NO FULL TEXT AVAILABLE. Access is restricted indefinitely. ----- Fatigue has been recognised as a significant factor in industrial accidents, such as transportation industries. Electroencephalography (EEG) has been found to be one of the most reliable indicators of fatigue (Artaud et al., 1994). Several indicators of fatigue can be observed from the increase in alpha and theta activities which are accompanied by a decrease of beta activities in EEG signals (Santamaria & Chiappa, 1987b; Lal & Craig, 2001).
This study attempted to identify possible algorithms derived from the changes in the brain activity that could be utilised as a means to detect fatigue in train drivers. This thesis has investigated changes in the brain activity (delta (δ), theta (θ), alpha (α), and beta (β)), and the equation (θ+α)/β activity. Participants from both train driver cohort, as well as non-professional driver cohort, were included in the study, in order to assess the changes in brain activity during a fatigue instigating monotonous driving session. This study has also explored the correlations between brain activity changes and the lifestyle and behavioural factors. Lastly, a case study of ten train drivers was presented with a possible technique proposed for fatigue detection technology.
A total of 50 male train drivers (aged 44±9.4 years) were recruited to participate in the 30-minute monotonous train driving experiment. All participants held current Rail Safety Worker Certificate (Driver) from the Department of Transport, Australia. In addition, a total of 52 non-professional drivers (36 males and 16 females, aged 27±9.4 years) were also recruited to participate in a monotonous driving session. All participants held a current driver’s license.
Simultaneous physiological recordings were obtained during the monotonous driving sessions. Two channels of EEG (frontal and temporal sites) and one channel of electrooculogram were acquired. The EEG data from the two channels were then subjected to Fast Fourier Transform (FFT) to derive the four EEG frequency components, which were delta (0–4 Hz), theta (4–8 Hz), alpha (8–13 Hz), and beta (13–35 Hz) (Rowan & Tolunsky, 2003). The EEG spectra were then sectioned into 10 equal sections in order to identify the changes in EEG activities over the period of the monotonous driving task.
Lifestyle Appraisal Questionnaire (Craig, Hancock, & Craig, 1996), Profile of Mood State (McNair, Lorr, & Doppleman, 1971), and Locus of Control of Behaviour (Craig, Franklin, & Andrews, 1984) questionnaires were administered prior to the study. Fatigue Likert scale, Fatigue questionnaire (Wessely & Powell, 1989), and State and Trait Anxiety Inventory (Spielberger, Gorsuch, Luschene, Vagg, & Jacobs, 1983) questionnaires were administered prior to and after the monotonous driving session.
The results showed that at the end of the driving session, the heart rate decreased significantly (p<0.0001), which indicated that the participants were considerably fatigued after the driving task (Malik et al., 1996).
At the end of the monotonous driving session, the delta activity showed a significant increase (p=0.04). There was a significant increase of theta activity during the driving session (p=0.002). Alpha activity also decreased significantly at the end of the monotonous driving session (p=0.00003), and a significant decrease in beta activity was also detected (p=0.003). A significant increase of activity computed using equation (θ+α)/β (p=0.01) was also found at the end of the monotonous driving session.
The results also showed that higher heart rate was positively associated with an increase in beta activity (r=0.87, p=0.01). A lower self-reported fatigue level was linked with higher beta (r=-0.78, p=0.04) and theta (r=-0.58, p=0.046) activities.
The results of the case study indicated that 40%-50% decrease in beta activity was recorded when participants were alert before the monotonous driving session and moderately fatigued at the end of the session. A 60%-70% decrease in beta activity was recorded when participants were extremely fatigued at the end of the driving session.
Future research needs to assess the reproducibility of the EEG of fatigue in train drivers and consider the development of robust fatigue countermeasure devices by combining the findings of this research with other technologies to increase the reliability or such systems.