A Pilot Study on the Evaluation of Brain Signal Features to Provide an Index Related to Visual Attention in Semi-Professional Table Tennis Players

For high-cognitive-load sports like table tennis, successful neurofeedback training needs accurate biomarkers for certain cognitive processes, e.g., visual attention. Yet, the majority of existing neurofeedback protocols are based on general EEG features that may not necessarily represent the required mental processes in these rapid sports. To bridge this gap, the current study explores EEG signal features related to visual attention. Therefore, in this research, EEG signals were registered in the Cz area in resting and decision-making conditions from 8 semi-professional male table tennis players (aged between 18 and 25). Features extracted were statistical parameters (mean and variance), Fourier transform mean amplitude and power, wavelet transform coefficients mean absolute value and standard deviation, and nonlinear features like entropy, fractal dimension, correlation dimension, and Lyapunov exponent. Then, correlations of these features with TOVA (Test of Variables of Attention) scores—reaction time and attention errors—were determined, and a t-test was conducted for observing pre- and post-test differences. It was observed that frequency and time-frequency features were most correlated, and their average correlation coefficients with reaction time scores were 0.84 and 0.89 respectively. In addition, frequency and time-frequency features also demonstrated the greatest changes pre- and post-test (P ≤ 0.02). These results indicate that these features can be used as good biomarkers for monitoring and improving attentional processes in sports individuals. Therefore, the use of these features in designing neurofeedback protocols for improving fast cognitive responses in sports individuals—particularly in sports such as table tennis—would not only be effective, but also feasible.