Abstract
This study is designed to test the hypothesis of whether writing and typing can be detected as different patterns within the same cognitive task. We designed this pilot study with five frequently-conducted learning-related tasks. We used the four electrode Muse Headset. Sixteen healthy subjects participated in this six-session experiment. In each session, we instructed them to conduct five different one-minute tasks, including reading, copying by writing, copying by typing, answering a question by writing and answering a question by typing. We compared the performance of classifiers of different categories within the same context, including the same users, the same feature extraction approach, and the same training and testing split. Most of the machine learning and deep learning algorithms could correctly classify the five tasks (20% by chance), the best algorithms achieved an accuracy for individual subjects of up to 70% for within session training and 44% for between session training.
