Pre-movement neural population activity in human motor cortex reflects the subsequent outcome of futsal free kicks

Recent sports science studies suggest that optimizing neural activity can enhance motor performance, but practical applications have been hindered by large-scale electroencephalography (EEG) setups and controlled laboratory protocols. In our study, we demonstrate that a portable 3-channel EEG device can predict trial-by-trial success in a futsal-based kicking task under conditions mimicking real-world sports. We employed a model that used a transformer-based encoder and multilayer perceptron decoder, with evaluations performed separately on all electrodes. The model achieved a mean score of approximately 0.68 under cross-validation. We found that significantly higher alpha– and beta-band power contributed to the classification. Our findings indicate that performance prediction is feasible using low-density EEG in a real-world sports scenario. This research provides practical insights that could lead to the development of new training methods to improve performance through the control of brain states.