Transfer Blocks Method on Multi-degrees Mental Workload Assessment with EEG.

Mental workload (MW) could be described as the cognitive resource that the human required to perform a specific task. An appropriate MW could increase the task performance, however, mental overload or underload would cause adverse effect. This paper recruited sixteen subjects in the experiment under four degrees workload tasks and Electroencephalogram (EEG) signals were recorded. Furthermore, in this work, the multi-degrees mental workload assessment was performed using Shannon entropy and power spectral density (PSD) with theta (4-7 Hz), alpha (8-13 Hz), betal (14-20 Hz) and beta2 (20-30 Hz) bands. Afterwards, the exploration of cross-block classification with transfer blocks was conducted. The results revealed that the energy of theta, betal and beta2 bands increased as MW degrees increased, while was obvious in theta band, and the multi-degrees mental workload assessment achieved an accuracy of 80% f 7.6% using SVM model. For cross-block classification, the Transfer Blocks method increased 23% accuracy for two-degrees mental workload assessment in comparison with the accuracy achieved by directly cross blocks method. It was concluded that the proposed Transfer Blocks method has better classification performance for mental workload assessment during cross blocks condition.