Towards task-independent workload classification: Shifting from binary to continuous classification

Passive Brain-Computer-Interfaces provide a promising approach to the continuous measurement of mental workload in realistic scenarios. Typically, a BCI is calibrated to discriminate between different levels of workload induced by a specific task. However, workload in realistic scenarios is typically a result of a mixture of different tasks. Here, we present a study on investigating the possibility of a task-independent classifier, which can be applied to classify mental workload induced by various tasks (including n-back, backward span, addition, word recovery and mental rotation). Furthermore, our approach is not limited to binary classification of workload but can discriminate it on a continuous metric