Method and apparatus for detection of nervous system disorders
First Claim
1. A computer-readable medium programmed with instructions for performing a method of predicting a seizure event, the medium comprising instructions for causing a programmable processor to:
- acquire an electroencephalograph (EEG) signal;
extract one or more EEG features from the EEG signal;
acquire a cardiovascular (CV) signal;
extract one or more CV features from the CV signal;
derive a discriminant measure from the EEG and CV features, including comparing at least one of the EEG and CV features to features extracted from a reference seizure to form at least one similarity measure, and including comparing at least one of the EEG and CV features to one or more features extracted from a reference baseline to form at least one dissimilarity measure, the discriminant measure being a logical function of the at least one similarity measure and the at least one dissimilarity measure;
compare the discriminant measure to a predetermined threshold; and
predict a seizure event when the discriminant measure exceeds the predetermined threshold.
1 Assignment
0 Petitions
Accused Products
Abstract
Systems and methods for detecting and/or treating nervous system disorders, such as seizures, are disclosed. Certain embodiments of the invention relate generally to implantable medical devices (IMDs) adapted to detect and treat nervous system disorders in patients with an IMD. Certain embodiments of the invention include detection of seizures based upon comparisons of long-term and short-term representations of physiological signals. Other embodiments include prediction of seizure activity based upon analysis of physiological signal levels. An embodiment of the invention monitors the quality of physiological signals, and may be able to compensate for signals of low signal quality. A further embodiment of the invention includes detection of seizure activity following the delivery of therapy.
79 Citations
19 Claims
-
1. A computer-readable medium programmed with instructions for performing a method of predicting a seizure event, the medium comprising instructions for causing a programmable processor to:
-
acquire an electroencephalograph (EEG) signal; extract one or more EEG features from the EEG signal; acquire a cardiovascular (CV) signal; extract one or more CV features from the CV signal; derive a discriminant measure from the EEG and CV features, including comparing at least one of the EEG and CV features to features extracted from a reference seizure to form at least one similarity measure, and including comparing at least one of the EEG and CV features to one or more features extracted from a reference baseline to form at least one dissimilarity measure, the discriminant measure being a logical function of the at least one similarity measure and the at least one dissimilarity measure; compare the discriminant measure to a predetermined threshold; and predict a seizure event when the discriminant measure exceeds the predetermined threshold. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
-
-
18. A system for predicting a seizure event, comprising:
-
means for acquiring electroencephalograph (EEG) signals; means for acquiring cardiovascular (CV) signals; and an implantable medical device (IMD) having a power supply, housing, memory, and a processor, the processor being adapted to receive EEG signals and extract one or more EEG features; receive CV signals and extract one or more CV features; derive a discriminant measure from the EEG and CV features, including comparing at least one of the EEG and CV features to features extracted from a reference seizure to form at least one similarity measure, and including comparing at least one of the EEG and CV features to one or more features extracted from a reference baseline to form at least one dissimilarity measure, the discriminant measure being a logical function of the at least one similarity measure and the at least one dissimilarity measure; and predict a seizure event when the discriminant measure exceeds a predetermined threshold.
-
-
19. A system for predicting a seizure event, comprising:
-
a programmable processor; and a computer readable storage medium coupled to the programmable processor, the storage medium comprising executable instructions for causing the programmable processor to extract one or more EEG features from an EEG signal, extract one or more CV features from a CV signal, compare at least one of the EEG and CV features to one or more features extracted from a reference baseline to form at least one dissimilarity measure, calculate a discriminant measure as a function of the at least one dissimilarity measure, and predict a seizure event when the discriminant measure meets a predetermined criteria.
-
Specification