Seizure detection, prediction and prevention using neurostimulation technology and deep neural network

1. A method for neuromodulation, the method comprising:

monitoring brain activity of a patient using one or more electrodes attached to the patient;

predicting, using a first machine learning model, whether a patient will have a seizure based on the monitored brain activity of the patient;

responsive to the first machine learning model predicting that the patient will have a seizure, determining, using a second machine learning model, a neuromodulation signal pattern for preventing the predicted seizure;

using a neurostimulator to apply the determined neuromodulation signal pattern to the patient;

after applying the determined neuromodulation signal pattern to the patient, detecting whether the patient had the predicted seizure based on the monitored brain activity of the patient; and

adjusting at least the second machine learning model based on whether the patient had the predicted seizure.

2. The method of claim 1 , further comprising adjusting the first machine learning model responsive to detecting that the patient had a seizure not predicted by the first machine learning model.

3. The method of claim 1 , wherein adjusting at least the second machine learning model based on whether the patient had the predicted seizure comprises:

if the patient did not have the predicted seizure, determining that the determined neuromodulation signal pattern was successful in preventing the predicted seizure; and

if the patient did have the predicted seizure, determining that the determined neuromodulation signal pattern was unsuccessful in preventing the predicted seizure.

4. The method of claim 1 , wherein the neuromodulation signal pattern determined by the second machine learning model is specific to the patient.

5. The method of claim 1 , wherein the first machine learning model predicts whether the patient will have a seizure by determining whether the monitored brain activity of the patient matches at least one pre-seizure pattern.

6. The method of claim 5 , wherein the pre-seizure pattern matched by the first machine learning model is specific to the patient.

7. The method of claim 5 , wherein the neuromodulation signal pattern determined by the second machine learning model is specific to the pre-seizure pattern matched by the first machine learning model.

8. The method of claim 1 , wherein at least the second machine learning model comprises a deep neural network.

9. The method of claim 1 , wherein at least the first machine learning model comprises a deep neural network.

10. The method of claim 1 , wherein detecting whether the patient had the predicted seizure comprises detecting, using at least a third machine learning model, whether the patient had the predicted seizure based on the monitored brain activity of the patient.

11. The method of claim 1 , wherein determining the neuromodulation signal pattern using the second machine learning model comprises choosing and combining from a local repository of potential neuromodulation signal patterns.

12. The method of claim 11 , further comprising transferring one or more potential neuromodulation signals from a remote storage into the local repository.

13. The method of claim 1 , further comprising:

transmitting the second machine learning model to a remote storage for updating; and

receiving an updated second machine learning model from the remote storage.

14. The method of claim 1 , further comprising:

transmitting the first machine learning model to a remote storage for updating; and

receiving an updated first machine learning model from the remote storage.

15. An apparatus for neuromodulation, comprising:

a memory; and

at least one processor coupled to the memory, the processor being operative:

to monitor brain activity of a patient using one or more electrodes attached to the patient;

to predict, using a first machine learning model, whether a patient will have a seizure based on the monitored brain activity of the patient;

responsive to the first machine learning model predicting that the patient will have a seizure, to determine, using a second machine learning model, a neuromodulation signal pattern for preventing the predicted seizure;

to use a neurostimulator to apply the determined neuromodulation signal pattern to the patient;

after applying the determined neuromodulation signal pattern to the patient, to detect whether the patient had the predicted seizure based on the monitored brain activity of the patient; and

to adjust at least the second machine learning model based on whether the patient had the predicted seizure.

16. The apparatus of claim 15 , wherein determining the neuromodulation signal pattern using the second machine learning model comprises choosing and combining from a repository of potential neuromodulation signal patterns stored within the memory of the apparatus.

17. The apparatus of claim 16 , further comprising transferring one or more potential neuromodulation signals from a remote storage into the memory of the device.

18. The apparatus of claim 15 , wherein the processor is further operative:

to transmit the second machine learning model to a remote storage for updating; and

to receive an updated second machine learning model from the remote storage.

19. The apparatus of claim 15 , wherein the processor is further operative:

to transmit the first machine learning model to a remote storage for updating; and

to receive an updated first machine learning model from the remote storage.

20. A computer program product comprising a non-transitory machine-readable storage medium having machine-readable program code embodied therewith, said machine-readable program code comprising machine-readable program code configured:

to monitor brain activity of a patient using one or more electrodes attached to the patient;

to predict, using a first machine learning model, whether a patient will have a seizure based on the monitored brain activity of the patient;

responsive to the first machine learning model predicting that the patient will have a seizure, to determine, using a second machine learning model, a neuromodulation signal pattern for preventing the predicted seizure;

to use a neurostimulator to apply the determined neuromodulation signal pattern to the patient;

after applying the determined neuromodulation signal pattern to the patient, to detect whether the patient had the predicted seizure based on the monitored brain activity of the patient; and

to adjust at least the second machine learning model based on whether the patient had the predicted seizure.