MixNet: Joining Force of Classical and Modern Approaches Toward the Comprehensive Pipeline in Motor Imagery EEG Classification

TL;DR

Low-density EEG MI classification results show MixNet’s superiority over state-of-the-art algorithms, offering promising implications for Internet of Thing (IoT) applications, such as lightweight and portable EEG wearable devices based on low-density montages.

Abstract

Recent advances in deep learning (DL) have significantly impacted motor imagery (MI)-based brain-computer interface (BCI) systems, enhancing the decoding of electroencephalography (EEG) signals. However, most studies struggle to identify discriminative patterns across subjects during MI tasks, limiting MI classification performance. In this article, we propose MixNet, a novel classification framework designed to overcome this limitation by utilizing spectral-spatial signals from MI data, along with a multitask learning architecture named MIN2Net, for classification. Here, the spectral-spatial signals are generated using the filter-bank common spatial patterns (FBCSPs) method on MI data. Since the multitask learning architecture is used for the classification task, the learning in each task may exhibit different generalization rates and potential overfitting across tasks. To address this issue, we implement adaptive gradient blending, simultaneously regulating multiple loss weights and adjusting the learning pace for each task based on its generalization/overfitting tendencies. Experimental results on six benchmark data sets of different data sizes demonstrate that MixNet consistently outperforms all state-of-the-art algorithms in subject-dependent and -independent settings. Finally, the low-density EEG MI classification results show that MixNet outperforms all state-of-the-art algorithms, offering promising implications for Internet of Thing (IoT) applications, such as lightweight and portable EEG wearable devices based on low-density montages.

Figures (9)

• Figure 1: Overall visualization of MixNet framework. (a) exhibits an overview of the preparation process for spectral-spatial signals. (b) demonstrates the multi-task autoencoder (AE) of MixNet, designed to simultaneously address three tasks: autoencoder, deep metric learning, and supervised learning. This framework is employed for EEG-MI classification. More in-depth insights into the network architecture can be found in \ref{['tab:model']}.
• Figure 2: The tangent slope of training and true losses at epoch $n$ and $n_0$.
• Figure 3: Frameworks of a) subject-dependent and b) subject-independent with stratified k-fold cross-validation for the classification models. Note that the pre-processing procedure comprises both downsampling and FBCSP methods.
• Figure 4: Impact of the number of training samples on binary classification performance, measured by AUC score, across two considered methods.
• Figure 5: Representation of raw and learned EEG features generated by each considered method using $t$-SNE projection. The image compares two-dimensional $t$-SNE projections designed for subject-dependent binary classification.