“…Additionally, Zhang et al (2020) proposed a deep learning framework that incorporates convolutional and recurrent neural networks. EEG-based BCI applications commonly rely on four main types of neurophysiological patterns, namely, steady-state visual evoked potential (SSVEP) (Autthasan et al, 2020;Kwak and Lee, 2020;Rivera-Flor et al, 2022;Zhang et al, 2022;Chailloux Peguero et al, 2023;Yan et al, 2023), event-related potential (ERP) (Cecotti and Graeser, 2011;Zou et al, 2016;Li et al, 2020), movement-related cortical potentials (MRCPs) (Xu et al, 2014;Jeong et al, 2020), and motor imagery (MI) (Siuly and Li, 2012;Higashi and Tanaka, 2013;Edelman et al, 2016;He et al, 2016;Chaisaen et al, 2020;Wu et al, 2020;Gaur et al, 2021;Ma et al, 2022;Fan et al, 2023;Zhang et al, 2023). Among these EEG applications, MI has garnered increasing attention within BCI systems due to its ability to elicit oscillatory neural activity in specific frequency bands over the motor cortex region without external stimuli.…”