Unified probabilistic framework for predicting and detecting seizure onsets in the brain and multitherapeutic device

摘要

The invention relates to a system for automatically predicting and preventing the electrographic onset of a seizure in an individual, comprising: a signal acquisition component to condition and digitize a plurality of raw signals received from a transducer implanted in the individual; a preprocessor to attenuate any artifacts in the plurality of digitized signals; a feature extraction component containing processing logic to select patient-specific seizure-predictive and seizure-indicative attributes from the preprocessed signals to form an optimal feature vector; a probability estimator component that synthesizes a probability vector as an estimator of the likelihood of seizure for a plurality of prediction times; a ultitherapy activation component containing processing logic to determine the therapy modalities that are to be activated or deactivated at any time; and an implanted device including a plurality of therapy actuators to automatically activate at least one associated therapy in response to an output signal from the multitherapy activation component.