Deep Learning-Based EEG Mental State Classification to Support Mental Focus in Female Cricketers

Full description

Mental focus is critical for optimal performance in cricket, especially for female athletes who often encounter unique psychological pressures and limited mental conditioning support. This study bridges this gap by combining Electroencephalography (EEG) with deep learning to classify three mental states, calm, focus, and neutral, aimed at enhancing personalized training protocols. EEG data were recorded using the portable Muse 2 headband during targeted mental tasks, and raw signals were transformed into wavelet-based time-frequency images. These images were then analyzed using pretrained convolutional neural networks, including AlexNet, ResNet, and XceptionNet. AlexNet achieved the highest classification accuracy of 99.34%, an Average of 92% with cross-validation, demonstrating the effectiveness of transfer learning and data augmentation for EEG-based mental state recognition. The findings provide actionable insights for coaches and athletes to tailor mental training and highlight the potential for real[1]time feedback systems that assist cricketers in managing their cognitive states under pressure. Future work will focus on integrating these models into wear[1]able technologies for in-game mental monitoring and adaptive coaching. This research underscores the promise of EEG-driven, data-informed approaches in modern sports training, advocating for improved mental preparation and support, tailored to female athletes.