System and method for detecting an epileptic seizure in a prone epileptic person

• US 9,161,711 B2
• Filed: 08/19/2009
• Issued: 10/20/2015
• Est. Priority Date: 08/19/2009
• Status: Active Grant

• Current Assignee: Commissariat a L'Energie Atomique (French Alternative Energies and Atomic Energy Commission), Movea, Move Incorporated (News Corporation)
• Original Assignee: French Alternative Energies and Atomic Energy Commission, Movea
• Inventors: Jallon, Pierre, Bonnet, Stphane
• Primary Examiner(s): Dougherty, Sean
• Assistant Examiner(s): Stout, Michael C
• Application Number:
  US13/391,239
• Publication Number:
  US 20120165705A1
• Time in Patent Office: 2,253 Days
• Field of Search: 600/595
• US Class Current: 1/1
• CPC Class Codes: A61B 2562/0219 Inertial sensors, e.g. accelerometers, gyroscopes, tilt switches
• CPC Class Codes: A61B 5/1118 Determining activity level
• CPC Class Codes: A61B 5/4094 Diagnosing or monitoring seizure diseases, e.g. epilepsy
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: G06F 18/295 Markov models or related models, e.g. semi-Markov models; Markov random fields; Networks embedding Markov models
• CPC Class Codes: G06F 2218/12 Classification; Matching

Abstract

A system for detecting an epileptic seizure in a prone person, comprising:

• • at least one motion sensor with at least one measurement axis having fastening means for securing said motion sensor to said person;
  • first means for determining a first probability of at least a first state transition diagram of the nocturnal activity of a prone person with respect to data representing the measurement signals of the motion sensor, said first diagram comprising predetermined probabilities of oriented transitions between two different or identical states, the probabilities of the states of said first diagram being predetermined; and
  • second means for determining a second probability of at least a second state transition diagram for an epileptic seizure with respect to data representing the measurement signals of the motion sensor, said second diagram comprising predetermined probabilities of oriented transitions between two different or identical states, the probabilities of the states of said second diagram being predetermined.

Claims

1. A system for detecting an epileptic seizure in a prone person comprising:

at least one motion sensor configured to couple to a person, the at least one motion sensor having at least one measurement axis; and

a computing device configured to perform steps comprising;

a first determination of a first probability of at least a first state transition diagram of nocturnal activity of a prone person with respect to data representing measurement signals of the motion sensor, said first diagram comprising predetermined probabilities of oriented transitions between two different or identical states, the probabilities of the states of said first diagram being predetermined;

a second determination of a second probability of at least a second state transition diagram for an epileptic seizure with respect to data representing the measurement signals of the motion sensor, said second diagram comprising predetermined probabilities of oriented transitions between two different or identical states, the probabilities of the states of said second diagram being predetermined;

associating a state of said person as a function of probabilities of the measurement signals of the motion sensor, said probabilities being calculated separately for high frequencies of the measurement signals above a first frequency threshold, and for low frequencies of the measurement signals below a second frequency threshold, the second frequency threshold being equal to or less than the first frequency threshold;

calculating relations between the first probability and the second probability; and

detecting an epileptic seizure when at least one of said calculated relations is below a threshold.

2. The system of claim 1, wherein at least one of said state transition diagrams is adapted to use a hidden Markov model.

3. The system of claim 2, wherein computing device is further configured to perform steps comprising:

filtering, for each measurement axis of the motion sensor, high frequencies above the first frequency threshold, and low frequencies below the second frequency threshold;

calculating a first value x equal to a linear combination of respective variations along each measurement axis, between two successive time intervals, of said low frequencies per time interval n;

calculating a second value y equal to a mean of energies, along each measurement axis, of said high frequencies;

determining a probability P_LF(x) of said first value x defined by a normal centered Gaussian distribution; and

determining a probability P_HF(x) of said second value y defined by a Chi-squared distribution with a degree of freedom equal to the number of measurement axes of the motion sensor taken into consideration.

4. The system of claim 3, wherein a probability density P(x(n),y(n)) of obtaining a pair of values (x(n), y(n)) for the low frequency component and the high frequency component being equal to a product of the probability density P_LF(x) of obtaining a value x(n) for the low frequency component and a probability density P_HF(x) of obtaining a value y(n) for the high frequency component, said probability densities P_LF(x) and P_HF(x) being defined by the following expressions:

5. The system of claim 2, wherein said hidden Markov model comprises no more than five states chosen from a group consisting of a rest activity, a slight agitation activity, a tremor activity, an agitation activity and a strong agitation activity.

6. The system of claim 2, wherein the probabilities of the states of said state transition diagrams are defined by the following relation:

7. The system of claim 1, wherein said second determination further determines a second probability of a second state transition diagram for an epileptic seizure with clonic manifestations and of a second state transition diagram for an epileptic seizure with tonic manifestations.

8. The system of claim 1, wherein the states and the probabilities of the states are identical for the first state transition diagram and for the second state transition diagrams.

9. The system of claim 6, wherein there are 18 of said pairs of variances (σ

_x⁽ⁱ⁾, σ

_y⁽ⁱ⁾), which are obtained by a combination of the following values;

σ

_x[0]=5×

10<sup>−

3</sup>, σ

_x[1]=1.8×

10<sup>−

2</sup>, σ

_x[2]=3.5×

10<sup>−

2</sup>, σ

_x[3]=5×

0.510<sup>−

2</sup>, σ

_x[4]=8×

10<sup>−

2</sup>, σ

_x[5]=1×

10<sup>−

1</sup>, and σ

_y[0]=1×

10<sup>−

2</sup>, σ

_y[1]=3×

10<sup>−

2</sup>, σ

_y[2]=8×

10<sup>−

2</sup>.

10. The system of claim 9, wherein said coefficients α

_i,p are defined as follows, for a pair of variances (σ

_x^(p)[m], σ

_y^(p)[n]), where p is an index such that p=m+6n, varying from 0 to 17 for the 18 pairs (σ

_x^(p), σ

_y^(p));

11. The system of claim 10, wherein a first state transition diagram of general nocturnal activity is defined by the following matrix:

12. The system of claim 10, wherein a second state transition diagram for an epileptic seizure with clonic manifestations is defined by the following matrix:

13. The system of claim 10, wherein a second state transition diagram for an epileptic seizure with tonic manifestations is defined by the following matrix:

14. The system of claim 1, further comprising an alarm configured to provide a warning of the detection of an epileptic seizure.

15. The system of claim 1, wherein said motion sensor comprises an accelerometer, a magnetometer or a gyrometer.

16. A method for detecting an epileptic seizure in a prone person, the method comprising:

obtaining measurement signals from a motion sensor configured to be coupled to a prone person, the motion sensor having at least one measurement axis;

determining a first probability of at least a first state transition diagram of nocturnal activity of the prone person, with respect to the measurement signals of the motion sensor, said first diagram comprising predetermined probabilities of oriented transitions between two different or identical states, the probabilities of the states of said first diagram being predetermined;

determining a second probability of at least a second state transition diagram for an epileptic seizure, with respect to the measurement signals of the motion sensor, where said second diagram comprises predetermined probabilities of oriented transitions between two different or identical states, the probabilities of the states of said second diagram being predetermined;

associating a state of said person as a function of probabilities of the measurement signals of the motion sensor, said probabilities being calculated separately for high frequencies of the measurement signals above a first frequency threshold, and for low frequencies of the measurement signals below a second frequency threshold, the second frequency threshold being equal to or less than the first frequency threshold;

calculating relations of the first probability and the second probability; and

detecting an epileptic seizure when at least one of said calculated relations is below a threshold.