Adversarial Spatio-Temporal Attention Networks for Epileptic Seizure Forecasting

TL;DR

This paper introduces STAN, an adversarial spatio-temporal attention network that effectively forecasts epileptic seizures by modeling brain connectivity and neural dynamics, achieving state-of-the-art accuracy on EEG datasets.

Contribution

The paper presents a novel unified cascaded attention architecture with adversarial training for seizure forecasting, capturing bidirectional spatio-temporal dependencies without assuming fixed preictal durations.

Findings

Achieves 96.6% sensitivity with 0.011 false alarms/hour on scalp EEG data.

Detects preictal states 15-45 minutes before seizure onset.

Maintains real-time computational efficiency suitable for edge deployment.

Abstract

Forecasting epileptic seizures from multivariate EEG signals represents a critical challenge in healthcare time series prediction, requiring high sensitivity, low false alarm rates, and subject-specific adaptability. We present STAN, an Adversarial Spatio-Temporal Attention Network that jointly models spatial brain connectivity and temporal neural dynamics through cascaded attention blocks with alternating spatial and temporal modules. Unlike existing approaches that assume fixed preictal durations or separately process spatial and temporal features, STAN captures bidirectional dependencies between spatial and temporal patterns through a unified cascaded architecture. Adversarial training with gradient penalty enables robust discrimination between interictal and preictal states learned from clearly defined 15-minute preictal windows. Continuous 90-minute pre-seizure monitoring reveals that the learned spatio-temporal attention patterns enable early detection: reliable alarms trigger at subject-specific times (typically 15-45 minutes before onset), reflecting the model's capacity to capture subtle preictal dynamics without requiring individualized training. Experiments on two benchmark EEG datasets (CHB-MIT scalp: 8 subjects, 46 events; MSSM intracranial: 4 subjects, 14 events) demonstrate state-of-the-art performance: 96.6% sensitivity with 0.011 false detections per hour and 94.2% sensitivity with 0.063 false detections per hour, respectively, while maintaining computational efficiency (2.3M parameters, 45 ms latency, 180 MB memory) for real-time edge deployment. Beyond epilepsy, the proposed framework provides a general paradigm for spatio-temporal forecasting in healthcare and other time series domains where individual heterogeneity and interpretability are crucial.