Unified probabilistic framework for predicting and detecting seizure onsets in the brain and multitherapeutic device
First Claim
1. A method for automatically predicting and preventing the electrographic onset of a seizure in an individual, comprising the acts of:
- monitoring a plurality of signals indicative of the activity of the brain of the individual;
extracting a set of features from the signals and forming an optimal feature vector;
synthesizing a probability vector based on the optimal feature vector as an estimator of the likelihood of seizure for a plurality of prediction time intervals; and
preventing the electrographic onset of a seizure by the automatic application of at least one intervention measure that is commensurate with the likelihood of seizure.
1 Assignment
0 Petitions
Accused Products
Abstract
A method and an apparatus for predicting and detecting epileptic seizure onsets within a unified multiresolution probabilistic framework, enabling a portion of the device to automatically deliver a progression of multiple therapies, ranging from benign to aggressive as the probabilities of seizure warrant. Based on novel computational intelligence algorithms, a realistic posterior probability function P(ST|x) representing the probability of one or more seizures starting within the next T minutes, given observations x derived from IEEG or other signals, is periodically synthesized for a plurality of prediction time horizons. When coupled with optimally determined thresholds for alarm or therapy activation, probabilities defined in this manner provide anticipatory time-localization of events in a synergistic logarithmic-like array of time resolutions, thus effectively circumventing the performance vs. prediction-horizon tradeoff of single-resolution systems. The longer and shorter prediction time scales are made to correspond to benign and aggressive therapies respectively. The imminence of seizure events serves to modulate the dosage and other parameters of treatment during open-loop or feedback control of seizures once activation is triggered. Fast seizure onset detection is unified within the framework as a degenerate form of prediction at the shortest, or even negative, time horizon. The device is required to learn in order to find the probabilistic prediction and control strategies that will increase the patient'"'"'s quality of life over time. A quality-of-life index (QOLI) is used as an overall guide in the optimization of patient-specific signal features, the multitherapy activation decision logic, and to document if patients are actually improving.
213 Citations
136 Claims
-
1. A method for automatically predicting and preventing the electrographic onset of a seizure in an individual, comprising the acts of:
-
monitoring a plurality of signals indicative of the activity of the brain of the individual;
extracting a set of features from the signals and forming an optimal feature vector;
synthesizing a probability vector based on the optimal feature vector as an estimator of the likelihood of seizure for a plurality of prediction time intervals; and
preventing the electrographic onset of a seizure by the automatic application of at least one intervention measure that is commensurate with the likelihood of seizure. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22)
-
-
23. A method for assessing a quality of life index in an individual subject to seizures in order to adjust an implanted device to optimize patient-specific feature signals and treatment therapies, comprising the acts of:
-
accumulating the energy of raw intracranial electroencephalograms (IEEG) for the individual over multiple data channels during seizures over a fixed time period;
accumulating the energy of a treatment control effort over the multiple data channels over all times of activation of the implanted device over a fixed time period;
weighting the accumulated energy of the IEEG and the accumulated energy of the control effort by seizure and treatment factors to determine a quality for the fixed period of time; and
determining a quality of life index as a weighted average of a current and previous qualities for a plurality of fixed time periods. - View Dependent Claims (24, 25, 26, 27, 28)
-
-
29. A method for periodic learning to improve and maintain the performance of a device implanted in an individual subject to seizures to provide treatment therapies, comprising the acts of:
-
assessing a quality of life index that penalizes the intensity, duration and frequency of seizures and treatments over a fixed period of time;
marking a time of unequivocal electrographic onset (UEO) in all recorded seizures over a previous fixed period of time;
creating learning sets of data based on the UEOs by clipping all the IEEG epochs immediately preceding seizures and labeling the clipped epochs as preseizure raw data;
clipping and labeling randomly chosen, non-overlapping data as non-preseizure or baseline raw data;
generating a time series of all features in a feature library from the preseizure and nonpreseizure raw data;
searching for an optimal feature vector in a power set of the feature library to minimize a classifier-based performance metric;
synthesizing a posterior probability estimator for the optimal feature vector; and
coupling an optimal therapy activation threshold to the probability estimator. - View Dependent Claims (30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68)
-
-
69. A computer readable medium containing a computer program product for automatically predicting and preventing the electrographic onset of a seizure in an individual, the computer program product comprising:
-
program instructions that monitor a plurality of signals indicative of the activity of the brain of the individual;
program instructions that extract a set of features from the signals and form an optimal feature vector;
program instructions that synthesize a probability vector based on the optimal feature vector as an estimator of the likelihood of seizure for a plurality of prediction time intervals; and
program instructions that prevent the electrographic onset of a seizure by initiating the automatic application of at least one intervention measure that is commensurate with the likelihood of seizure. - View Dependent Claims (70, 71, 72, 73, 74, 75, 83)
-
-
76. A computer readable medium containing a computer program product for assessing a quality of life index in an individual subject to seizures in order to adjust an implanted device to optimize patient-specific feature signals and treatment therapies, the computer program product comprising:
-
program instructions that accumulate the energy of raw intracranial electroencephalograms (IEEG) for the individual over multiple data channels during seizures over a fixed time period;
program instructions that accumulate the energy of a treatment control effort over the multiple data channels over all times of activation of the implanted device over a fixed time period;
program instructions that weight the accumulated energy of the IEEG and the accumulated energy of the control effort by seizure and treatment factors to determine a quality for the fixed period of time; and
program instructions that determine a quality of life index as a weighted average of a current and previous qualities for a plurality of fixed time periods. - View Dependent Claims (77, 78)
-
-
79. A computer readable medium containing a computer program product for periodic learning to improve and maintain the performance of a device implanted in an individual subject to seizures to provide treatment therapies, the computer program product comprising:
-
program instructions that assess a quality of life index that penalizes the intensity, duration and frequency of seizures and treatments over a fixed period of time;
program instructions that collect a time of unequivocal electrographic onset (UEO) in all recorded seizures over a previous fixed period of time;
program instructions that create learning sets of data based on the UEOs by clipping all the IEEG epochs immediately preceding seizures and labeling the clipped epochs as preseizure raw data;
program instructions that clip and label randomly chosen, non-overlapping data as non-preseizure or baseline raw data;
program instructions that generate a time series of all features in a feature library from the preseizure and nonpreseizure raw data;
program instructions that search for an optimal feature vector in a power set of the feature library to minimize a classifier-based performance metric;
program instructions that synthesize a posterior probability estimator for the optimal feature vector; and
program instructions that couple an optimal therapy activation threshold to the probability estimator. - View Dependent Claims (80, 81, 82, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106)
-
-
107. 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 multitherapy 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. - View Dependent Claims (108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124)
-
-
125. A system for assessing a quality of life index in an individual subject to seizures in order to adjust an implanted device in order to optimize patient-specific feature signals and treatment therapies comprising:
-
a signal acquisition component to condition and digitize a plurality of raw signals received on multiple data channels from a transducer implanted in an individual;
a first storage for accumulating the energy of raw intracranial electroencephalograms (IEEG) for the individual on multiple data channels during seizures over a fixed time period;
a second storage for accumulating the energy of a treatment control effort on the multiple data channels over all times of activation of the implanted device over a fixed time period;
a processor including;
a first logic module for weighting the accumulated energy of the IEEG and the accumulated energy of the control effort by seizure and treatment factors to determine a quality for the fixed period of time; and
a second logic module for determining a quality of life index as a weighted average of a current and previous qualities for a plurality of fixed time periods. - View Dependent Claims (126)
-
-
127. A system for periodic learning to improve and maintain the performance of a device implanted in an individual subject to seizures in providing treatment therapies comprising:
-
a signal acquisition component to condition and digitize a plurality of raw signals received from a transducer implanted in the individual;
a processor coupled to the signal acquisition component and including a learning and training module for;
assessing a quality of life index that penalizes the intensity, duration and frequency of seizures and treatments over a fixed period of time;
marking a time of unequivocal electrographic onset (UEO) in all recorded seizures over a previous fixed period of time;
creating learning sets of data based on the UEOs by clipping all the IEEG epochs immediately preceding seizures and labeling the clipped epochs as preseizure raw data;
clipping and labeling randomly chosen, non-overlapping data as non-preseizure or baseline raw data;
generating a time series of all features in a feature library from the preseizure and nonpreseizure raw data;
searching for an optimal feature vector in a power set of the feature library to minimize a classifier-based performance metric;
synthesizing a posterior probability estimator for the optimal feature vector; and
coupling an optimal therapy activation threshold to the probability estimator. - View Dependent Claims (128, 129, 130, 131, 132, 133, 134, 135, 136)
-
Specification