AAU Student Projects - visit Aalborg University's student projects portal
A master's thesis from Aalborg University
Book cover


Timely Constraints of EEG in Cognitive Distraction and Focus Detection while Driving

Authors

; ;

Term

4. term

Publication year

2018

Abstract

EEG (elektroencefalografi) kan bruges til at skelne mellem kognitivt fokus og kognitiv distraktion, men det er mindre klart, om denne forskel kan genkendes på tværs af dage. Vi udviklede Animus, et system der med EEG-data vurderer, om en bilist er mentalt fokuseret eller distraheret. I en køresimulator indsamlede vi EEG-data fra 8 deltagere over to dage med 7 dages mellemrum. Når Animus blev trænet og evalueret på data fra samme dag, opnåede systemet i gennemsnit 98,01% klassifikationsnøjagtighed på tværs af deltagerne. Når modellen blev trænet på dag 1 og evalueret på dag 2, var gennemsnittet 77,95%. Dette tyder på, at hjernesignaler er tilstrækkeligt stabile over tid til, at Animus kan skelne mellem fokus og distraktion på senere dage, om end med lavere præcision end samme dag. For at undersøge betydningen af EEG-hjelmens placering gennemførte vi et andet forsøg med to køresessioner: én hvor hjelmen blev taget af og sat på igen, og én hvor den blev siddende. Genplacering reducerede nøjagtigheden fra 84,06% til 77,11%, hvilket indikerer, at hjelmens pasform og placering påvirker resultaterne.

Electroencephalography (EEG), which measures the brain’s electrical activity via sensors on the scalp, can distinguish cognitive focus from cognitive distraction. Less is known about whether this distinction holds across days. We developed Animus, a system that uses EEG data to determine whether a driver is mentally focused or distracted. Using a driving simulator, we collected EEG data from 8 participants across two days separated by 7 days. When Animus was trained and evaluated on data from the same day, it achieved an average classification accuracy of 98.01% across participants. When trained on day 1 and evaluated on day 2, the average accuracy was 77.95%. These results suggest that brainwave patterns are consistent enough over time for Animus to distinguish focus from distraction on future days, though with reduced accuracy compared to same-day evaluation. To assess the impact of EEG helmet placement, we ran a second experiment with two driving sessions: one where the helmet was removed and reattached, and one where it stayed in place. Reattachment reduced accuracy from 84.06% to 77.11%, indicating that fit and placement of the EEG helmet affect performance.

[This abstract was generated with the help of AI]