Back to results list
Please use this identifier to cite or link to this item:
|Title:||Neurophysiological correlates of performance and fatigue in study of mental workload||Authors:||Leung, Wing-sze Ada||Keywords:||Hong Kong Polytechnic University -- Dissertations
|Issue Date:||2006||Publisher:||The Hong Kong Polytechnic University||Abstract:||Rest break schedule is a commonly used strategy to alleviate fatigue at work. However, previous studies examining rest break schedule on low demanding work task, particularly data entry, have revealed inconsistent results. These could be attributable to the task was of low demand on mental effort and the behavioral variables were not sensitive enough to reveal the anticipated differences. Based on the energetical approach proposed by Gaillard (1993) and Hockey (1997), the combination of neurophysiological variables such as energy regulation, mental effort and subjective fatigue feeling may offer new insights into the problem. This study explores the relationships between common behavioral variables and neurophysiological correlates in a 6-hour data entry task under two rest break schedules (longer and less frequent versus shorter and more frequent). The behavioral variables were task performance, subjective fatigue, and eye blink rate. The neurophysiological correlates were brain activities in terms of the spectral powers of electrocephalogram (EEG) wavebands, and mental effort in terms of heart rate variability (HRV). The study was divided into two phases. Phase I was a validation of the Chinese version of the Swedish Occupational Fatigue Inventory (SOFI-C) for measuring subjective fatigue in this study. Phase II conducted the work-simulated experiment by which both the behavioral and neurophysiological results were obtained. In Phase I, the Swedish Occupational Fatigue Inventory was translated and the content validity evaluated with respect to its five subscales describing fatigue. The Chinese version was field-tested among 104 sedentary workers. Exploratory factor analysis revealed a five-factor solution, which was comparable to that of the original English version. Cronbach's alphas for the five factors were between 0.88 and 0.95. Test-retest reliability was satisfactory, with intraclass correlations (ICC) ranging from 0.69 to 0.83. The results indicate that the SOFI-C was valid and reliable for use in this study. In Phase II, 40 university students (20 females and 20 males) with a mean age of 20.9 ± 1.3 years were randomly assigned to a longer and less frequent rest break (LM) or shorter and more frequent rest break (SM) schedule. The experimental task consisted of two 3-hour blocks with a 1-hour break in-between. The LM schedule consisted of three 50-minute sessions with 10-minute rest breaks between each. The SM schedule consisted of five 30-minute sessions with 5-minute rest breaks between each. Participants were to type two-digit numbers at maximal speed and accuracy. The fatigue of the participants was assessed using the SOFI-C at the beginning and end of the morning and afternoon sessions. The spectral power of different EEG wavebands at the left frontal (F₃), right frontal (F₄), left occipital (O₁), and right occipital (O₂) areas, the HRV (0.1 Hz component), and eye blink rate were captured at the end of each session. The LM subjects were found to show the highest speed by end of the morning session, and the SM subjects by end of the afternoon session. No significant differences in error rates were revealed. Both groups showed significant increases in SOFI-C scores across time on all subscales except the Physical Exertion subscale. There were, however, no significant differences between the two schedules. The spectral power of alpha activities at F4 was significantly higher in the LM than in the SM group throughout the task (p < 0.05). For both groups, the heart rate variability which reflected mental effort did not vary across time and differ between schedules. Eye blink rates remained steady throughout the task sessions for both groups. For the LM subjects, the speed of data entry correlated positively with beta activities (r = 0.59) at F₃ at end of the morning, but negatively with delta activities (r = -0.64) at F4 in midafternoon. In the SM schedule, the speed correlated only with alpha activities (r = 0.64 to 0.70) at O₁ at the end of the afternoon.
The SOFI-C scores correlated with EEG spectral powers in different ways. For the LM group, the Physical Discomfort and Sleepiness subscales obtained at the end of the morning session correlated negatively with alpha and beta activities (r = -0.53 to -0.64) at O₁ and O₂ at midmorning. As scores continued to increase by end of the afternoon, they correlated positively with the delta and theta activities (r = 0.53 to 0.70) at F₃ in the middle and end of the afternoon. In contrast, for the SM group, the Physical Discomfort subscale obtained at the end of the morning correlated positively with theta, alpha, and beta activities (r = 0.59 to 0.69) at O₁ only in the middle of the morning session. Delta activities, which correlated positively with Physical Discomfort and Lack of Motivation at F₃ at the end of morning session, became negatively correlated with Sleepiness and Lack of Energy (r = -0.53 to -0.54) at O₁ at the end of the afternoon. The results of this study indicate that the two different rest schedules were likely to impose different levels of mental load and processing on individual subjects. At the group level, lower alpha activities throughout the SM group suggested that a higher mental demand might have been imposed on the subjects of this group. At the individual subject level, subjects generally found the task to be more demanding after the midway point of the morning sessions. The errors they committed typing were associated with an increase in mental processing (increase fast wave) in the SM and a decrease in mental effort (increased heart rate variability) in the LM groups. The SM subjects who were more actively engaged and performed better in the data entry task elicited an earlier fatigue state than did the LM subjects. Nonetheless, the SM subjects continued to show a significant increase in their typing speed in the afternoon, whereas the LM subjects did not. Behavioral and neurophysiological variables are found to be useful for demonstrating the differential effects of long (10 minutes) versus short (5 minutes) rest breaks on facilitating an individual's performance on a data entry task. The findings reveal the inadequacy of using merely behavioral variables for studying mental workloads, particularly on tasks that demand a low level of information processing. The results further support the importance of addressing mental load, cognitive processes, and energetic state simultaneously. The implications of the differences in performance and mental workload patterns between the two rest schedules are also discussed. The present study has several implications. First, the use of neurophysiological correlates would be useful for revealing the underlying fatigue phenomenon associated with prolonged engagement in low information processing tasks like data entry. Second, effectiveness of the rest schedule was found not only dependent on the task content but also on how long the task has to be performed. Third, solely subjective fatigue rating would not be adequate for revealing the fatigue phenomenon across time and correlation of physiological data with subjective fatigue offers new insights into the phenomenon. Lastly, the results obtained shed light on the theoretical model for explaining prolonged task engagement which can be applied in future research.
|Description:||xix, 250 leaves : col. ill. ; 30 cm.
PolyU Library Call No.: [THS] LG51 .H577P RS 2006 Leung
|URI:||http://hdl.handle.net/10397/3217||Rights:||All rights reserved.|
|Appears in Collections:||Thesis|
Show full item record
Files in This Item:
|b20696796_link.htm||For PolyU Users||162 B||HTML||View/Open|
|b20696796_ir.pdf||For All Users (Non-printable)||2.96 MB||Adobe PDF||View/Open|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.