Optimization of spatio-temporal pattern processing for seizure warning and prediction

• US 7,461,045 B1
• Filed: 08/18/2004
• Issued: 12/02/2008
• Est. Priority Date: 08/18/2003
• Status: Expired due to Fees

• Current Assignee: University of Florida Research Foundation, Incorporated (State University System of Florida), Arizona Board of Regents (University of Arizona)
• Original Assignee: University of Florida Research Foundation, Incorporated (State University System of Florida), Arizona Board of Regents (University of Arizona)
• Inventors: Iasemidis, Leonidas D., Shiau, Deng-Shan, Sackellares, James C., Pardalos, Panos M., Chaovalitwongse, Wanpracha A.
• Primary Examiner(s): Holmes; Michael B
• Application Number:
  US10/920,446
• Time in Patent Office: 1,567 Days
• Field of Search: None
• US Class Current: 706/52
• CPC Class Codes: A61B 5/245 specially adapted for magnetoencephalographic [MEG] signals
• CPC Class Codes: A61B 5/316 Modalities, i.e. specific diagnostic methods
• CPC Class Codes: A61B 5/369 Electroencephalography [EEG] devices for psychotechnics A61B5/16
• CPC Class Codes: A61B 5/4094 Diagnosing or monitoring seizure diseases, e.g. epilepsy
• CPC Class Codes: G06F 2218/08 Feature extraction

Abstract

An exemplary method for analyzing behavior of a system includes receiving dynamical measurement data regarding the system, applying a quadratically constrained quadratic 0-1 problem to identify data among the received data, and storing the identified data. An exemplary seizure warning method includes continuously calculating STL_max values, identifying critical sites by applying a quadratically constrained quadratic 0-1 solution to data (e.g. STL_max profiles) derived from pre-seizure onset and post-seizure onset EEG signals, monitoring a T-index curve of the identified critical sites, warning of an impending seizure, observing a seizure and then repeating the cycle by identifying critical sites using the new data, and then monitoring them.

Claims

1. A computer-implemented method for predicting an impending seizure, comprising:

receiving dynamical measurement data regarding the system;

applying a quadratically constrained quadratic 0-1 program to identify data components among the received data, wherein the applying comprises linearizing the quadratically constrained quadratic 0-1 program, and wherein the linearizing comprises introducing a new 0-1 variable for each product of two variables and then formulating the quadratically constrained quadratic 0-1 program as a linear 0-1 problem;

storing the identified data components; and

predicting an impending seizure based on the identified data components.

6. The method of claim 1, wherein the dynamical measurement data comprise multi-dimensional time series data.

2. A computer-implemented method for predicting an impending seizure, comprising:

receiving dynamical measurement data regarding the system;

applying a quadratically constrained quadratic 0-1 program to identify data components among the received data, wherein the applying comprises linearizing the quadratically constrained quadratic 0-1 program and comparing measures of statistical distance between mean values of the dynamical measurement data;

storing the identified data components; and

predicting an impending seizure based on the identified data components.

3. A computer-implemented method for predicting an impending seizure, comprising:

receiving dynamical measurement data regarding the system, wherein the dynamical measurement data includes maximum Lyapunov exponents;

applying a quadratically constrained quadratic 0-1 program to identify data components among the received data, wherein the applying comprises linearizing the quadratically constrained quadratic 0-1 program;

storing the identified data components; and

predicting an impending seizure based on the identified data components.

4. The method of claim 3, wherein the applying comprises applying a T-statistic to the dynamical measurement data.

5. A computer-implemented method for predicting an impending seizure, comprising:

receiving dynamical measurement data regarding the system;

applying a quadratically constrained quadratic 0-1 program to identify data components among the received data, wherein the applying comprises minimizing statistical distances between mean values of the dynamical measurement data;

storing the identified data components; and

predicting an impending seizure based on the identified data components.

7. A computer-implemented method for predicting an impending seizure, comprising:

receiving dynamical measurement data regarding the system;

applying a quadratically constrained quadratic 0-1 program to identify data components among the received data, wherein the applying comprises applying Karush Kuhn optimality conditions and then formulating the quadratically constrained quadratic 0-1 program as a linear mixed integer 0-1 problem;

storing the identified data components; and

predicting an impending seizure based on the identified data components.

8. The method of claim 7, comprising displaying the predicted future behavior.

9. The method of claim 7, wherein the applying comprises comparing measures of statistical distance between mean values of the dynamical measurement data.

10. The method of claim 7, wherein the dynamical measurement data includes maximum Lyapunov exponents.

11. The method of claim 10, wherein the applying comprises applying a T-statistic to the dynamical measurement data.

12. A computer-implemented method for predicting an impending seizure, comprising:

receiving dynamical measurement data regarding the system, wherein the receiving comprises receiving largest Lyapunov exponent (STLmax) values calculated based on multi-channel data signals recorded from a brain;

applying a quadratically constrained quadratic 0-1 program to identify data components among the received data;

storing the identified data components; and

predicting an impending seizure based on the identified data components.

13. The method of claim 12, wherein the applying comprises applying the quadratically constrained quadratic 0-1 program to profiles of the calculated STLmax values to identify critical, data channels recorded from a brain area.

14. The method of claim 13, further comprising:

monitoring a T-index profile of the identified critical data channels; and

warning of the impending seizure, based on the monitoring.

15. The method of claim 13 wherein the applying comprises linearizing the quadratically constrained quadratic 0-1 program, wherein the linearizing comprises introducing a new 0-1 variable for each product of two variables and then formulating the quadratically constrained quadratic 0-1 program as a linear 0-1 problem.

16. The method of claim 13, wherein the applying comprises applying Karush-Kuhn optimality conditions and then formulating the quadratically constrained quadratic 0-1 program as a linear mixed integer 0-1 problem.

17. The method of claim 13, wherein the identifying comprises minimizing statistical distances between mean values and/or other synchronization degrees of the calculated STLmax values.

18. The method of claim 13, wherein the applying comprises comparing measures of statistical distance between mean values and/or other synchronization degrees of the calculated STLmax values.