TFOC-Net: A Short-time Fourier Transform-based Deep Learning Approach for Enhancing Cross-Subject Motor Imagery Classification

2025-07-03

AI Summary

Cross-subject motor imagery (CS-MI) classification suffers from limited performance due to high inter-subject variability in EEG signals, severely hindering the practical deployment of calibration-free brain–computer interfaces (BCIs). To address this, we propose an STFT-CNN co-optimization framework: a learnable, parameterized short-time Fourier transform (STFT) enhances discriminative time-frequency representation, while a multi-temporal-window input scheme and balanced-batch training strategy improve cross-subject generalization. The method operates end-to-end on STFT spectrograms without subject-specific calibration. Evaluated on BCI Competition IV datasets 1, 2a, and 2b, it achieves accuracies of 67.60%, 65.96%, and 80.22%, respectively—significantly outperforming existing calibration-free approaches and establishing new state-of-the-art benchmarks. Additionally, we publicly release a high-quality preprocessed dataset to promote standardization and reproducibility in CS-MI research.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Learning & Optimization for CV

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Cross-subject motor imagery (CS-MI) classification in brain-computer interfaces (BCIs) is a challenging task due to the significant variability in Electroencephalography (EEG) patterns across different individuals. This variability often results in lower classification accuracy compared to subject-specific models, presenting a major barrier to developing calibration-free BCIs suitable for real-world applications. In this paper, we introduce a novel approach that significantly enhances cross-subject MI classification performance through optimized preprocessing and deep learning techniques. Our approach involves direct classification of Short-Time Fourier Transform (STFT)-transformed EEG data, optimized STFT parameters, and a balanced batching strategy during training of a Convolutional Neural Network (CNN). This approach is uniquely validated across four different datasets, including three widely-used benchmark datasets leading to substantial improvements in cross-subject classification, achieving 67.60% on the BCI Competition IV Dataset 1 (IV-1), 65.96% on Dataset 2A (IV-2A), and 80.22% on Dataset 2B (IV-2B), outperforming state-of-the-art techniques. Additionally, we systematically investigate the classification performance using MI windows ranging from the full 4-second window to 1-second windows. These results establish a new benchmark for generalizable, calibration-free MI classification in addition to contributing a robust open-access dataset to advance research in this domain.

Problem

• Enhancing cross-subject motor imagery classification accuracy
• Addressing EEG variability across individuals in BCIs
• Developing calibration-free BCIs for real-world applications

Innovation

• STFT-transformed EEG data classification
• Optimized STFT parameters for preprocessing
• Balanced batching strategy in CNN training