RECTOR: Masked Region-Channel-Temporal Modeling for Affective and Cognitive Representation Learning

2026-06-13

AI Summary

This work addresses the challenge of constructing robust clinical representations from electroencephalography (EEG) and stereotactic EEG (sEEG) data, where affective and cognitive impairments manifest as dynamic network alterations spanning brain regions, channels, and time. To overcome the limitations of fixed anatomical priors, the authors propose RECTOR, a self-supervised framework that jointly models the region–channel–time triad. RECTOR introduces adaptive functional parcellation to generate dynamic brain region structures and integrates masked topological prediction, hierarchical block-sparse self-attention, and cross-view consistency constraints for unified representation learning. The method achieves new state-of-the-art performance in EEG-based emotion recognition and sEEG-based task engagement classification, while demonstrating strong robustness to missing channels, generalizability across electrode montages, and compatibility with large-scale pretraining on heterogeneous EEG datasets.

Abstract

Affective and cognitive disorders manifest as distributed, time-varying brain network dynamics across regions, channels, and time, challenging robust representation learning from EEG/sEEG for clinical diagnosis. We propose RECTOR (Masked Region-Channel-Temporal Modeling), an end-to-end self-supervised framework that unifies joint region-channel-temporal representation learning beyond fixed anatomical priors. At its core, RECTOR-SA is a hierarchical, block-sparse self-attention induced by Adaptive Functional Partitioning that evolves region structures from static anatomical definitions to adaptive functional regions. The self-supervision is driven by Masked Topology and Representation Learning, which jointly optimizes three complementary objectives: Masked Predictive Modeling, Topological Structure Modeling, and Cross-View Consistency. Across diverse benchmarks, RECTOR sets a new state-of-the-art in EEG emotion recognition and sEEG task-engagement classification. Crucially, its strong robustness to missing channels and cross-montage generalization underscores its potential for large-scale pre-training on heterogeneous EEG/sEEG, providing interpretable insights at both region and channel levels.

Problem

• affective disorders
• cognitive disorders
• EEG representation learning
• brain network dynamics
• clinical diagnosis

Innovation

• Masked Region-Channel-Temporal Modeling
• Adaptive Functional Partitioning
• Self-supervised Learning
• Block-sparse Self-attention
• Cross-montage Generalization