System For Real-Time Measurement Of The Activity Of A Cognitive Function And Method For Calibrating Such A System

Patent metadata

Field	Value
Publication number	US-2021290142-A1
Application number	US-201816766087-A
Country	US
Kind code	A1
Filing date	Nov 21, 2018
Priority date	Nov 21, 2017
Publication date	Sep 23, 2021
Grant date	—

Abstract

The invention concerns a method for calibrating a system for real-time measurement of the activity of a cognitive function of a test subject, the method comprising the successive steps of: acquiring electrical signals representative of a neural activity of a test subject; calculating values of markers of the cognitive function activity; generating a plurality of copies of calculated values of markers and adding noise to the generated copies; and, constructing a classifier by machine learning, based on the calculated marker values and noisy copies, the classifier being suitable for measuring the activity of the cognitive function of the test subject by calculating a probability that an electrical signal representative of the neural activity of the test subject results from a predetermined activity state of the cognitive function of the test subject.

First claim

1 . A method for calibrating a system for real-time measurement of an activity of a cognitive function of a test subject, the method comprising the following successive steps:
a) acquiring electrical signals representing a neural activity of the test subject during an execution of a first task by the test subject, the first task being configured so that the execution of the first task by the test subject leads to different activity states of the cognitive function of the test subject; b) calculating markers values of the activity of the cognitive function from the electrical signals acquired in step a) and reference electrical signals, each reference electrical signal being representative of a neural activity of a reference subject during an execution of the first task by the reference subject, the reference subject being part of a first reference population, the markers values being representative of an activity state of the cognitive function of the test subject; c) generating copies of the markers values and adding noise to the copies so as to produce noisy copies; d) constructing a classifier by automatic learning from the markers values and the noisy copies, the classifier being configured to measure the activity of the cognitive function by calculating a value representing a first probability that an electrical signal representative of the neural activity of the test subject results from a predetermined activity state of the cognitive function of the test subject.

2 . Method according to claim 1 , wherein the cognitive function is working memory.

3 . Method according to claim 1 , wherein the markers values are representative of a low activity state or a high activity state of a cognitive function of a reference subject.

4 . Method according to claim 1 , including after step c) and before step d), a step of determining for each marker value, from the markers values and the noisy copies, a correlation value representing a correlation of the marker value with the cognitive function activity states, ordering the markers values according to the correlation value, and further selecting markers values based on a rank of the ordered markers values, wherein step d) is implemented based on the selected markers values.

5 . Method according to claim 4 , wherein step d) is implemented only based on the selected markers values or only based on the selected markers values and noisy copies of the selected markers values.

6 . Method according to claim 1 , wherein the first task is configured so that execution of the first task leads alternately to at least two different activity states of the cognitive function.

7 . Method according to claim 1 , wherein the first task is configured so that execution of the first task leads alternately to a low activity state and a high activity state of the cognitive function.

8 . Method according to claim 1 , wherein a second task is configured so that execution of the second task by the test subject leads to simultaneous states, each simultaneous state simultaneously corresponding to a low activity of the cognitive function and a high activity of a confusion function, the method comprising steps of:
e) calculating, by the classifier, a representative value for a second probability that an electrical signal representative of a neural activity of a second subject executing the second task results from the predetermined activity state of the test subject's cognitive function, the second subject being part of a-second reference population, the electrical signal representative of the neural activity of the second subject being then acquired during the execution of the second task by the second subject; and f) comparing the value representative of the second probability and a fixed threshold value.

9 . Method according to claim 8 , wherein the first task is configured so that execution of the first task leads alternately to at least two different activity states of the cognitive function.

10 . Method according to claim 8 , wherein the second task is configured so that execution of the second task leads alternately to a low activity state and a high activity state of the cognitive function.

11 . Method according to claim 1 , wherein one of the marker values is a representative value of a spectral power of an electrical signal, the representative value being calculated over a part of the frequency spectrum of the electric signal.

12 . Method according to claim 11 , wherein the part of the frequency spectrum of the electric signal is chosen among the α range, the β range, the γ range and the θ range.

13 . Method according to claim 1 , wherein electrical signals acquired in step a) are acquired by means of electrodes arranged in positions Fp1 and/or Cz and/or Oz and/or CP5 of a 10-20 system of the international standard for electrode placement.

14 . Method for measuring in real time an activity of a cognitive function of a test subject comprising:
a calibration step according to the method of claim 1 , a step of acquiring electrical signals representative of a neural activity of the test subject and a step of measuring in real time the activity of the cognitive function of the test subject by calculating a value representative of the first probability using a system for real-time measurement of the activity of the cognitive function of the test subject.

15 . System for real-time measurement of an activity of a cognitive function of a test subject comprising:
a subsystem for electrical signal acquisition; a processing unit;
wherein the processing unit is configured for:
a) acquiring electrical signals representing a neural activity of the test subject during an execution of a first task by the test subject, the first task being configured so that the execution of the first task by the test subject leads to different states of activity of the cognitive function of the test subject;
b) calculating markers values of the activity of the cognitive function from the electrical signals acquired and reference electrical signals, each reference electrical signal being representative of a neural activity of a reference subject during an execution of the first task by the reference subject, the reference subject being part of a first reference population, the markers values being representative of an activity state of the cognitive function of the test subject;
c) generating copies of the markers values and adding noise to the copies so as to produce noisy copies;
d) constructing a classifier by automatic learning from the markers values and the noisy copies, the classifier being configured to measure the activity of the cognitive function by calculating a value representing a first probability that an electrical signal representative of the neural activity of the test subject results from a predetermined activity state of the cognitive function of the test subject.

16 . System according to claim 15 , wherein the cognitive function is working memory.

Assignees

• Paris Sciences Lettres Quartier Latin
• Centre Nat Rech Scient
• Ecole Superieure Physique & Chimie Ind Ville De Paris
• Univ Sorbonne

Inventors

• Vialatte François
• Mora-Sanchez Aldo
• Dreyfus Gérard
• Gaume Antoine
• Pulini Alfredo Aram

Classifications

• G06F3/015

  Input arrangements based on nervous system activity detection, e.g. brain waves [EEG] detection, electromyograms [EMG] detection, electrodermal response detection · CPC title

• A61B5/374 Primary

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

• A61B5/7246

  using correlation, e.g. template matching or determination of similarity · CPC title

• A61B5/7221

  Determining signal validity, reliability or quality (preventing, reducing or removing noise induced by motion artefacts A61B5/7207; noise originating from a therapeutic or surgical apparatus A61B5/7217) · CPC title

• A61B5/7267

  involving training the classification device · CPC title