Performances-Sleep and Circadian Rhythms

38 downloads 0 Views 406KB Size Report
light for 4 hours, dim light for 2 hours followed by bright light for 2 hours or dim light for 4 hours. ... experienced by night workers depends, at least in part,.
\

Sleep, 17(2): 140--145 © 1994 American Sleep Disorders Association and Sleep Research Society

Performances-Sleep and Circadian Rhythms Two- and 4-Hour Bright-Light Exposures Differentially Effect Sleepiness and Performance the Subsequent Night *tVeronica C. Thessing, tAo Michael Anch, *Mark J. Muehlbach, *Paula K. Schweitzer and *tJames K. Walsh *Deaconess Medical Center, St. Louis, Missouri, U.S.A.; and tSt. Louis University, St. Louis, Missouri, U.S.A.

Summary: The effect of two durations of bright light upon sleepiness and performance during typical night shift hours was assessed. Thirty normal, healthy young adults participated in a 2-night protocol. On the I st night subjects were exposed to bright or dim light beginning at 2400 hours, under one of the following three conditions: bright light for 4 hours, dim light for 2 hours followed by bright light for 2 hours or dim light for 4 hours. Following light exposure, subjects remained awake until 0800 hours in a dimly lit room and slept in the laboratory between 0800 and 1600 hours, during which time sleep was estimated with actigraphy. Throughout the 2nd night, the multiple sleep latency test (MSLT), simulated assembly line task (SALT) performance, and subjective sleepiness were recorded. The single, 4-hour exposure to bright light was found to significantly increase MSLT scores and improve SALT performance during the early morning hours on the night following bright-light exposure. No significant effects were noted with a 2-hour exposure. The most likely explanation for these findings is a phase delay in the circadian rhythm of sleepiness-alertness. Key Words: Bright light-Shift work-Circadian rhythm-PerformanceSleepiness.

The incongruity between sleep/wake schedules and the circadian system of shift workers often results in impaired alertness and decreased performance and has a negative impact on safety during night work (1,2). Moreover, intolerance to shift work is postulated to be associated with internal de synchronization of circadian rhythms (3). Thus, in order to improve the health, safety and performance of night workers, it is highly desirable to maximize concordance between sleep/wake schedules and circadian rhythms. Previous research demonstrated that appropriately timed, bright-light exposure can phase shift the human circadian system (4-6), leading to recommendations for use with shift workers (7-10). In fact, such previous laboratory studies using bright-light treatment to counter the negative effects of night work have been prom-

Accepted for publication November 15, 1993. Address correspondence and reprint requests to Veronica Thessing, Sleep Disorders and Research Center, Deaconess Medical Center, 6150 Oakland Ave., St. Louis, MO 63139, U.S.A.

ising. Bright light can cause an adaptive phase shift in circadian rhythms, resulting in increased nocturnal alertness and cognitive performance and eqhanced daytime sleep (6,8,10,11). Bright light also appears to have direct alerting effects during exposure (12-15), which results in improved performance. Although the phase shifting and alerting effects of bright light have been clearly demonstrated, the optimal parameters of bright-light exposure (e.g. duration, intensity, number of exposures) are not well established. The practicality of bright-light treatment to counter the sleepiness and performance decrements experienced by night workers depends, at least in part, on the specific dimensions of light exposure required to produce the desired effect. Dawson and Campbell (11) have reported that a single 4-hour bright-light exposure (from 2400 to 0400 hours) significantly improved alertness 2 nights later during typical night shift hours. A related issue, which was investigated in the present study, is the effect of the duration of a single bright-light exposure necessary to produce a significant decrease in sleepiness during the subsequent night.

140

EFFECTS OF DIFFERENT BRIGHT LIGHT DURATIONS

In this study, sleepiness and performance of subjects working simulated night shifts were compared after being exposed to either 4, 2 or 0 hours of bright light on the preceding night. METHODS Subjects Thirty normal, healthy young adults (19 females, 11 males; mean age = 21 years, range 18-29 years) participated in a 2-night protocol. Subjects were free of any active medical or psychological illness or sleep complaints. Smokers, those with a history of extreme photosensitivity, and those who had worked a night shift in the past 3 months were excluded from the study. One potential subject reported extreme photosensitivity based on the need to wear protective eyewear and thus did not participate. All subjects were required to score between 30 and 70 on the HorneOstburg Morningness-Eveningness Questionnaire (16). Prior to participation, all subjects reported sleeping at least 8 hours per night between 2200 and 0800 hours for at least 2 nights. Subjects signed an informed consent form prior to entering the study and were paid for their participation. Procedures Subjects were trained on the simulated assembly line task (SALT) (17) to a minimum criterion of 90% correct responses before entering the study. Subjects were then randomly assigned to one of three conditions: bright light for 4 hours (BL-4), dim light for 2 hours followed by bright light for 2 hours (BL-2) or dim light for the entire 4 hours (BL-O). Light was measured with a light meter (Luna-Pro F Light Meter, Gossen, Germany) at the subject's eye level (approximately 1 ft from the light source) at the beginning and end of each bright or dim light session for each subject. These measurements resulted in a mean illuminance of 9,258.4 lux for the BL-4 condition; mean illuminance of 355.8 lux for the dim-light session of the BL-2 condition and mean illuminance of 8,772.6 lux for the bright-light session of the BL-2 condition; and mean illuminance of 315.8 lux for the BL-O condition. Relatively high illuminance levels were used to maximize the possibility of a significant effect of a single exposure. On night 1, light (bright or dim) exposure began at 2400 hours for all subjects. Subjects were seated in a chair approximately 1 ft from the light sources, which consisted of three V-shaped cool white florescent light bulbs encased in a 24 x 24 x 3.5-in box (Medic Light, Inc., Lake Hopatcong, NJ) and a 36 x 18 x 5-in box (The Sun box Co., Gaithersburg, MD). All surfaces sur-

141

rounding the light source were white in order to reflect the bright light. During exposure subjects were allowed to read, study, watch a 5-in screen television placed below the desk stand of the light source, or engage in any stationary activity while keeping themselves directly facing the light source (i.e. not turning from side to side). Subjects were also instructed, and frequently reminded, to occasionally look into the light source, but not to stare directly into the light for an extended period of time. At 0230 hours, subjects received a light snack containing no caffeine. Following exposure, subjects were required to remain awake until 0800 hours in a dimly lit laboratory « 300 lux) where they remained throughout the study while being continuously monitored by staff personnel. During this time (0400-0800 hours) subjects were allowed to move about the dimly lit laboratory and continue their previous activities. However, subjects were neither permitted to ingest any stimulating substances, such as caffeine, nor were they allowed to partake in physical exercise to help keep themselves awake. At 0700 hours, subjects ate breakfast. From 0800 to 1600 hours they slept in the laboratory wearing a wrist activity monitor (actigraph, Ambulatory Monitoring, Inc.). Actigraph measures of wrist movement were taken at I-minute epochs. Actigraph records were scored using previously validated software (18). Sleep was defined as the first of three successive epochs with activity level less than or equal to 20 units (accumulated movements). Wake was defined as any epoch with activity level greater than 20 units and those epochs with activity level less than or equal to 20 units that did not meet the above criterion for sleep. After sleeping in the laboratory, from 1600 to 2300 hours subjects were allowed to engage in indoor sedentary activities, but were not exposed to sunlight. On night 2, the following dependent measurements were made throughout the night according to the schedule depicted in Fig. 1. Illumination levels in the laboratory were maintained below 300 lux throughout the day and night. Multiple sleep latency test (MSLT). MSLT was performed at 2-hour intervals, beginning at 2300 hours for a total offive tests. The MSLT is a well-established, widely used measure of physiological sleepiness that provides an electro physiological determination of the onset of sleep in a sleep-conducive environment. Methods and scoring followed established guidelines (19). Simulated assembly line task (SALT). The SALT was performed every other hour for 60 minutes beginning at 2330 hours for a total of four sessions. The SALT visual performance measure presents subjects with images of electronic circuit boards that pass across a video Sleep, Vol. 17, No.2, 1994

L

v.

142 '.1M.

(sleepiness)

2300 2330 2400 0030 0100 0130 0200 0230 0300 0330 0400 0430 0500 0530 0600 0630

X

ozoo

x

C. THESSING ET AL. ~ (begins)

~

(performance)

X

x

15 ----.--

BL-O BL-2

-

BL-4

-0-

10

X X

20

X

x x

x

Snack

x x x x

X

x

5

o+---~----~--~----~--~

0100 0300 0500 Time of Night

0700

FIG. 2. Mean MSLT sleep latencies for each condition (vertical ba.rs = standard errors) at five time points during the night. BL-4 = bnght hght exposure for 4 hours, BL-2 = dim light for 2 hours followed by bright light for 2 hours, BL-O = dim light for 4 hours. Subte~ts were conducted at the same time during all conditions; graphical displacement of time points in all figures for clarity only.

x

x

FIG. 1. ~ch.edule for each subject on the 2nd night of the present study begmnmg at 2300 and ending at 0700 hours. ~onitor as objects might travel on a typical assembly Ime conveyor belt. The participant uses a mouse interface with the computer and is required to perform quality control inspections on each object in order to identify and reject faulty "products" or repair certain types of defective boards. Subjects must also respond to a reaction-time test consisting of "alarms", which represent assembly line down time, by pressing the buttons on the mouse interface as quickly as possible. Three SALT variables were identified for analysis: (1) correct responses: the percentage of discardable or repairable boards correctly handled; (2) correction time: the mean number of seconds from appearance of a faulty board until appropriate action was taken; and (3) alarm response time: the mean number of seconds from alarm onset to operator intervention to resume assembly line progress. Visual analog scale (VAS). Subjective sleepiness was measured with a VAS 5 minutes prior to each MSLT. Subjects estimated their level of sleepiness by making a vertical mark through a 1OO-mm horizontal line, with one pole labeled "extremely sleepy" and the other labeled "extremely alert". Subjects also rated their performance on a VAS after each SALT session. Subjects placed a vertical mark on a lOO-mm horizontal line with one pole labeled "extremely poor" and the othe; labeled "extremely well".

Data analysis

Data were analyzed with repeated-measures analysis of variance (ANOVA) with light condition as the between-groups factor and time of night as the withingroups factor, using the multivariate general linear hypothesis module ofSYST AT (SYST AT Inc., Evanston, Sleep, Vol. /7, No.2, 1994

IL) statistical software. Planned comparisons were made among groups at each time point for the objective measures: MSLT and SALT data.

RESULTS . Because of technical problems, actigraph data were obtained for only 24 subjects (six in the BL-4 group, nine in both the BL-2 and BL-O groups). Estimated sleep duration did not significantly differ among conditions, averaging 415 (SE = 28.0) minutes, 430 (SE = 27.4) minutes and 437 (SE = 36.7) minutes for BL0, BL-2 and BL-4, respectively. MSLT. Figure 2 contains a graph of mean MSLT values and their standard errors for each condition across the night. As expected, 3 x 5 (light condition x time of night) repeated-measures ANOVA revealed a significant main effect of time of night for the MSLT, with sleepiness increasing as the night progressed in all conditions (F = 19.44, p < 0.01). There was a trend (F = 3.06, p = 0.06) toward a main effect of condition with BL-4 (mean sleep latency = 9.74 minutes) being more alert during the night as compared to the other two conditions (mean sleep latency for BL-2 = 5.75 minutes, for BL-O = 5.68 minutes). No significant time of night x condition interaction was noted. Planned comparisons revealed that sleep latencies at 0500 and 0700 hours were significantly greater for BL-4 than for BL-O (p < 0.04 for each), with a trend in the same direction at 0300 hours (p = 0.11). Comparisons at 0500 and 0700 hours between BL-4 and BL-2 approached significance (p = 0.07 and 0.14, respectively) and were also in the same direction. SALT. All SALT performance variables were analyzed with a 3 x 4 (light condition x time of night) repeated measures ANOV A. Mean SALT correct responses and their standard errors for each condition across the night are shown in Fig. 3. The percentage

143

EFFECTS OF DIFFERENT BRIGHT LIGHT DURATIONS 100 8 90

-

----

-----...--

C OJ

2 OJ

c..

80

7

BL-O BL-2 BL-4

0130

0330

0

"

OJ

6

(J)

5

2330

0530

of discardable or repairable boards correctly handled decreased throughout the night (F = 8.86, p < 0.01). There was no main effect among conditions (F = 2.068, p = 0.15). Planned comparisons showed significantly better performance later in the night at 0530 hours for BL-4 (mean = 93.94%) as compared to BL-2 (mean = 80.09%; p < 0.02) and a trend when compared to BL-O (mean = 81.95%; p = 0.08). No interaction was found. SALT correction time showed a significant main effect of time of night (F = 11.60, p < 0.01). The mean number of seconds from appearance of a faulty board until appropriate action was taken increased throughout the night (see Fig. 4). There was no main effect among conditions (F = 1.899, p = 0.17). Planned comparisons revealed that correction time at 0130 hours was significantly less for BL-4 than BL-O (p < 0.05), and at 0330 hours was significantly less for BL-4 than for BL-2 (p < 0.05). Mean response time to SALT "alarms" showed a trend toward a main effect for time of night (F = 2.18, p = 0.097). The mean number of seconds from alarm onset to operator intervention to resume assembly line progress slightly increased throughout the night (see Fig. 5). There was no main effect of condition (F = 0.85, p = 0.44). However, planned comparisons showed a trend for faster responses later in the night at 0530 hours for BL-4 (mean = 1.12 seconds) as compared to BL-2 (mean = 3.19 seconds; p = 0.116). VAS. Subjective ratings of sleepiness and performance both showed significant time-of-night effects [F(4,24) = 24.05, p < 0.01 and F(3,25) = 39.69, p < 0.01, respectively]. From 2300 to 0700 hours, sleepiness ratings increased 52.4%, whereas from 0100 to 0630 hours, performance ratings declined 50.2%. No condition differences were found for either rating scales. The means for each condition on the sleepiness scale were BL-O = 41.78 mm (SD = 23.77), BL-2 = 38.76 mm (SD = 20.03) and BL-4 = 53.02 mm (SD = 24.01). The means for each condition on the performance scale

0130

0330

BL-O BL-2 BL-4

0530

Time of Night

Time of Night

FIG. 3. Mean (and standard error) percentages of discardable or repairable boards correctly handled (correct responses) on SALT for each condition at four time points during the night. BL-4 = bright light exposure for 4 hours, BL-2 = dim light for 2 hours followed by bright light for 2 hours, BL-O = dim light for 4 hours.

---..-

c:

o+-----~----~----~----__.

2330

-

----