Table 3. Summary of key methodological characteristics and findings of individual papers.
| Paper Author(s) | Country/Year | Design/objective | Sample /gender | Age range | Sleep loss | Drive setting/ drive duration | Frequency /time (s) of drive | Driving performance outcomes (specific definition) | Findings/effect size |
|---|---|---|---|---|---|---|---|---|---|
| Acute sleep loss | |||||||||
| Moderate (2 h <sleep loss <4h) and severe (4h <sleep loss <6h) | |||||||||
| Rupp, et al., 2004) | US /2004 | Repeated measures between- groups design, comparison of normal sleep with moderate and severe sleep loss (5 and 3 h sleep respectively) while doing single driving or dual driving and subtraction tasks. |
26 (13 men and 13 women) | 18–26 yrs. | Sleep loss for 3 h, and Sleep for 5 h | Simulator/30 min | Night time from 1a.m and 9 a.m., and from 3 a.m. and 9 a.m. | a) Lateral position, lane deviation (deviation in road position from lane centre), b) SD a of lateral position (lane variability), c) Lane crossings (car crosses one of lane markers), d) Lane incidents (count of times car crossed into lane edge), e) mean deviation from speed limit, f) SD of deviation from speed limit (Speed variability), g) Wind reaction time: seconds taken to correct road position in presence of wind gust |
a) No significant main effects or interactions on lane deviation (F = 0.3), b) Interaction of group and task type: greater lane variability for sleep deprived group for the dual driving and subtraction task versus the single driving task (medium effect size, Cohen’s d = 0.79). c) Interaction of group and session: greater lane variability for sleep deprived group vs the control group, Cohen’s d = 0.85 (large effect size), e) No effect on mean and SD of deviation from speed limit (F = 0.58, P > 0.05) and (F = 3.47, P > 0.05), f) Increase of 1.4 in lane crossings after sleep loss for single driving task, g) Increase of 1.6 in lane crossings after sleep loss on dual driving and subtracting task, h) Interaction of sleep condition and experimental session on lane crossings (large effect size; Cohen’s d = 0.98). |
| (Philip, et al., 2005(b)) | France /2004 | Quasi-experimental (Cross-over), comparison of 6 times of drives on real road and simulator after habitual sleeps (8 h) with those after only sleep for 2 h in 6 times of the day. | 12 men | 19–24 yrs. | Sleep for 2 h, from 11p.m. to 1 a.m.) | a) On a highway b) on a Divided Attention Steering Simulator/ 1.5 h | 6 times/ day between 9 a.m. to 9:30 p.m. | a) Lateral position (car distance from lateral lanes), b) Line crossing (car crosses one of lane markers). |
a) Lateral position not reported, b) Main effect of sleep loss (8- fold increase) on inappropriate line crossings (F1,10 = 60.013, P < .001) On the simulator, c) No effect of time of day on lane crossing (F5,50 = 1.274, P = 0.301) on the simulator, d) On real road, an increase of 8 cases in line crossing compared with no line crossing in rested condition. |
| (Philip, et al., 2005(a)) | France /2005 | Randomized open cross-over design, comparison of 5 times of real driving after habitual sleep (8 h) with those after only sleep for 2 hours | 22 men | 18–24 yrs. | Sleep for 2 hours, from 11 p.m. to 1 a.m.) | Open Highway/1.5 h | 5 times/ day from 9 a.m. to 7:30 p.m. | a) Lateral position (car distance from lateral lanes), b) Lane deviation (mean lateral deviation from centre of the road), c) Line crossing (car crosses one of lane markers). |
a) Lateral position not reported, b) Lane deviation not reported, c) Cumulative number of line crossing after 5 times of drives Increased from 66 cases (rested) to 535 cases (sleep-restricted); (incidence rate ratio (IRR): 8.1(95% CI): 3.2–20.5; p < 0.001). |
| (Lowden, et al., 2009) | Sweden /2008 | Repeated measures Between-subjects design, Comparison of performance of young and elderly drivers after extended wake time. |
10 young (5 male, 5 female), 10 elderlies (5 male, 5 female). | 18–24 yrs., 55–64 yrs. |
5.5 h (extended wake) and 2 h sleep. | Hi-Fi moving base simulator/1.5 h | Single drive either in afternoon or at night time. | a) Lateral position (perpendicular distance between the right side of the right front wheel and the left side of the right-hand lane boarder), b) SD of lateral position, c) Mean and SD of speed, |
a) Lateral position not reported, b) Mean and SD of speed not reported, c) SD of lateral position increased from the 30th minutes of drive onwards. |
| (Sagaspeb, et al., 2008), (Verster, et al., 2011) | France /2008 | Quasi-experimental (Cross-over), comparison effects of 2, 4 and 8 h sleep loss (extended wake) | 14 men | 21–25 yrs. | 2, 4 and 8 h sleep loss (extended wake) | Two- lane highway/2 h, 4 h, 8 h |
Reference session (9–10 p.m.), Midnight at the wake time | Inappropriate line crossing, lane crossing (car crosses one of lateral lane markers) | The incidence rate ratios of inappropriate line crossings, compared to the reference session (9–10 p.m.), were 6.0 (95% CI, 2.3 to 15.5; P,.001), 15.4 (CI, 4.6 to 51.5; P,.001) and 24.3 (CI, 7.4 to 79.5; P,.001), for 2 h, 4h and 8 h driving durations respectively. |
| (Filtness, et al., 2012) | UK /2011 | Repeated measures between- subjects, comparing of sleep in normal and restricted to 5 h (extended wake) conditions among young and elderly drivers | 20 young men, 20 old men | 20–26 yrs., 52–74 yrs. |
Sleep loss (extended wake) for 3 h | Immobile car with a computer- generated road projection/1.5 h | afternoon | Lane crossings; lane departure (all four wheels came out of the driving lane) | Interaction of sleep condition and age group; lane crossing increased in the last 30 min of the drive in both young and old drivers, with more impairment in young drivers. |
| (Anderson & Horne, 2013) | UK /2012 | Repeated measures within-subjects design, comparison of driving after normal sleep and extended wake | 8 men | 20–26 yrs. | Sleep loss (extended wake) for 3 h, | Immobile car with a computer-generated road projection/2 h | Afternoon at 2 p.m. | a) Driving incidents; lane crossings (when at least two wheels of the vehicle leaving the carriageway), b) Distraction |
a) positive correlation between number of distractions and number of lane crossings under sleep restriction (large effect size r = 0.74); b) from 2308 distractions under sleep deprivation 474 distraction directly resulted in incidents (t = 2.73; df = 7; p < 0.03) |
| Total sleep loss (6 <sleep loss <8) | |||||||||
| (Pizza, et alc., 2004), | Italy /2004 | Repeated measures within-subjects design, comparison of normal sleep and sleep deprivation. | 10 (5 men and 5 women) | Mean 24.9 (± 0.6) yrs. | One night total sleep loss | STISIM 300 Driving Simulator/30 min | 4 times/day between morning and afternoon | a) Lateral position (distance from the car to the left lane marker), b) Mean lane position (mean distance from lane centre), c) Lane position variability; SD of lateral position (deviation in distance from the car to the midline), d) Number of crashes, e) Mean and SD of speed, f) Deviation from the speed limit. |
a) Lateral position not reported, b) No change in mean lane position across drives (χ2 = 0.99), c) Increase of lane position variability with the highest of 0.20 at 2 p.m. (χ2 = 0.003, p<0.05), d) Increased of number of crashes with the highest of 0.8 at 2 p.m., e) No change in mean and SD of speed, even worsening of mean speed (χ2 = 0.98, p > 0.05) and the SD of speed (χ2 = 0.21, p > 0.05). |
| (Contardi, et al., 2004) | Italy /2004 | Repeated measures within-subjects design, Comparison normal sleep and sleep deprivation. |
10 (5 men and 5 women) | Mean 24.9 (± 0.6) yrs. | One night total sleep loss | STISIM 300 Driving Simulator/30 min | 4 times/day between morning and afternoon | a) Lateral position, b) SD of lateral position, c) Number of crashes, d) mean and SD of speed, e) Deviation from the speed limit (frequency of exceeding the speed limit 120km/h). |
Increase in deviation from the speed limit during daytime drive (χ2 = 0.018, p < 0.05). |
| (Morris, et al., 2015) | USA /2015 | Repeated measures within -subjects design, To suggest a better- quality indicator of driving errors around curves in early stage of sleepiness. |
20 (9 men, 11 women) | Mean 20.55 (± 2.44) yrs. | One night of total sleep loss | Simulated highway drive on a high fidelity KQ-Vection fixed-base simulator/20-min | 5 times of drive from 8 pm. to 10 am. | a) Lateral lane position variability (deviation in lateral lane position), b) SD of Lateral lane position, c) Performance on the curves, e) Heading difference variability; yaw in degrees (the momentary difference between the direction of the vehicle and the tangential direction of the lane). |
a) Increase of deviation in lateral lane position (absolute values) F (4,76) = 10.011, p<0.001, η2p = 0.345, b) Increase in lateral lane deviation across the night (5 test sessions). c) Increase in vehicle heading differences variability (raw data not absolute values) F(4,76) = 15.989, p<0.001, η2p = 0.345457. |
| (Jackson, et al., 2016) | Australia /2016 | Repeated measures within- subjects design, To compare newer algorithms of driver sleepiness over PERCLOS, to compare impairment of total sleep loss with blood alcohol concentrations |
22 (3 men, 19 women) | 18–26 yrs., mean 20.8(± 1.9) yrs. | One night of total sleep loss | The AusEd simulated driving task (computer-based divided attention task)/ 30 minutes | Night time | a) SD of lateral position (the distance from middle of left-hand lane), b) Speed variability (variations of speed from 60–80 km/h), c) Mean number of crashes (off road events, stoppage events, truck collisions). |
a) Increase in SD of lateral position (from 97.98 in baseline to 115.40 cm in sleep deprived condition ((F = 6.81, P = 0.016), b) No change in speed variability (P = 0.214), c) No change in mean number of crashes (P = 0.348). |
| Chronic sleep loss | |||||||||
| Mild sleep loss (1 <sleep loss <2) | |||||||||
| (Matthews, et al., 2012(b)) | Australia /2012 | Between-groups design, Comparison of control group with medium and severe sleep deprived groups |
41 men | Mean 21.8 (±3.8) yrs. | Chronic sleep loss of 1 h and 3 h. | York Driving Simulator/ 10 min | Rotating sleep/wake in to forced descynchrony time, 8–9 times/day. | a) Lateral position (distance from centre of the car to the left lane marker), b) SD of lateral position |
a) Lateral position not reported, b) Increase in SD of lateral position in both moderate and severe sleep restriction. |
| (Garner, et al., 2015) | USA /2015 | Repeated measures within-subjects design, effect of sleep loss on adolescent with various vulnerabilities to sleep loss (based on their attentional decline outside driving setting after sleep loss and parent’s report) in various types of roads (urban/suburban vs rural) |
17 adolescents (8 men, 9 women) | 16–18 yrs., (mean 17.4(± 0.9) yrs. | 5 nights of 6.5 h in bed vs 5-night of 10 h in bed (randomised), two nights washout between them | STISIM M300 simulator/2 counterbalanced drives; a 20-min suburban drive, and a 30-min rural drive | Afternoon at 2pm. or 4 pm. | a) SD of lateral position (distance relative to centre line), b) Mean speed, c) SD of speed, d) Occurrence of crashes: 0 = no crash, 1 = crash event. |
a) No effect of sleep condition on variables, b) A three-way interaction of sleep x drive x vulnerability (P = 0.019), effect size η2 = 0.33; on rural drive, greater SD of lateral position in sleep loss than sleep extension, in urban drive worsened SD of lateral position only in vulnerable people to sleep loss, c) A three-way interaction of sleep x drive x vulnerability (P = 0.015), effect size η2 = 0.36; Vulnerable people had less speed during sleep restriction than sleep extension in rural drive. |
| Moderate sleep loss (2 <sleep loss <4) | |||||||||
| (Matthews, et al., 2012(a)) | Australia /2011 | Repeated measures, within-subjects design, Comparison of chronic moderate sleep loss with normal sleep |
14 men | mean 21.8 (±3.8) yrs. | Chronic sleep loss, 3 h sleep loss (5 h sleep) | York Driving Simulator/10 min | 9 times/day, both day and night time during a 7 forced- desynchronized periods of 23.33 h of wake followed by 4.67 h of time in bed) | a) Lane position (distance from centre of the car to the left lane marker), b) SD of lane position, c) Lane violation, d) Line crossing, e) Crash (when centre of the car leaves the road or car hits the adjacent car), f) Mean and SD of speed, g) Mean deviation from the speed limit, h) Speed variability (SD of deviation from speed limit), i) Speed violation (cumulative time that speed was 5 km/h more than speed limit). |
a) No effect of day (sleep debt) on mean lane position, b) Increase in SD of lane position by time of day or prior wake times or days with growing sleep debt, c) Increase in lane violation by time of day or prior wake times or days with growing sleep debt, d) Change in mean speed after sleep loss over different days, e) Increase in SD of speed after sleep loss at different times of day (mostly after nadir) and on different days, but not at various prior wake times, f) Increase in deviation from speed limit and speed violation by the ‘Day’ variable, capturing the growing sleep debt, g) Speed variability not reported explicitly, but increased by time of day (from 180 degrees after nadir to circadian phase 60 degree) per the figure. |
| (Kosmadopoulos, et al., 2015) | Australia /2015 | Repeated measures between-subjects design, to examine the effect of sleep loss on neurobehavioral and subjective tasks |
32 men | Mean 22.8(± 2.9) yrs. | 9 days of forced-desynchronized protocol with 4h sleep per 24 h | York desktop driving simulator/10 min | 5 times in every 24 hours, every 2.5 h, beginning 1.5 h after awakening | SD of lateral position (distance in metres from the centre point of the car to the centre of a two-lane road) |
a) SD of lateral position was most sensitive to sleep loss F (1, 30) = 38.52, P<0.001, and circadian phase F(5,534) = 61.48, P<0.001, and interaction of circadian phase with sleep dose F(5,534) = 31.98, P<0.001, b) Large effect sizes of sleep loss (f2 = 1.28) and circadian phase (f2 = 0.58) on SD of lateral position. c) A large interaction of circadian phase with sleep loss on SD of lateral position (f2 = 0.85) during circadian nadir. |
aStandard deviation
bSame as study of Verster., et al 2011
cSame as study of Contardi., et al 2004