Method and apparatus for detecting and classifying seizure activity

1. A method of monitoring a patient for seizure activity comprising:

monitoring a patient using one or more electromyography electrodes to obtain an electromyography signal during a monitoring period;

processing with a processor said electromyography signal to determine values of one or more features of said electromyography signal;

inputting said values of the one or more features of said electromyography signal into a neural network trained to identify an increased risk of central nervous system depression during recovery from a detected seizure;

said neural network being trained to identify said increased risk of central nervous system depression during recovery from said seizure based on levels of oxygen saturation as measured during one or more training periods;

processing said values of the one or more features of said electromyography signal using the neural network to determine one or more outputs of one or more output nodes of the neural network, wherein said one or more outputs include a first output indicating an increased risk of central nervous system depression during recovery from said seizure; and

initiating a response based on said detected seizure.

2. The method of claim 1 wherein at least one output among the one or more outputs indicates an identification of clonic phase seizure activity.

3. The method of claim 1 wherein at least one output among the one or more outputs indicates an identification of tonic phase seizure activity.

4. The method of claim 1 wherein at least one output among the one or more outputs indicates detection of an electromyography signal pattern that is indicative of normal recovery from a seizure.

5. The method of claim 4 said neural network being trained to detect normal recovery from a seizure based on a correlation between said electromyography signal pattern and normal oxygen saturation level measured for one or more patients during the one or more training periods.

6. The method of claim 1 wherein at least one output among the one or more outputs is a prediction of a value of a physiological parameter that may be experienced by the patient following some time delay.

7. The method of claim 6 wherein said physiological parameter is oxygen saturation.

8. The method of claim 1 further comprising monitoring the patient using one or more saturated oxygen sensors during said monitoring period.

9. The method of claim 1 wherein said increased risk of central nervous system depression during recovery from said seizure is associated with a level of oxygen saturation of less than 80% to 90%.

10. The method of claim 1 wherein at least one output among the one or more outputs indicates an identification of a non-epileptic psychogenic seizure.

11. The method of claim 1 wherein said processing includes decomposition of the electromyography signal using a wavelet transform.

12. The method of claim 1 wherein said one or more of features of said electromyography signal are appropriate to identify changes in clonic-phase burst activity.

13. The method of claim 1 wherein said one or more of features of said electromyography signal include a clonic-phase burst repetition rate or clonic-phase burst amplitude.

14. An apparatus for monitoring a patient for seizure activity comprising:

one or more electromyography electrodes configured to provide an electromyography signal of seizure-related muscle activity of a patient during a patient monitoring period;

a processor configured to receive the electromyography signal and process the electromyography signal to determine values of one or more features of said electromyography signal;

said processor further configured to input said values of the one or more features of said electromyography signal into a neural network trained to identify increased risk of central nervous system depression related to a detected seizure;

said neural network being trained to identify said increased risk of central nervous system depression during recovery from said seizure based on levels of oxygen saturation as measured during one or more training periods;

said processor further configured to process said values of the one or more features of said electromyography signal using the neural network to determine one or more outputs of one or more output nodes of the neural network, wherein said one or more outputs include a first output indicating detection of an increased risk of central nervous system depression during recovery from said seizure; and

said processor further configured to initiate a response based on said detected seizure.

15. The apparatus of claim 14 wherein at least one output among the one or more outputs indicates an identification of clonic phase seizure activity.

16. The apparatus of claim 14 wherein at least one output among the one or more outputs indicates identification of tonic phase seizure activity.

17. The apparatus of claim 14 further comprising a sensor suitable for measuring saturated oxygen levels for said patient.

18. The apparatus of claim 14 wherein at least one output among the one or more outputs indicates an identification of anon-epileptic psychogenic seizure.

19. The apparatus of claim 14 wherein said increased risk of central nervous system depression during recovery from said seizure is associated with a level of oxygen saturation of less than 80% to 90%.

20. The apparatus of claim 14 wherein at least one output among the one or more outputs indicates detection of an electromyography signal pattern that is indicative of normal recovery from a seizure.