Meta-cognitive Multi-scale Hierarchical Reasoning for Motor Imagery Decoding

TL;DR

This work tackles the challenge of unreliable motor imagery EEG decoding due to noise and cross-subject variability. It introduces a meta-cognitive, multi-scale hierarchical framework that combines a multi-scale hierarchical signal processing (MHSP) module with an introspective uncertainty estimation (IUE) head to improve subject-independent MI classification. The MHSP component builds multi-scale temporal representations via a two-level GRU hierarchy, while IUE provides per-cycle confidence and guides iterative refinement through Monte-Carlo tree search and adaptive halting. In leave-one-subject-out evaluations on the BCI Competition IV-2a dataset, the approach yields consistent accuracy gains and reduced inter-subject variance across EEGNet, ShallowConvNet, and DeepConvNet backbones, indicating enhanced robustness for practical MI-based BCI deployment.

Abstract

Brain-computer interface (BCI) aims to decode motor intent from noninvasive neural signals to enable control of external devices, but practical deployment remains limited by noise and variability in motor imagery (MI)-based electroencephalogram (EEG) signals. This work investigates a hierarchical and meta-cognitive decoding framework for four-class MI classification. We introduce a multi-scale hierarchical signal processing module that reorganizes backbone features into temporal multi-scale representations, together with an introspective uncertainty estimation module that assigns per-cycle reliability scores and guides iterative refinement. We instantiate this framework on three standard EEG backbones (EEGNet, ShallowConvNet, and DeepConvNet) and evaluate four-class MI decoding using the BCI Competition IV-2a dataset under a subject-independent setting. Across all backbones, the proposed components improve average classification accuracy and reduce inter-subject variance compared to the corresponding baselines, indicating increased robustness to subject heterogeneity and noisy trials. These results suggest that combining hierarchical multi-scale processing with introspective confidence estimation can enhance the reliability of MI-based BCI systems.

Figures (1)

• Figure 1: Overview of the proposed framework. Backbone features are partitioned into adaptive windows and patches. A low--level GRU with top--down gating summarizes patches per window, a high--level GRU integrates across windows to yield per--cycle logits and a halting score.