abstract
- This study explores the feasibility of using machine learning algorithms to distinguish between young and older adults based on electroencephalographic (EEG) signals acquired during a cue-based motor imagery mental task. Event-related potential (ERP) analysis revealed significant latency and amplitude differences in the N200 and P300 peaks between the two groups, with older adults exhibiting delayed and lower amplitude peaks. Based on these results, time-domain features extracted from the EEG signals and three machine learning classification models, support vector machine with linear kernel (SVML), support vector machine with radial basis function kernel (SVMR), and extreme gradient oosting (XGB) were used to evaluate the bi-class classification between young and older adults. The results indicate that it is possible to differentiate between age groups with accuracies above the empirical chance level, although there is a slight tendency to misclassify young adults as older adults. Future work includes investigating group differences in the oscillatory responses and exploring frequency domain and spatial patterns as features, along with deep learning models, to improve classification performance. This study shows age-related changes in neural responses and suggests that EEG signals can be used for age-related studies and applications. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.