20 results about "Seizure detection" patented technology

The invention discloses an epileptic seizure detection system based on a multi-dimensional hypergraph fusion network, which comprises a multi-modal data collector, a signal preprocessor, a feature extractor, a multi-dimensional hypergraph builder, a hypergraph feature extractor, a feature fusion device, a classifier and a time sequence prediction corrector, the multi-dimensional hypergraph constructor constructs three hypergraph structures, namely an intra-modal hypergraph, an inter-modal hypergraph and a time sequence hypergraph; the feature fusion device integrates the features extracted by the three hypergraphs to generate a comprehensive feature vector; the classifier classifies the fusion features through a two-layer feedforward neural network, and outputs probability prediction of epileptic seizure; the time sequence prediction corrector applies a majority voting smoothing filter to carry out post-processing on a classification result, isolated misclassification is eliminated, and the time sequence consistency of prediction is enhanced. The invention provides an innovative multi-dimensional hypergraph fusion network, and the robustness and accuracy of epileptic seizure detection are remarkably improved by simultaneously modeling high-order relationships in modals, between modals and time sequence dimensions.

The invention discloses an unbalanced classification method for electroencephalogram data in epilepsy detection, and relates to the technical field of data optimization and big data processing. According to the method, firstly, covariance matrixes of majority class samples and minority class samples are calculated, linear transformation is achieved through matrix decomposition, and the minority class samples inherit global distribution characteristics of the majority class samples; and then, in the transformed feature space, sorting samples based on mahalanobis distance and performing partition pairing, selecting sample pairs with large difference to generate convex combination synthesis samples, and ensuring sample diversity and boundary consistency. Experimental results show that on a CHB-MIT electroencephalogram data set, the method effectively solves the problems that a traditional oversampling technology is prone to expanding minority class decision boundaries and generated samples are lack of diversity, and the reliability of epileptic seizure detection is remarkably improved.

Embodiments described herein relate to systems, devices, and methods for monitoring brain activity and delivering electrical brain stimulation to a patient. In some embodiments, a system can deliver responsive electrical stimulation in a closed-loop manner and can offer real-time or near real-time monitoring of induced neurophysiological effects. After detecting a targeted brain pattern (i.e., an epileptic seizure), the system may deliver high-intensity ultra-short electrical stimulation impulses non-invasively or minimal-invasively to diminish or stop the neural oscillations underlying the epileptic seizure. The stimuli are delivered in time and space, relative to the emerging seizure rhythm patterns, such that they diminish or terminate the seizure. The system may include an implantable device including one or more electrodes electrically coupled to one or more processors. The processor(s) may be operatively coupled to a memory, one or more communication modules, and may optionally be coupled to a battery and one or more additional sensor(s).

Systems and method for seizure detection. A seizure detection device utilizes frequency discrimination, time series features, and machine learning clustering algorithms to distinguish between the EEG signal of those experiencing a seizure compared to those who are not experiencing a seizure.

This specification describes techniques for seizure detection in mammals. A first set of electrodes is used to stimulate a region of a brain of a mammal, and an electrical signal sensed by a second subset of electrodes is acquired during stimulation. The electrical signal is segmented into time-segmented portions according to a value of a time parameter. For each time-segmented portion of the electrical signal, a power is determined for the portion of the signal within a frequency band defined by a value of a frequency parameter. A classifier is used to classify the time-segmented portion of the signal as seizure positive or seizure negative based on the determined power within the frequency band. The values of the time and frequency parameters can result in the classifier achieving an effective level of performance defined by an area under a precision-recall curve (AU-PRC) for the classifier of at least 0.5.

We report a method of determining an occurrence of an epileptic convulsive seizure in a patient, comprising: receiving body data from a patient during a first time period, determining a work level relating to said first time period at least based partially upon said body data; determining whether said work level exceeds an extreme work level threshold; performing a responsive action, in response to a determination that said work level exceeds said extreme work level threshold. We also report a medical device system configured to implement the method. We also report a non-transitory computer readable program storage unit encoded with instructions that, when executed by a computer, perform the method.

This specification describes techniques for seizure detection in mammals. A first set of electrodes is used to stimulate a region of a brain of a mammal, and an electrical signal sensed by a second subset of electrodes is acquired during stimulation. The electrical signal is segmented into time-segmented portions according to a value of a time parameter. For each time-segmented portion of the electrical signal, a power is determined for the portion of the signal within a frequency band defined by a value of a frequency parameter. A classifier is used to classify the time-segmented portion of the signal as seizure positive or seizure negative based on the determined power within the frequency band. The values of the time and frequency parameters can result in the classifier achieving an effective level of performance defined by an area under a precision-recall curve (AU-PRC) for the classifier of at least 0.5.

Methods, systems, and apparatus for determining probabilistic measures of seizure activity (PMSA) values based on a plurality of seizure detection algorithms and / or body signals used as inputs by the seizure detection algorithms. Use of the PMSA values to detect seizure activity based on a consensus of the algorithms and / or body signals, and / or warn, log, administer a therapy, or assess the efficacy of a therapy.

The present invention relates to a brain dysfunction and seizure detector monitor and system, and a method of detecting brain dysfunction and / or seizure of a subject. Preferably, the present invention also includes one or more seizure detection algorithms. The analysis method is specifically optimized to amplify abnormal brain activity and minimize normal background activity yielding a seizure index directly related to the current presence of ictal activity in the signal. Additionally, a seizure probability index based on historical values of the aforementioned seizure index, is derived for diagnostic purposes. The seizure probability index quantifies the probability that the patient has exhibited abnormal brain activity since the beginning of the recording. These indexes can be used in the context of emergency and / or clinical situations to assess the status and well-being of a patient's brain, or can be used to automatically administer treatment to stop the seizure before clinical signs appear.

The application relates to a single-channel seizure detection device, comprising: a data acquisition unit configured to acquire single-channel electroencephalogram data to be detected and to pre-process the single-channel electroencephalogram data to be detected; an electroencephalogram feature extraction unit configured to extract time-frequency features of the pre-processed single-channel electroencephalogram data and to perform standardization processing on the extracted time-frequency features; and a seizure detection unit configured to sequentially pass the standardization-processed time-frequency features through an input layer, an Embedding layer, a feature generalization module, a global pooling layer, a 1x1 convolution layer and a Linear output layer to obtain a detection result; the application combines the light convolution calculation advantages of the partial convolution idea and the advantages of the Style-domain generalization algorithm, and further divides and processes frequency domain information to reduce the increase of calculation amount, so that the detection model enhances the cross-patient detection effect of single-channel seizure electroencephalogram and makes the network lighter, and the generalization and real-time performance of the detection model are greatly improved.

The application provides a method and system for seizure detection and automatic annotation based on electroencephalogram data, comprising the following steps: step 1, acquiring electroencephalogram records with the possibility of seizures and pre-processing; step 2, building an SC-LSTM model based on deep learning, including two parallel feature extraction modules and a classification module, for feature extraction, feature selection fusion and seizure detection and classification of electroencephalogram. The deep learning technology used in the application can automatically capture effective features from EEG signals and realize end-to-end seizure detection. The application of deep learning technology to seizure detection can reduce the workload of neurologists, improve work efficiency and improve detection accuracy.

The present disclosure provides systems and methods for seizure detection. The epileptic seizure detection method may include receiving a plurality of electroencephalogram (EEG) signals for a subject via a plurality of channels, preprocessing the plurality of EEG signals by segmenting the plurality of EEG signals for each channel into a plurality of time data segments, extracting a plurality of features from each time data segment for each channel, and applying a machine learning algorithm to the plurality of features to perform seizure binary classification for each time data segment of each channel. A control strategy can be employed to determine the epileptic seizure burden of the aggregated epileptic seizure binary classification. A notification may be generated when the seizure burden equals or exceeds a threshold. A healthcare practitioner may use the notification to assess whether the subject is likely to be at risk of having an epileptic seizure.

The invention discloses an epileptic seizure detection system based on a double-branch space-time diagram neural network. The epileptic seizure detection system comprises an electroencephalogram signal collector, a signal preprocessor, a double-branch diagram constructor, a double-branch diagram neural network feature learning device, a feature fusion device and a classifier. Wherein the double-branch graph constructor comprises a time branch and a space branch; the time branch maps a down-sampled electroencephalogram sequence into a graph structure for each channel by adopting a horizontal visibility graph algorithm; the spatial branch constructs node features based on multi-scale visual graph network statistical features, and constructs edge weights based on dispersion indexes and cross-channel phase amplitude coupling. The double-branch graph neural network feature learner automatically learns time dynamic features and space correlation features from two types of graph structures, and aggregates multi-channel information through an attention mechanism. And the feature fusion device adopts an attention mechanism to carry out weighted fusion on the double-branch features. According to the invention, the accuracy of epilepsy detection is significantly improved by fusing spatial-temporal features and graph neural network automatic learning through a double-branch architecture.

The application discloses a PICU seizure detection system and a storage medium, and a program is stored on the storage medium. When the program is executed by a processor, the following steps are implemented: real-time acquisition of electroencephalogram data, electrocardiogram data and electromyogram data of a subject with epilepsy; segmentation of the electroencephalogram data into multiple electroencephalogram data segments, and detection of each electroencephalogram data segment to obtain suspicious electroencephalogram data segments; acquisition of electrocardiogram data segments and electromyogram data segments according to the suspicious electroencephalogram data segments, and feature extraction of the suspicious electroencephalogram data segments, the electrocardiogram data segments and the electromyogram data segments to obtain electroencephalogram suspicious segment features, electrocardiogram suspicious segment features and electromyogram suspicious segment features; and detection of whether the suspicious electroencephalogram data segments are real seizure segments.

The application discloses a seizure detection system based on EEG feature distribution adaptation transfer learning, comprising: a data acquisition module, which acquires source domain electroencephalogram data and target domain electroencephalogram data; a data preprocessing module, which performs data segmentation, time-frequency decomposition, feature extraction and feature vector construction on the electroencephalogram data of the source domain and the target domain; a feature transfer learning and electroencephalogram pattern classification module, which iteratively calculates a feature space transformation matrix based on field distribution adaptation to minimize the distribution difference between the electroencephalogram feature samples of the source domain and the target domain; in the feature transfer learning process, the transformed source domain electroencephalogram feature samples are used to train a classifier to realize pattern classification of the target domain samples; the weight factor of the distribution adaptation and the feature space transformation matrix are iteratively updated to realize training of the classifier and pattern classification of the target domain electroencephalogram data and seizure detection. The problem of complex model and time-consuming training of the existing seizure detection technology is solved.

The application discloses a cross-patient seizure detection method based on a graph-enhanced pulse state space model, first, the brain electrical signal data of a patient is acquired and an electrode space topology graph is constructed; then, a pulse state space model is used to extract the time sequence features of physiological channels, and a pulse graph convolution network is used to perform message aggregation between adjacent nodes in a discrete pulse domain across time steps; finally, feature aggregation is performed through a global average pooling operation at a graph level, and a two-class prediction probability is output through a fully connected classifier. The application follows the design logic of "first time and then space", the pulse state space model models the potential brain state dynamics of the time sequence change, and realizes stable long context feature processing through the cyclic updating mode of the adaptive hardware; the pulse graph neural network completes the information transmission and fusion between channels based on the electrode correlation graph, can capture the spatial dependence relationship which is crucial for the representation of seizures, reduces the calculation energy consumption, and improves the overall performance.

Systems and methods to detect seizures using analog circuitry. One example method generally includes obtaining, at a seizure detection system, one or more electroencephalogram (EEG) signals, detecting a plurality of features associated with each of the one or more EEG signals, generating a bitstream indicating a seizure probability associated with each feature of the plurality of features to yield a plurality of bitstreams indicating a plurality of seizure probabilities, and generating a seizure detection output based on the plurality of bitstreams indicating the plurality of seizure probabilities of the plurality of features.

The invention relates to the technical field of epilepsy diagnosis, in particular to an epileptic seizure detection method and system based on dynamic and static feature fusion. The method comprises the following steps: collecting a standardized input sequence, reconstructing a brain region functional topological structure by utilizing a Copula-Graph modeling method, extracting global static space characteristics, performing state recursion modeling on a time sequence of the standardized input sequence by adopting a selective state space encoder, performing weighted integration on the sequence after state recursion through a time attention mechanism, and performing state recursion modeling on the time sequence after state recursion. The method comprises the steps that global static spatial features are obtained, dynamic path features are obtained, an OT-CoAttention dynamic and static fusion mechanism is adopted to carry out spatial domain and time domain alignment and weighted fusion on the global static spatial features and the dynamic path features, fusion features are obtained, and an epileptic seizure or non-seizure result and a confidence interval thereof are output; according to the method, space-time two-way interaction is realized, different time scale features can be adaptively captured, feature splitting is avoided, feature discrimination is improved, and the false drop rate is remarkably reduced.

The invention relates to the technical field of biomedical signal processing, in particular to an epileptic seizure detection method and system based on time-space diagram attention and characteristic distillation. The method comprises the steps of collecting a multi-channel electroencephalogram, calculating Pearson correlation coefficients among channels of the multi-channel electroencephalogram to generate an adjacency matrix, extracting a channel feature matrix based on TCN, fusing the adjacency matrix and the feature matrix to generate a graph, inputting the graph, conducting self-adaptive weighted aggregation on the channel relation through GAT, and obtaining the multi-channel electroencephalogram. Global features and local features are extracted, feature complementation and semantic alignment of the global features and the local features are carried out through a two-way knowledge distillation mechanism based on KL divergence, outputs of global branches and local branches are spliced and transmitted into a full connection layer and pooling for classification, and seizure / non-seizure judgment is output; the method is high in detection accuracy, the system is light in weight, the attack classification result, the confidence score and the attention visualization graph can be output, and the diagnosis credibility and practicability are enhanced.