Artificial Intelligence Approaches for EEG Signal Acquisition and Processing in Lower-Limb Motor Imagery: A Systematic Review.

BACKGROUND

Motor imagery (MI) is defined as the cognitive ability to simulate motor movements while suppressing muscular activity. The electroencephalographic (EEG) signals associated with lower limb MI have become essential in brain-computer interface (BCI) research aimed at assisting individuals with motor disabilities.

OBJECTIVE

This systematic review aims to evaluate methodologies for acquiring and processing EEG signals within brain-computer interface (BCI) applications to accurately identify lower limb MI.

METHODS

A systematic search in Scopus and IEEE Xplore identified 287 records on EEG-based lower-limb MI using artificial intelligence. Following PRISMA guidelines (non-registered), 35 studies met the inclusion criteria after screening and full-text review.

RESULTS

Among the selected studies, 85% applied machine or deep learning classifiers such as SVM, CNN, and LSTM, while 65% incorporated multimodal fusion strategies, and 50% implemented decomposition algorithms. These methods improved classification accuracy, signal interpretability, and real-time application potential. Nonetheless, methodological variability and a lack of standardization persist across studies, posing barriers to clinical implementation.

CONCLUSIONS

AI-based EEG analysis effectively decodes lower-limb motor imagery. Future efforts should focus on harmonizing methods, standardizing datasets, and developing portable systems to improve neurorehabilitation outcomes. This review provides a foundation for advancing MI-based BCIs.