A Comparative Analysis of Indexing of Mental Workload by using Neuro-Driving Tools based on EEG Measurements Coupling with the Eye-Tracking System

Abstract

Event Abstract Back to Event A Comparative Analysis of Indexing of Mental Workload by using Neuro-Driving Tools based on EEG Measurements Coupling with the Eye-Tracking System Mayu Ichiki1*, Guangyi Ai1, Jonathan S. Ooi2 and Hiroaki Wagatsuma1, 3, 4 1 Kyushu Institute of Technology, Graduate School of Life Science and Systems Engineering, Japan 2 Universiti Putra Malaysia, Malaysia 3 RIKEN BSI, Japan 4 AIRC, AIST, Japan Recently, driving with smart devices like cellphones is becoming a social problem. According to the 2015 report of American Automobile Association (AAA), 60 percent for teenage drivers crashed related to driver distractions and top three causes are interactions with passengers (15%), on the cellphone (12%) and looking at something in the vehicle (10%) [1], suggesting an importance of the measurement of cognitive workloads during the driving [2]. Strayer et al. [3-5] investigated levels of its workload called "mental workload" ranging from a single-task to the complex cognitive task such as OSPAN task to be the highest level, which is assumed to exceed the workload in Siri-based interactions. They used P300 amplitude for the criterion of EEG based indexing of the mental workload. The criterion is statistically evaluated in the Oddball paradigm in principle; however it is not suitable for an on-line evaluation in driving without guarantee of rareness as an Oddball task. As a comparative study, we investigated different indexing methods of the mental workload using EEGs such as alpha, beta and theta [6-7] (or its combinations [8-11] and phase lock value [12-15]), for applying to different conditions, in-laboratory and driving simulator (DS) and the real-vehicle environment. In the natural driving condition, transitions of multiple conditions, such as play music, read messages, destination setting by voice recognition as distractions needs to be investigated for validated analysis using EEGs and exploring of possible relations to with psychological measures obtained from the Eye-tracking system including pupil size, blink number and saccade amplitude [16-17]. Acknowledgements This work was partly supported by JSPS KAKENHI 16H01616 and the collaborative research project with FUJITSU TEN LIMITED.