Selection of electroencephalography (eeg) channels valid for determining cognitive load of a subject

Patent metadata

Field	Value
Publication number	US-2016128593-A1
Application number	US-201514665476-A
Country	US
Kind code	A1
Filing date	Mar 23, 2015
Priority date	Nov 6, 2014
Publication date	May 12, 2016
Grant date	—

Abstract

Disclosed is a method and system for selection of Electroencephalography (EEG) channels valid for determining cognitive load of subject. According to one embodiment, EEG signals are obtained from EEG channels associated with subject performing cognitive tasks are received. Time-frequency features of EEG signals are extracted for a frequency band comprise maximum energy value, minimum energy value, average energy value, maximum frequency value, minimum frequency value, and average frequency value. Weight of an EEG channel associated with time-frequency feature is derived using statistical learning technique. Binary values for EEG channels corresponding to time-frequency feature are assigned using weight of EEG channel associated with time-frequency feature. Intersections of binary values of EEG channels corresponding to maximum energy value and average energy value, minimum energy value and average energy value, maximum frequency value and average frequency value, and minimum frequency value and average frequency value are computed. Unions of intersections are computed, wherein the unions represent EEG channels valid to determine cognitive load of subject.

First claim

What is claimed is:

1 . A method for selecting a set of Electroencephalography (EEG) channels valid for determining a cognitive load of a subject, the method comprising:
receiving, by a processor, EEG signals obtained from a plurality of EEG channels associated with a subject performing one or more cognitive tasks; extracting, by the processor, time-frequency features of the EEG signals for at least one frequency band of each EEG channel, wherein the time-frequency features include at least one of a maximum energy value, a minimum energy value, an average energy value, a maximum frequency value, a minimum frequency value, or an average frequency value; deriving, by the processor, a weight for each time-frequency feature associated with each EEG channel using a statistical learning technique; assigning, by the processor, a binary value for each EEG channel using the weight of each time-frequency feature associated with each EEG channel; computing, by the processor, a first intersection of the binary value of each EEG channel corresponding to the maximum energy value and the average energy value, a second intersection of the binary value of each EEG channel corresponding to the minimum energy value and the average energy value, a third intersection of the binary value of each EEG channel corresponding to the maximum frequency value and the average frequency value, and a fourth intersection of the binary value of each EEG channel corresponding to the minimum frequency value and the average frequency value; computing, by the processor, a first union of the first intersection and the second intersection, and a second union of the third intersection and the fourth intersection; and computing, by the processor, a third union of the first union and the second union, wherein the third union represents a set of EEG channels valid to determine a cognitive load of the subject.

2 . The method of claim 1 , wherein the EEG signals are initially processed to remove one or more artifacts from the EEG signals.

3 . The method of claim 1 , wherein the plurality of EEG channels are connected to a low resolution EEG device comprising a maximum of 10 to 20 EEG channels.

4 . The method of claim 1 , wherein the one or more cognitive tasks comprise low load cognitive tasks or high load cognitive tasks.

5 . The method of claim 1 , wherein the binary value is assigned to the at least one frequency band of each EEG channel using the weight associated with each time-frequency feature of each EEG channel.

6 . The method of claim 1 , wherein assigning the binary value comprises:
setting the binary value of each EEG channel corresponding to each of the maximum energy value, the minimum energy value, the average energy value, the maximum frequency value, the minimum frequency value, and the average frequency value equal to 1 when the corresponding weight associated with the maximum energy value, the minimum energy value, the average energy value, the maximum frequency value, the minimum frequency value, and the average frequency value of each EEG channel is greater than a threshold value; or setting the binary value equal to 0, wherein the binary value equal to 1 represents a valid EEG channel and the binary value equal to 0 represents a not valid EEG channel.

7 . The method of claim 1 , wherein the statistical learning technique is at least one of a connectionist framework based Adaptive Neural Network technique, a Maximal Information Coefficient (MIC) based technique, a minimum Redundancy Maximum Relevance (mRMR) feature selection technique, or a Hilbert-Schmidt Independence Criterion Least absolute shrinkage and selection operator technique (HSIC Lasso).

8 . The method of claim 1 , wherein the frequency band of the EEG channels is at least one of alpha, beta, theta, or delta.

9 . The method of claim 1 , wherein the third union representing the set of EEG channels valid to determine the cognitive load of the subject are assigned binary values equal to 1.

10 . The method of claim 1 , further comprising determining a weighted channel activation index for each EEG channel to determine a discrimination power of each EEG channel, wherein the weighted channel activation index is computed using binary values for each EEG channel computed for a plurality of subjects and a maximum of F values as a weighting factor, wherein the F values represent a ratio of between group variability and within group variability, and the F values are computed using Analysis of Variance (ANOVA) analysis, and the F values are greater than zero.

11 . A system for selecting a set of Electroencephalography (EEG) channels valid for determining a cognitive load of a subject, the system comprising:
a processor; and a memory coupled to the processor, wherein the processor executes programmed instructions stored in the memory to:
receive EEG signals obtained from a plurality of EEG channels associated with a subject performing one or more cognitive tasks;
extract time-frequency features of the EEG signals for at least one frequency band of each EEG channel, wherein the time-frequency features include at least one of a maximum energy value, a minimum energy value, an average energy value, a maximum frequency value, a minimum frequency value, or an average frequency value;
derive a weight for each time-frequency feature associated with each EEG channel using a statistical learning technique;
assign a binary value for each EEG channel using the weight of each time-frequency feature associated with each EEG channel;
compute a first intersection of the binary value of each EEG channel corresponding to the maximum energy value and the average energy value, a second intersection of the binary value of each EEG channel corresponding to the minimum energy value and the average energy value, a third intersection of the binary value of each EEG channel corresponding to the maximum frequency value and the average frequency value, and a fourth intersection of the binary value of each EEG channel corresponding to the minimum frequency value and the average frequency value;
compute a first union of the first intersection and the second intersection, and a second union of the third intersection and the fourth intersection; and
compute a third union of the first union and the second union, wherein the third union represents a set of EEG channels valid to determine a cognitive load of the subject.

12 . The system of claim 11 , wherein the EEG signals are initially processed to remove one or more artifacts from the EEG signals.

13 . The system of claim 11 , wherein the processor executes programmed instructions stored in the memory to assign the binary value by:
setting the binary value of each EEG channel corresponding to each of the maximum energy value, the minimum energy value, the average energy value, the maximum frequency value, the minimum frequency value, and the average frequency value equal to 1, when the corresponding weight associated with the maximum energy value, the minimum energy value, the average energy value, the maximum frequency value, the minimum frequency value, and the average frequency value of each EEG channel for the at least one frequency band respectively is greater than a threshold value; and setting the binary value is set equal to 0, wherein the binary value equal to 1 represents a valid EEG channel and the binary value equal to 0 represents a not valid EEG channel.

14 . The system of claim 11 , wherein the statistical learning technique includes at least one of a connectionist framework based Adaptive Neural Network technique, a Maximal Informati

Assignees

• Tata Consultancy Services Ltd

Inventors

• Sinharay Arijit
• Sinha Aniruddha
• Das Diptesh
• Chatterjee Debatri

Classifications

• A61B5/4088

  Diagnosing of monitoring cognitive diseases, e.g. Alzheimer, prion diseases or dementia · CPC title

• A61B5/0006

  ECG or EEG signals · CPC title

• A61B2560/0475

  Special features of memory means, e.g. removable memory cards · CPC title

• A61B5/165 Primary

  Evaluating the state of mind, e.g. depression, anxiety · CPC title

• A61B5/374

  Detecting the frequency distribution of signals, e.g. detecting delta, theta, alpha, beta or gamma waves · CPC title