CT-MIFNet: Convolutional transformer-based multi-view interaction and fusion network for EEG decoding

Yibo Xiong

doi:10.1016/j.bspc.2025.108421

CT-MIFNet: Convolutional transformer-based multi-view interaction and fusion network for EEG decoding

2025-08-08

0

PRE

AI

DOI：10.1016/j.bspc.2025.108421

Original

Original request for help

Save

Abstract

En 中文

• A novel parallel dual-branch convolutional Transformer-based multi-view interaction and fusion network is proposed for EEG decoding. • A cross-covariance attention mechanism is introduced to facilitate feature interaction and fusion across diverse perspectives while reducing computational load. • Extensive comparisons with the state-of-the-art models on three datasets with two distinct BCI paradigms demonstrate the generalization and robustness of the proposed model. • T-distributed stochastic neighbor embedding and heatmap are employed to enhance the interpretability of the proposed model.

Keywords:

EEG decoding

convolutional Transformer

multi-view interaction

cross-covariance attention

brain-computer interface

CT-MIFNet: Convolutional transformer-based multi-view interaction and fusion network for EEG decoding

Abstract

Journal

Organization