brainwaves, theta brainwaves, and gamma brainwaves.
3 . The method of claim 1 , wherein an average amplitude of the first component signal either exceeds a first threshold that lies within the first bandwidth or falls below a second threshold that lies within the first bandwidth.
4 . The method of claim 1 , further comprising automatically performing, via the brain activity interpretation system, the first action.
5 . The method of claim 4 , wherein the first action includes playback of music or initiation of an audio-based guided meditation session.
6 . The method of claim 1 , further comprising identifying a first neurotransmitter associated with brain waves characterized by the first component signal, wherein the first action promotes improved regulation of the first neurotransmitter.
7 . The method of claim 1 , further comprising classifying, using the machine learning model, features of the first EEG recording into five component signals including an alpha brainwave signal, a beta brainwave signal, a delta brainwave signal, a theta brainwave signal, and a gamma brainwave signal.
8 . The method of claim 7 , further comprising determining the first component signal is the most dominant of the five component signals.
9 . The method of claim 7 , wherein the machine learning model incorporates a gradient boosting ensemble method to classify the EEG recording.
10 . A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to interpret brain wave data for medical research, diagnosis, or behavioral therapy by performing the following:
receive, at a brain activity interpretation system, a first electroencephalogram (EEG) recording for a first person;
pass the first EEG recording as input to a machine learning model of the brain activity system, the machine learning model trained to identify a user's brain states based on distinct bandwidths detected in EEG data;
separate, via the machine learning model, the first EEG recording into at least a first component signal falling within a first bandwidth and a second component signal falling within a second bandwidth that differs from the first bandwidth;
determine, at the brain activity interpretation system, the first component signal is outside of a predefined sub-range, the predefined sub-range lies within the first bandwidth, and wherein the first component signal is associated with a first type of brain state dysregulation;
select, at the brain activity interpretation system and based on the first type of brain state dysregulation, a first action from an action repository,
wherein the first action includes adjusting their nutritional intake at lunch, meditation or taking a walk after lunch, or refraining from writing emails or engaging in communications that require an emotional investment from the user;
display, via a dashboard for the brain activity interpretation system, the first action;
receive, at the brain activity interpretation system, a second EEG recording for the first person after display of the first action;
determine, at the brain activity interpretation system, the first component signal is outside of a predefined sub-range, the pre-defined sub-range lies within the first bandwidth, and wherein the first component signal is associated with a first type of brain state dysregulation;
separate, using the machine learning model of the brain activity interpretation system, the second EEG recording into at least a third component signal, and wherein the third component signal falling within the first bandwidth;
determine, at the brain activity interpretation system, the third component signal is within a predefined sub-range, the predefined sub-range lies within the first bandwidth,
wherein the third component signal is associated with a normal brain activity, and
wherein in the normal brain activity wave shapes are measured from 0.5 μV to 100 μV in amplitude peak to peak; and
display, via the dashboard, brain wave data indicates a numerical comparison between the first component signal and third component signal, wherein the brain wave data provides instructions to the user to improve the user's mental health.
11 . The non-transitory computer-readable medium storing software of claim 10 , wherein the first bandwidth corresponds to a frequency range associated with one of alpha brainwaves, beta brainwaves, delta brainwaves, theta brainwaves, and gamma brainwaves.
12 . The non-transitory computer-readable medium storing software of claim 10 , wherein an average amplitude of the first component signal either exceeds a first threshold that lies within the first bandwidth or falls below a second threshold that lies within the first bandwidth.
13 . The non-transitory computer-readable medium storing software of claim 10 , wherein the instructions further cause the one or more computers to automatically perform, via the brain activity interpretation system, the first action.
14 . The non-transitory computer-readable medium storing software of claim 10 , wherein the first action includes playback of music or initiation of an audio-based guided meditation session.
15 . A system for interpreting brain wave data for medical research, diagnosis, or behavioral therapy, the system comprising one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to:
receive, at a brain activity interpretation system, a first electroencephalogram (EEG) recording for a first person;
pass the first EEG recording as input to a machine learning model of the brain activity system, the machine learning model trained to identify a user's brain states based on distinct bandwidths detected in EEG data;
separate, via the machine learning model, the first EEG recording into at least a first component signal falling within a first bandwidth and a second component signal falling within a second bandwidth that differs from the first bandwidth;
determine, at the brain activity interpretation system, the first component signal is outside of a predefined sub-range, the predefined sub-range lies within the first bandwidth, and wherein the first component signal is associated with a first type of brain state dysregulation;
select, at the brain activity interpretation system and based on the first type of brain state dysregulation, a first action from a action repository,
wherein the first action includes adjusting a nutritional intake at lunch, meditation or taking a walk after lunch, or refraining from writing emails or engaging in communications that require an emotional investment from the user;
display, via a dashboard for the brain activity interpretation system, the first action;
receive, at the brain activity interpretation system, a second EEG recording for the first person after display of the first action;
determine, at the brain activity interpretation system, the first component signal is outside of a predefined sub-range, the predefined sub-range lies within the first bandwidth, and wherein the first component signal is associated with a first type of brain state dysregulation;
separate using the machine learning model of the brain activity interpretation system, the second EEG recording into at least a third component signal, and wherein the third component signal is falling within the first bandwidth;
determine, at the brain activity interpretation system, the third component signal is within a predefined sub-range, the predefined sub-range lies within the first bandwidth,
wherein the third component signal is associated with a normal brain activity, and
wherein in the normal brain activity wave shapes are measured from 0.5 μV to 100 μV in amplitude peak to peak; and
display, via the dashboard, brain wave data indicates a numerical comparison between the first component signal and third component signal, wherein the brain wave data provides instructions to the user to improve the user's mental health.
16 . The system of claim 15 , wherein the instructions further cause the one or more computers to classify, using the machine learning model, features of the EEG recording into five component signals including an alpha brainwave signal, a beta brainwave signal, a delta brainwave signal, a theta brainwave signal, and a gamma brainwave signal.
17 . The system of claim 16 , wherein the instructions further cause the one or more computers to determine the first component signal is the most dominant of the five component signals.
18 . The system of claim 16 , wherein the machine learning model incorporates a gradient boosting ensemble method to classify the EEG recording.