Subject-Independent EEG Classification of Motor Imagery Based on Dual-Branch Feature Fusion

Abstract

A brain computer interface (BCI) system helps people with motor dysfunction interact with the external environment. With the advancement of technology, BCI systems have been applied in practice, but their practicability and usability are still greatly challenged. A large amount of calibration time is often required before BCI systems are used, which can consume the patient’s energy and easily lead to anxiety. This paper proposes a novel motion-assisted method based on a novel dual-branch multiscale auto encoder network (MSAENet) to decode human brain motion imagery intentions, while introducing a central loss function to compensate for the shortcomings of traditional classifiers that only consider inter-class differences and ignore intra-class coupling. The effectiveness of the method is validated on three datasets, namely BCIIV2a, SMR-BCI and OpenBMI, to achieve zero calibration of the MI-BCI system. The results show that our proposed network displays good results on all three datasets. In the case of subject-independence, the MSAENet outperformed the other four comparison methods on the BCIIV2a and SMR-BCI datasets, while achieving F1_score values as high as 69.34% on the OpenBMI dataset. Our method maintains better classification accuracy with a small number of parameters and short prediction times, and the method achieves zero calibration of the MI-BCI system.

Two-Phase Multitask Autoencoder-Based Deep Learning Framework for Subject-Independent EEG Motor Imagery Classification

“…Several deep learning models employing MTAEs for the classification of MI tasks have been proposed in recent literature [39], [50]. Autthasan et al [39] proposed a multitask endto-end AE (MIN2Net) to learn a latent representation with enhanced discriminative information of EEG data achieved simultaneously by optimizing three loss functions: reconstruction, cross-entropy, and triplet loss.…”

“…Autthasan et al [39] proposed a multitask endto-end AE (MIN2Net) to learn a latent representation with enhanced discriminative information of EEG data achieved simultaneously by optimizing three loss functions: reconstruction, cross-entropy, and triplet loss. Yanqing et al [50] presented a dual-branch network (MSAENet) that combines a multiscale branch with an AE. The multiscale feature extraction enables the learning of latent features using different receptive fields, while the AE extracts discriminative features to improve classification performance.…”

Exploring the Potential of Attention Mechanism-Based Deep Learning for Robust Subject-Independent Motor-Imagery Based BCIs

“…Although these models can be quite accurate, they require ongoing calibration for each new user, leading to resource-intensive implementations [15], [16]. The second approach focuses on subject-independent models, aiming to identify features that are consistent across different individuals to avoid the need for calibration [17], [18]. While this approach is less resource-intensive, it often results in compromised performance [19].…”