ThinkNet: Compact Architecture Selection and Validation-Gated Ensembles for Subject-Independent MI-EEG Decoding

2026-09-27

AI Summary

This study addresses the insufficient generalization and inflated evaluation performance in subject-independent MI-EEG decoding caused by data and computational constraints. We propose a validation control framework coupled with a gated inference strategy. Methodologically, test information is strictly isolated to prevent data leakage, while train-only normalization, evolutionary architecture search, and gated ensembling are integrated to select compact models, further accelerated via CUDA-based inference. Experimental results demonstrate that a lightweight model with merely 4.9K parameters achieves 44.35% accuracy within 0.99 ms of inference time, outperforming larger architectures. This work reveals that validation reliability constitutes a critical bottleneck under subject shift, offering an efficient decoding paradigm for resource-constrained scenarios.

Abstract

Practical assistive and rehabilitative brain--computer interfaces require subject-independent motor-imagery EEG (MI-EEG) decoders that generalize to new users under limited target-user data and constrained compute. However, held-out-subject performance can be overstated when test-subject information influences preprocessing, model selection, or ensemble selection. We present \textit{ThinkNet}, a validation-controlled framework that combines train-only normalization, validation-guided evolutionary search, and validation-gated inference to identify compact decoders and inference policies for held-out subjects. We evaluate four-class BCI Competition IV-2a (session T) decoding with nine Leave-One-Subject-Out (LOSO) folds, three seeds, seven fixed decoder entries, and a broader search over ten representative decoder families; the held-out subject is never used for normalization, hyperparameter, architecture, or ensemble-policy selection. In the fixed benchmark, the validation-selected compact decoder achieved 44.35$\pm$15.41\% accuracy with 4.9K parameters, 19 KB FP32 weights, and 0.99 ms batch-1 Orin CUDA inference. Across the broader search, compact models ($\leq$25K parameters) achieved higher mean held-out accuracy than mid-size and large alternatives after selected retraining (40.10\% vs. 35.09\% and 34.78\%). Validation-gated ensembling improved over validation-selected single-model inference, reaching 43.98$\pm$16.25\% in the fixed benchmark and 43.31$\pm$15.88\% for the compact six-family ensemble. A non-deployable oracle analysis revealed a 6.1-point family-selection gap and near-zero validation--test correlation, showing that validation reliability remains a key bottleneck under subject shift. Thus, ThinkNet is a validation-controlled framework for compact MI-EEG model and inference-policy selection, rather than a single-architecture benchmark.

Problem

• Subject-independent MI-EEG decoding
• Brain-computer interface
• Compact architecture
• Cross-subject generalization
• Validation reliability

Innovation

• Subject-Independent MI-EEG Decoding
• Compact Architecture Search
• Validation-Gated Ensembles
• Evolutionary Search
• Leave-One-Subject-Out