Measurement of EEG reactivity

• US 7,844,324 B2
• Filed: 02/14/2007
• Issued: 11/30/2010
• Est. Priority Date: 02/14/2007
• Status: Active Grant

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Viertiö-Oja, Hanna E., Virtanen, Juha, Särkelä, Mika, Salmi, Tapani
• Primary Examiner(s): Lacyk; John P
• Assistant Examiner(s): DORNA, CARRIE R
• Application Number:
  US11/674,732
• Publication Number:
  US 20080194981A1
• Time in Patent Office: 1,385 Days
• Field of Search: 600544-546
• US Class Current: 600/544
• CPC Class Codes: A61B 5/377 using evoked responses
• CPC Class Codes: A61B 5/7203 for noise prevention, reduction or removal
• CPC Class Codes: A61B 5/7221 Determining signal validity, reliability or quality A61B5/7207 and A61B5/7217 take precedence
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records

Abstract

The invention relates to a method and apparatus for assessing the reactivity observable in a certain physiological signal, especially the EEG signal, of a comatose subject. In order to obtain an objective and a reliable measure of the reactivity automatically and without the presence of a trained EEG specialist, a valid signal model is constructed for an EEG signal obtained from the subject. A time reference corresponding to a stimulus is applied and further signal data is obtained from the time series, the further signal data being subsequent to the time reference. By employing the further signal data, the method tests whether the signal model remains to be a valid signal model for the EEG signal also after the stimulus, and indicates, based on the test, whether reactivity is present in the physiological signal.

Claims

1. A method for assessing physiological signal reactivity of a subject in a single-trial analysis, the method comprising:

obtaining a single continuous physiological signal from the subject;

supplying a stimulus to the subject while the physiological signal is being obtained;

constructing, based only on a time series of the physiological signal obtained from the subject, a signal model for the physiological signal using a processor;

applying a single time reference to the time series corresponding to the time the stimulus is supplied;

obtaining further signal data from the time series of the physiological signal, the further signal data being obtained from the subject subsequent to the time reference;

testing whether the signal model is a valid signal model for the physiological signal, wherein the testing includes employing the signal model constructed from the physiological signal obtained at least partially prior to the single time reference and the further signal data obtained subsequent to the single time reference; and

indicating, based on the testing, whether reactivity is present in the physiological signal.

2. A method according to claim 1, wherein the constructing includes constructing the signal model based on the time series of the physiological signal, in which the physiological signal is an EEG signal.

3. A method according to claim 1, wherein the constructing includes constructing the signal model based on the time series of the physiological signal, in which the signal model is an autoregressive model.

4. A method according to claim 1, wherein the testing includes

calculating a prediction value based on the signal model;

and

defining a prediction error based on the further signal data and the prediction value.

5. A method according to claim 4, wherein the indicating includes comparing the prediction error with a predetermined threshold value and deciding, based on the comparing, whether reactivity is present in the physiological signal.

6. A method according to claim 4, wherein the defining includes defining the prediction error, in which the prediction error is dependent on the difference between the prediction value and a value obtained in the further signal data.

7. A method according to claim 1, wherein

the constructing includes deriving the signal model based on a first time window of the time series and first data based on the signal model;

and

the testing includes deriving (i) a further signal model based on a second time window of the time series and second data based on the further signal model, wherein the second time window includes the further signal data, and (ii) an error value dependent on the difference of the first and second data.

8. A method according to claim 1, wherein the constructing includes deriving first data based on the signal model and the testing includes deriving second data based on the further signal model, in which the signal model and the further signal model are dependent on an autocorrelation function.

9. A method according to claim 1, wherein the supplying includes supplying the stimulus to the subject, in which the stimulus is a selected stimulus in a sequence comprising a plurality of stimuli in ascending order of intensity.

10. A method according to claim 9, further comprising repeating the constructing, applying, obtaining, and testing for the next stimulus in the sequence, wherein the repeating is performed as long as the testing indicates that reactivity is present in the physiological signal.

11. A method according to claim 10, wherein the indicating further includes determining, based on the intensity of a stimulus, a numeric value for the reactivity, in which the stimulus is the first stimulus in the sequence for which the determining indicates that reactivity is present in the physiological signal.