Balancing Interpretability and Performance in Motor Imagery EEG Classification: A Comparative Study of ANFIS—FBCSP—PSO and EEGNet

Achieving both accurate and interpretable classification of motor-imagery EEG remains a key challenge in braincomputer interface (BCI) research. In this paper, we compare a transparent fuzzy-reasoning approach (ANFIS—FBCSP—PSO) with a well-known deep-learning benchmark (EEGNet) using the publicly available BCI Competition IV-2a dataset. The ANFIS pipeline combines filter-bank common spatial pattern feature extraction with fuzzy IF-THEN rules optimized via particleswarm optimization, while EEGNet learns hierarchical spatialtemporal representations directly from raw EEG data. In withinsubject experiments, the fuzzy-neural model performed better (68.58% ± 13.76% accuracy,$\kappa=58.04 \% \pm 18.43$), while in cross-subject (LOSO) tests, the deep model exhibited stronger generalization (68.20% ± 12.13% accuracy,$\kappa=57.33 \% \pm 16.22)$. The study therefore provides practical guidance for selecting MIBCI systems according to the design goal: interpretability or robustness across users. Future investigations into transformerbased and hybrid neuro-symbolic frameworks are expected to further advance transparent EEG decoding.