More accessible version
ICS > News > Lectures  Site Map.Search.Help.GreekEnglish
.Printer friendly version
 Institute of Computer Science

Lecture

Development and assessment of annotated databases in actual driving conditions
for the assessment of sleep related disorders

Speaker:
Nicos Maglaveras
PhD, Associate Professor, Aristotle University of Thessaloniki (AUT). Director of the Medical Informatics Lab, Medical School (AUT).
Date:
Monday, 19 June 2006
Time:
16:00 - 17:30
Location:
"Stelios Orphanoudakis" Seminar Room, FORTH. Heraklion, Crete
Host:
Franco Chiarugi, Ioannis Tollis

Abstract:
Sleep loss and disturbed sleep can result in impaired performance. Sleep deprivation can reduce attention and vigilance by 50%, decision-making ability, communication skills, and memory. The most sensitive tasks are those, which are long and monotonous, such as driving, which become very vulnerable to the effects of sleep deprivation.
Studies have affirmed that sleep-deprived drivers are just as dangerous as drunk drivers. It has been shown that people who drive after being awake for 17 to 19 hours performed worse than those with a blood alcohol level of .05 percent. Driver sleepiness due to sleep deprivation is a causative factor in 1% to 3% of all motor vehicle crashes. Reducing the extent of the drowsy driver problem is critical to improving the safety of EU highways. Recent studies have shown the importance of developing driver fatigue countermeasure devices to help prevent driving accidents and errors. Although numerous physiological indicators are available to describe an individual’s level of alertness, the EEG signal has been shown to be one of the most predictive and reliable as a direct measure of brain activity. Recent EEG studies in drivers attempt to describe the development of an EEG based fatigue countermeasure algorithm, and test the reliability of this algorithm to detect different phases of fatigue in ‘offline data analysis’ mode. However, these studies were performed in driver simulators, and could not assess the effects of the real driving environmental conditions.
In the present study, twenty subjects with average driving experience participated. Subjects stayed awake for at least 24 hours before the experiment. Upon arrival the subjects’ level of sleepiness was estimated by using the Karolinska Sleepiness Scale (KSS) test. The subjects were seated on the driver seat and the attached electrodes were connected to an ambulatory polysomnography monitoring system. An experienced driving instructor was seated at the co-driver’s seat. During the experiment, 8-channel EEG, EOG, ECG, EMG, and driving behaviour data were collected.
EEG recordings were firstly band-pass filtered, and then the Infomax Independent Component Analysis (ICA) technique was used to remove artefacts stemming from eye-blinks and muscle noise, especially in the frontal EEG channels (Fp1, Fp2). ICA decomposition was performed on EEG+EOG measurements. Components contaminated by artefacts were rejected, and the remained components mixed and projected back onto the scalp-channels. Two experts assessed visually, the EEG data that were collected. The annotation of the EEG data revealed an increase of slowing activity during driving events (i.e. driving errors, stepping on the opposite road lane), and an acute increase of the alpha waves just before the event. The most important finding of this study is the synchronization among the EEG channels that was observed during the driving event. Two basic synchronization patterns were recognized: a fast one and a slow one among the EEG channels. From the EEG data that were collected, we estimate the Relative Band Ratio (RBR) of the EEG frequency bands, the Shannon entropy, and the Kullback-Leibler (KL) entropy for each segment. Moreover, time-frequency analysis was performed. Analysis revealed a significant increase of theta waves RBV and a decrease of alpha and gamma waves RBV during the driving events with the decrease of the alpha waves RBV being opposite to the changes in these waves in simulated driving experiments. Moreover, Shannon and KL entropy revealed a significant decrease during the driving events.
Conclusively, EEG can assess effectively the brain activity alterations that occur during sleeping/drowsy events in driving. An effective quantitative EEG measure that assesses the synchronization among the EEG channels that was visually observed by the medical experts should be developed and tested in the near future.
 
Bio:
Nicos Maglaveras B.Eng.`82 Arist. Univ. of Thessaloniki, Greece, MSc `85, PhD `88 Northwestern Univ, IL, USA in Electrical Engineering with emphasis in Biomedical Engineering. In 1990 he joined the faculty of the medical school in the Lab of Medical Informatics, Aristotelian University, Thessaloniki, Greece where he is currently an Associate Professor of medical informatics.
He has performed research & development in nonlinear biological systems simulation, cardiac electrophysiology, medical expert systems, ECG/EEG analysis, physiological mapping techniques, parallel processing, medical imaging, medical informatics, health telematics and neural networks. He has also developed graduate and undergraduate courses in the areas of medical informatics, computer architecture and programming, biomedical signal processing and biological systems simulation. He has contributed in more than 150 publications in refereed international journals, books and conferences in the above mentioned areas, as well as in a number of white papers such as the healthgrid white paper and the EAMBES white paper on research directives for the FP7.
He has served as a reviewer in the CEC AIM technical reviews in the past, and in a number of international journals. Currently he is a member of the editorial board of the Methods of Information in Medicine Journal, and an associate editor in the IEEE Transactions on Information Technologies in Biomedicine. Nicos Maglaveras has participated in more than 20 Greek national research projects, the HEALTH TELEMATICS, the LEONARDO, the TMR, IST and the ESPRIT programmes of the CEC dealing mainly with biomedical informatics, computer patient records, and medical information processing and management. He has served as the scientific coordinator of the CEC funded IST projects ‘Distance information technologies for home care. The citizen health system (CHS)’ and of the IST project with acronym PANACEIA-iTV.
He has served as the chairman of the Computers in Cardiology 2003 conference, the ISBMDA 2006 conference and the pHealth 2007 workshop. He has also served as the vice-chair of the MIE 1997, and has served as a member of the scientific program committee in over 10 international conferences held in Greece or abroad. He is currently the chair of the societies division of the European Alliance for Medical and Biological Engineering and Sciences (EAMBES). Prof. Maglaveras has been a member of the IEEE, the EAMBES, the Greek Technical Chamber, the New York Academy of Sciences, the CEN/TC251-WG5 and the Eta Kappa Nu and he is a fluent speaker of Greek, English and French.