Jozef Crosland

BE (Hons)

PhD Student

New Zealand Brain Research Institute, Christchurch

PhD Student

Department of Medicine, University of Otago, Christchurch

Working Thesis Title:  “Real-time EEG-based detection and prediction of microsleeps from deep neuroelectric features in the brain”

 

Lapses of responsiveness – comprising microsleeps and attention lapses – are complete losses of responsiveness for ~0.5-15 s. They can result in injurious or fatal accidents, especially in transport sectors (pilots, air-traffic controllers, truck and car drivers, train drivers, etc.). The Christchurch Neurotechnology Research Programme (NeuroTech™) is a world leader in lapse research in terms of behavioural and physiological characterization and EEG-based detection of microsleeps and of determining underlying mechanisms in the brain.

 

Earlier NeuroTech projects have (i) used fMRI to identify regions with changes in brain activity associated with microsleeps and (ii) investigated and improved techniques for zooming in on deep neuroelectric activity, specifically related to microsleeps. This project aims to use a combination of these to achieve substantial improvements in detection/ prediction of microsleeps from scalp EEG. This will involve exploring and improving techniques for fMRI-guided enhancement of deep neuroelectric activity (source space) to identify new features for feeding into a classification system, on their own or with scalp-EEG features (sensor space). The classification system will be trained with, and tested against, behavioural gold standards from NeuroTech’s previous studies. If the new deep features are proven to be of considerable value, it will then be important to derive these features in real-time for implementing real-world environments