Scalable biomarkers of Parkinson's disease: insights from mobile EEG in Peru

Mobile electroencephalography (EEG) may offer a scalable, cost-effective way to capture neural signatures of Parkinson’s disease (PD) in low-resource settings. In a cross-sectional cohort of 44 individuals with PD in Lima, Peru evaluated “off-medication”, we recorded EEG data using a 32-channel wireless system Emotiv Flex2 saline EEG with integrated gyroscope and collected standard clinical measures including motor (UPDRS-III) and cognitive (MoCA) evaluations. We quantified relative band power and tested associations with clinical outcomes using partial Spearman correlations. Reduced left frontal low beta (18–22 Hz) power correlated with worse motor severity (ρ = −0.539, P < 0.001) and lower cognitive performance (MoCA ρ = 0.430, P = 0.005), including a MoCA sub-score emphasizing attention/executive function (ρ = 0.650, P < 0.001). Decreased left parietal gamma (30–40 Hz) power also correlated with UPDRS-III (ρ = −0.618, P < 0.0001). We also explored the possibility to assess fall risk using this device. Linear discriminant analysis with leave-one-out cross-validation achieved 97% accuracy (AUC = 0.95) using combined EEG-gyroscope features versus 78% with only EEG features and 72% with only gyroscope features. These results suggest mobile EEG may identify clinically meaningful PD biomarkers and, when paired with inertial features, may support fall-risk assessment.