# US 10,945,654 B2 · Methods, systems, and apparatus for self-calibrating EEG neurofeedback

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 10,945,654 B2
- Document ID: US10945654B2
- Status: Granted
- Kind: B2
- Application number: 15/043,909
- Filed: 2016-02-15
- Published: 2016-08-18
- Granted: 2021-03-16
- Earliest claimed: 2015-02-14
- Assignees: Massachusetts Institute of Technology
- Applicants: Massachusetts Institute of Technology
- Inventors: Andreas Mershin; Thrasyvoulos Karydis
- Web page: https://patentgrep.com/patent/US10945654B2

## CPC

- A61B 5/375: using biofeedback
- A61B 5/4824: {Touch or pain perception evaluation (A61B5/4821 takes precedence)}
- A61B 5/6803: {Head-worn items, e.g. helmets, masks, headphones or goggles}
- A61B 5/6831: {Straps, bands or harnesses}
- A61B 5/7257: {using Fourier transforms}
- A61B 5/7267: {involving training the classification device}
- A61B 5/743: {Displaying an image simultaneously with additional graphical information, e.g. symbols, charts, function plots}
- G16H 50/70: for mining of medical data, e.g. analysing previous cases of other patients
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA

## Abstract

Methods, systems, and apparatus implementing a generalizable self-calibrating protocol coupled with machine learning algorithms in an exemplary setting of classifying perceptual states as corresponding to the experience of perceptually opposite mental states (including pain or no pain) are disclosed. An embodiment presented represents inexpensive, commercially available, wearable EEG sensors providing sufficient data fidelity to robustly differentiate the two perceptually opposite states. Low-computational overhead machine learning algorithms that can be run on a mobile platform can be used to find the most efficient feature handles to classify perceptual states as self-calibrated by the user. The invention is generalizable to states beyond just pain and pave the way towards creating EEG NFB applications targeting arbitrary, self-calibrated perceptual states in at-home and wearable settings.

## Claims

### Claim 1

A system for classifying perceptual states comprising:
an EEG sensor and signal conditioning platform;
a classifier, the classifier constructed using EEG data of only a single user; and
a mobile neurofeedback apparatus coupled to the signal conditioning platform comprising:
an input interface constructed to allow the single user to enter a first single user time-stamped marker on first EEG data taken from only the single user, the single user marker configured to be used in the classifier for machine learning of arbitrary, user-defined perceptual brain states,
the classifier constructed to receive and operate on second EEG data from the single user, the single user marker entered for one arbitrary, user-defined perceptual brain state perceived only by the single user during a calibration phase;
a processor configured to predict the one arbitrary, user-defined perceptual brain state for the same single user by using the classifier to analyze incoming second EEG signals of the same single user during an operational phase and to generate a classification result; and
a display configured for viewing of EEG data and the classification result;
wherein the classifier is further constructed to be a multi-class classifier to receive a second single user marker entered for at least one additional arbitrary, user-defined perceptual brain state opposite the one arbitrary, user-defined perceptual brain state;
and wherein the classifier also classifies non-binary intermediate brain states between the one arbitrary, user-defined perceptual brain state and the one additional arbitrary, user-defined perceptual brain state.

### Claim 2 (depends on 1)

The system of claim 1, wherein the classifier is a maximum posterior probability classifier.

### Claim 3 (depends on 1)

The system of claim 1, wherein the classification result is refined through a feedback loop in which the single user answers whether the classifier has correctly predicted one or more perceptual brain states.

### Claim 4 (depends on 1)

The system of claim 1, wherein the classifier is a k-nearest neighbor classifier.

### Claim 5 (depends on 1)

The system of claim 1, wherein the classifier uses a Fast-Fourier transform (FFT) to separate alpha frequency responses from beta frequency responses.

### Claim 6 (depends on 5)

The system of claim 5, wherein the FFT is a short FFT taken every 22 samples with a window of 256 producing feature vectors at a rate of 10 observations per second.

### Claim 7 (depends on 1)

The system of claim 1, further comprising a link between the signal conditioning platform and the neurofeedback apparatus, wherein the link is wireless.

### Claim 8 (depends on 1)

The system of claim 1, wherein the classifier includes a machine learning algorithm, the machine learning algorithm chosen from one of k-nearest neighbor, support vector machines, neural networks, or decision trees.

### Claim 9

An apparatus for classifying perceptual states, comprising:
an EEG sensor and signal conditioning platform;
a classifier, the classifier constructed using EEG data of only a single user; and
a mobile neurofeedback apparatus comprising:
an input interface for allowing the single user to enter time-stamped markers on first EEG data from the single user, the time-stamped markers used in a classifier for machine learning of arbitrary, user-defined perceptual brain states;
a processor to generate the classifier for the prediction of an arbitrary, user-defined perceptual brain state from a second EEG signal from the same single user, and to generate a classification result based on said first EEG data and said time-stamped markers from the same single user;
a display configured for viewing of the second EEG signal and the classification result; and
a link between the EEG sensor and neurofeedback apparatus;
wherein the classifier is further constructed to receive a second single user marker entered for at least one additional arbitrary, user-defined perceptual brain state opposite the one perceptual brain state;
and wherein the classifier also classifies non-binary intermediate brain states between the one arbitrary, user-defined perceptual brain state and the one additional arbitrary, user-defined perceptual brain state.