System and method for deep learning for tracking cortical spreading depression using EEG

Inventors

Chamanzar, Alireza • Liu, Xujin • Jiang, Lavender Y. • Vogt, Kimon A. • Moura, José M.F. • Grover, Pulkit

Abstract

Disclosed herein is a system and method implementing an automated, generalizable model for tracking cortical spreading depressions using EEG. The model comprises convolutional neural networks and graph neural networks to leverage both the spatial and the temporal properties of CSDs in the detection. The trained model is generalizable to different head models such that it can be applied to new patients without re-training. Further, the model is scalable to different densities of EEG electrodes, even when trained on a specific electrode density.

Core Innovation

The invention provides a method of detecting cortical spreading depression (CSD) waves from EEG signals. A computer coupled to an EEG machine receives a time series of readings taken from a series of time windows from a plurality of EEG electrodes. Trained deep learning models executing on the computer extract temporal features for each EEG electrode from the time series of readings from that electrode.

The method further uses a graph neural network executing on the computer, where the graph neural network has trained attention layers to aggregate the extracted temporal features to a node in the graph neural network to extract spatial information. Each node represents a physical location of brain activity as determined from one or more of the plurality of EEG electrodes. In this way, temporal electrode features are combined with spatial information derived from electrode locations in a graph form.

Classifiers executing on the computer provide a binary value for each electrode for each time window based on a probability threshold representing the presence or absence of a CSD wavefront at each respective electrode. The binary values are stitched together using a sliding time window, producing a temporal binary value based on a number of binary values representing the presence of CSD wavefronts within the sliding time window. A CSD episode is determined to have occurred by temporal binary values indicating the presence of the CSD wavefronts in consecutive sliding time windows.

Claims Coverage

Independent claim clm-00001 covers a full pipeline for detecting CSD waves from EEG: electrode-wise temporal feature extraction using trained deep learning models, spatial aggregation using a graph neural network with attention, per-electrode binary classification using probability thresholding, and CSD episode determination using sliding-window stitching across consecutive time windows. One independent claim is explicitly present in the provided claim text.

Overall claim coverage centers on detecting CSD waves by combining per-electrode temporal feature extraction, spatial aggregation via an attention-based graph neural network mapping electrode-derived information into graph nodes representing physical brain activity locations, and per-time-window probability-threshold classification with post-processing stitching across sliding time windows.

Stated Advantages

Documented Applications