A portable cognitive tool for engagement level and activity identification

• Main Authors: Teo, Terry, Lye, Sun Woh, Li, Yu Fei, Zakaria, Zainuddin
• Other Authors: School of Mechanical and Aerospace Engineering
• Format: Conference or Workshop Item
• Language: English
• Published: 2024
• Subjects: Engineering; Assessment; Neurophysiology; Monitoring; EEG; Outliers
• Online Access: https://hdl.handle.net/10356/173503; https://icecce.com/
• Institution: Nanyang Technological University

Wearable devices such as Electroencephalography (EEG) hold immense potential in the monitoring and assessment of a person’s task engagement. This is especially so in remote or online sites. Research into its use in measuring an individual's cognitive state while performing task activities is therefore expected to increase. Despite the growing number of EEG research into brain functioning activities of a person, key challenges remain in adopting EEG for real-time operations. These include limited portability, long preparation time, high number of channel dimensionality, intrusiveness, as well as level of accuracy in acquiring neurological data. This paper proposes an approach using a 4-6 EEG channels to determine the cognitive states of a subject when undertaking a set of passive and active monitoring tasks of a subject. Air traffic controller (ATC) dynamic-tasks are used as a proxy. The work found that using a newly developed channel reduction and identifier algorithm, good trend adherence of 89.1% can be obtained between a commercially available BCI 14 channel Emotiv EPOC+ EEG headset and that of a carefully selected set of reduced 4-6 channels. The approach can also identify different levels of engagement activities ranging from general monitoring, ad hoc and repeated active monitoring activities involving information search, extraction, and memory activities.