Towards Mental Workload Time Series Classification and Interpretation for Real-Time Feedback in Brain-Computer Interfacing Video Games

Mhalgi, Shreya Milind

Etd

Towards Mental Workload Time Series Classification and Interpretation for Real-Time Feedback in Brain-Computer Interfacing Video Games

Public Deposited

Analytics

open in viewer

One of the important factors contributing to the creation of engaging and pleasurable video game experiences is immersion. New Brain-Computer Interfaces (BCIs) enable the computer interfaces to engage with the player’s mind such as detecting the player’s real-time mental status (high/low mental workload) and using it as a real-time input for an immersive game experience. First, this research aims to find the most suitable classifier with the objective of classifying high and low mental workload by tuning numerous machine learning algorithms on the mental workload time series (fNIRS) dataset using our customized classification tool NaML. Second, this research aims to explore high and low workload intensity clusters using our novel brain signal exploration tool BrainEx to enhance our understanding of the dataset. Contributions to this research introduce a dashboard, NeuroHub, which will enable researchers who do not have an extensive Computer Science or Data Science background to conduct data parsing, data mining, and machine learning efficiently on the Functional Near-Infrared Spectroscopy (fNIRS) data. The results from this research lay a foundation for using exploratory insights to improve the classification of time series brain data.

Last modified