Feasibility of DS for supplementing in-cabin daylight
Daytime light exposure
In-cabin illuminances were measured continuously from 08:30 to 15:45 close to the eyes of the truck drivers with a wearable device (LuxBlick 2.039). To better account for non-visual effects of light on the melanopsin photoreceptor system, the MEDI (CIE S 026/E:201835) at eye level was additionally estimated.
Light exposure during driving rest periods (DS++)
DS++ was applied before and after test drives (from 09:00 to 10:00 and from 14:00 to 15:40). In the morning, participants were exposed to 531 ± 221 lx under DS++ compared to 47 ± 60 lx in the placebo condition. This corresponded to an estimated mean 1459 MEDI under DS++ versus 19 MEDI in the placebo intervention. Similarly, participants were exposed to higher illuminances in the afternoon under DS++ compared to the placebo (DS++: 538 ± 253 lx, estimated mean 1478 MEDI; placebo: 28 ± 55 lx, estimated mean 11 MEDI).
Light exposure during driving (DS
+
)
During the 4-h driving periods, truck drivers were exposed to 650 ± 379 lx under DS+ compared to 247 ± 374 lx in the placebo condition with no additional artificial in-cabin light, resulting in a 2.5-fold increase in photopic stimulation (Fig. 3). Moreover, the drivers were exposed to an estimated average of 1,346 MEDI under DS+ compared to 247 MEDI in the placebo condition, indicating an increase in melanopic stimulation by a factor of 5.4. Without additional in-cabin lighting (i.e., in the placebo condition), only 14% of the windshield illuminances reached the drivers’ eyes.
In the second half of the eight study weeks starting in early January, outdoor illuminances varied considerably, barring two extremely bright days (during the placebo condition) in the last study week in early February (Fig. 3; study week 8). Figure 4 shows all measured illuminances at eye level and estimated MEDIs with high temporal resolution (12-min intervals) across the 4-h test drives.
Light conditions during driving; (a) mean outdoor illuminance (estimated with a photosensor behind the windshield) is depicted during the driving periods for the placebo and the active light (DS+) condition over the course of eight study weeks (SW) from the end of November until early February; (b) mean illuminances at eye level are compared for each corresponding study weekday of the placebo and the DS+ condition; error bars indicate standard errors of the means; on two bright days (marked with hashtags), illuminances at eye level were above 1,000 lx in the placebo condition.
Light exposure during the 4-h test drives; (a) mean illuminances at eye level; (b) estimated mean melanopic daylight equivalent illuminances (MEDIs) at eye level; data are shown per hour (divided by 5 separate 12-min intervals) across five weekdays; the error bars show the standard errors of the mean.
Comfort ratings for the interventions
At the end of the study week, the truck drivers provided comfort ratings for each intervention and were asked to share any negative side effects. Table 2 summarizes the comfort ratings for the different intervention components while driving, including DS and the two co-occurring placebo components, inactive UV light exposure and the mock air-refreshing system.
In contrast to the placebo components, the truck drivers rated DS+ as more “pleasant” and “activating” on a semantic differential (list of bipolar adjective pairs; range from 1 to 7; Table 2). On a 5-point Likert scale, DS+ was perceived as significantly more effective in improving well-being than the other intervention components. In addition, DS+ improved the truck drivers’ subjective self-assessment of their fitness. Furthermore, participants did not perceive DS+ as intrusive and did not rate the adjective pairs “familiar–unfamiliar” and “unobtrusive–obtrusive” differently for the three intervention components (Table 2).
Finally, truck drivers rated DS+ with an average score of 1.3 (graded from 1 = very good to 6 = insufficient), and thus significantly better than the placebo intervention components (UV, 3.8; AIR, 3.1). With an average recommendation score of 5.5 (on a scale of 1 = not at all to 6 = absolutely), the participants strongly recommended the DS+ system to other truck drivers.
Importantly, no study participants reported any negative visual side effects (e.g., restrictions due to glare, dazzling, or eye irritation) while driving in the DS+ or placebo conditions. The full distribution of the ratings results is provided in the Supplementary Materials C.1, Table C.1.
Efficacy of DS on daytime alertness before and after driving
On four of the five study days in the DS++ condition (except for Thursday morning), truck drivers were exposed to bright light before the drives started and after they ended, and subjective and objective alertness data were recorded.
Effects on subjective alertness
The subjective sleepiness measures (KSS) are shown in Fig. 5a. We observed a significant interaction for morning KSS scores (F [1,31] = 4.534, p = 0.041, partial η2 = 0.128). Post-hoc tests showed that the KSS values did not differ between the two interventions at both morning measurement times (both p > 0.05). However, subjective sleepiness increased over time in the placebo condition (08:45, 2.56 ± 0.21; 09:55, 3.28 ± 0.27; p < 0.001), while it remained at the same level in the DS++ condition (08:45, 2.75 ± 0.17; 09:55, 3.03 ± 0.23; p = 0.372).
Subjective measure of sleepiness and objective measures of alertness during driving rests in the morning and afternoon; (a) self-reported sleepiness assessed with the Karolinska Sleepiness Scale (KSS); (b) reaction speed (in s−1) on the Psychomotor Vigilance Task (PVT); (c) number of lapses on the PVT; error bars indicate 95% confidence intervals of the mean.
A second ANOVA was run to determine the effects of DS+ on subjective sleepiness during the test drives by analyzing data from three measurement points (just before the drive started [09:55], during the drive break [12:00], and immediately after the drive ended [14:05]). We did not observe significant interactions or main effects of Intervention or Measurement Time on KSS scores (all p > 0.05).
A third ANOVA examined the effects of DS++ on subjective sleepiness in the afternoon after driving ended. We did not find a significant interaction effect or a main effect of Intervention on KSS scores (both p > 0.10). However, subjective sleepiness significantly increased over time in the afternoon under both study arms (14:05, 3.38 ± 0.28; 15:35, 4.23 ± 0.33; F [1, 31] = 18.140, p < 0.001, partial η2 = 0.369). Thus, the highest sleepiness ratings (scores ranged between rather alert and neither alert nor sleepy) were obtained at the end of the simulated working day in both study conditions.
Effects on objective measures of alertness
A PVT was performed twice in the morning before driving and twice in the afternoon after driving. In the morning, the PVT took place just before DS++ started (at 08:45) and after 45 minutes of exposure to DS++ (at 09:45). In the afternoon, the PVT was performed immediately after the start of the exposure to DS++ (at 14:05) and after 90 minutes of DS++ exposure (at 15:15). Due to skewed data, the reciprocal reaction time (1/RT; i.e., PVT reaction speed) was subjected to statistical analysis.
PVT reaction speed
PVT reaction speeds are shown in Fig. 5b. In the mornings, we observed an increase in reaction speeds under DS++ and a decrease in the placebo condition between the two tests. A two-way ANOVA revealed a significant interaction effect during the morning under DS++ (F [1, 31] = 25.324, p < .001, partial η2 = .450). Both interventions did not differ at 08:45 (DS++ 3.55 ± 0.07 s−1; placebo: 3.52 ± 0.07 s−1; p = .999) but showed a significant difference at 09:45 (p < .001) with higher reaction speed under DS++ (3.68 ± 0.07 s−1) compared to the placebo condition (3.41 ± 0.06 s−1). Moreover, we observed a significant increase in reaction speed under DS++ and a significant decrease in reaction speed under the placebo condition (both p < .05).
In the afternoon, there was a comparable decline in reaction speed in both conditions, but on different levels. There was no significant interaction effect of DS++ (F [1, 31] = 1.509, p < .229, partial η2 = .046). However, there were significant main effects of Intervention (F [1, 31] = 20.446, p < .001, partial η2 = .397) and Measurement Time (F [1, 31] = 26.839, p < .001, partial η2 = .464), with higher reaction speed under DS++ (3.58 ± 0.06 s−1) than under the placebo condition (3.37 ± 0.07 s−1) and higher reaction speed at the first measurement time at 14:05 (3.55 ± 0.07 s−1) than at the second measurement time at 15:15 (3.40 ± 0.06 s−1).
Number of lapses on PVT
The number of lapses on the PVT across the four measurement points during the day are depicted in Fig. 5c. A two-way ANOVA showed a significant interaction effect of DS++ in the morning (F [1, 31] = 6.293, p = .018, partial η2 = .169). Both interventions did not differ at the first measurement point at 08:45 (p = .999) but showed a significant difference at 09.45 (p = .030), with fewer lapses under DS++ (1.63 ± 0.50) than the placebo condition (2.34 ± 0.56). Moreover, there was a significant increase in the number of lapses in the placebo condition between these measurement times (p = .005). In addition, we observed no significant interaction or main effects of Measurement Time on the number of lapses under DS++ in the afternoon (Interaction: F [1, 31] = 0.459, p = .503, partial η2 = .015; Measurement Time: F [1, 31] = 0.413, p = .525, partial η2 = .013). However, the main effect of Intervention reached significance (F [1, 31] = 7.952, p = .008, partial η2 = .204), with fewer lapses under DS++ (1.78 ± 0.47) than under the placebo condition (3.45 ± 1.00).
Carry-over effects of DS beyond the working day
To assess any potentially adverse activating effects of DS in the evenings, subjective alertness measures and melatonin levels were assessed on three study days (Monday, Wednesday, and Friday) on an hourly basis from 17:00 to 22:00 as well as once immediately before going to bed, at 22:25. Results from two-factor repeated-measures ANOVAs with the factors Intervention and Measurement Time are reported here together with results of sleep quality assessments.
Effects on evening alertness levels
Figure 6c shows similar significant increases in self-reported sleepiness levels (KSS ratings) in both study conditions over time (F[6,138] = 45.874, p < 0.001, partial η2 = 0.666), starting with alert (KSS score: 3) to some signs of sleepiness (KSS score: 6) before bedtime (DS: from 2.75 ± 0.24 to 6.08 ± 0.45; placebo: from 2.83 ± 0.20 to 6.13 ± 0.34). No significant differences of Intervention or interaction effects were found between both experimental conditions (Intervention: F[1,23] = 0.596, p = 0.448; Interaction: F[6,138] = 0.710, p = 0.642).
Carry-over effects of DS beyond the working day; (a,b) two objective alertness measures of the Psychomotor Vigilance Task (PVT) at 22:00. (PVT reaction speed and number of lapses); (c) self-reported sleepiness using the Karolinska Sleepiness Scale (KSS); (d) saliva melatonin levels (pg/mL) in the evening.
In addition, objective alertness before bedtime was measured with the PVT on Monday and Friday at 22:00. By applying a t-test for two dependent samples, we observed no differences in reaction speed between the two interventions (t[15] = 0.107, p = 0.748; DS: 3.23 ± 0.09; placebo: 3.27 ± 0.12). In contrast, the number of lapses on the PVT was different (t[15] = 5.064, p = 0.040) with more lapses under DS (4.94 ± 1.20) than under the placebo condition (3.50 ± 1.17) (Fig. 6a,b).
Effects on evening melatonin levels
Saliva melatonin was sampled on three evenings (Monday, Wednesday, and Friday). As shown in Fig. 6d, mean saliva melatonin levels increased from 17:00 to 22:25 in both study conditions (DS: 1.19 ± 0.18 pg/mL to 3.80 ± 0.44 pg/mL; Placebo: 1.31 ± 0.13 pg/mL to 4.38 ± 0.53 pg/mL; F[6,138] = 37.341, p < 0.001, partial η2 = 0.619), with no significant effect of Intervention or Interaction (Intervention: F[1,23] = 3.518, p = 0.073; Interaction: F[6,138] = 1.950, p = 0.077).
Effects on sleep quality
Truck drivers were asked to document their sleep onset and offset times and rate their subjective sleep quality after each of the five study nights (Monday to Friday) in both study conditions.
Subjective sleep duration did not differ between the two interventions (DS: 389 ± 53 min; placebo: 389 ± 57 min; p = 0.998). In addition, overall subjective sleep quality as measured by the SSA did not differ between conditions (DS: 30.5 ± 9.2, placebo: 31.0 ± 6.2; p = 0.348). This remained true for the three SSA subscales of awakening quality (DS: 13.0 ± 4.3, placebo 12.9 ± 3.5; p = 0.880), somatic complaints after awakening (DS: 5.7 ± 1.2, placebo: 5.7 ± 0.9; p = 0.781), and specific sleep quality (DS: 11.9 ± 3.9, placebo: 12.6 ± 3.0; p = 0.348).
Results of the objective sleep quality measurements using PSG twice per study week (Monday and Friday) are shown in the Supplementary Materials (C.3 and Table C.3). In general, the data quality was limited, as the PSG data were often either missing or noisy (e.g., electrodes lost contact during the night). Based on the existing results, however, DS showed no relevant negative effects on global sleep parameters.
Credit: Source link
