Conceived and designed the experiments: JT PS PP AB OR JFD HA. Performed the experiments: JT PS AB OR. Analyzed the data: PS AC JT. Contributed reagents/materials/analysis tools: PS JT. Wrote the paper: PS JT HA AB OR JFD AC PP.
The authors have declared that no competing interests exist.
The contribution of circadian system and sleep pressure influences on executive performance as a function of age has never been studied. The aim of our study was to determine the age-related evolution of inhibitory motor control (i.e., ability to suppress a prepotent motor response) and sustained attention under controlled high or low sleep pressure conditions.
14 healthy young males (mean age = 23±2.7; 20–29 years) and 11 healthy older males (mean age = 68±1.4; 66–70 years) were recruited. The volunteers were placed for 40 hours in “constant routine”. In the “Sleep Deprivation SD” condition, the volunteer was kept awake for 40 hours to obtain a high sleep pressure condition interacting with the circadian process. In the “NAP” condition, the volunteer adopted a short wake/sleep cycle (150/75 min) resulting in a low sleep pressure condition to counteract the homeostatic pressure and investigate the circadian process. Performances were evaluated by a simple reaction time task and a Go/Nogo task repeated every 3H45.
In the SD condition, inhibitory motor control (i.e., ability to inhibit an inappropriate response) was impaired by extended wakefulness equally in both age groups (P<.01). Sustained attention (i.e. ability to respond accurately to appropriate stimuli) on the executive task decreased under sleep deprivation in both groups, and even more in young participants (P<.05). In the NAP condition, age did not influence the time course of inhibitory motor control or sustained attention. In the SD and NAP conditions, older participants had a less fluctuating reaction time performance across time of day than young participants (P<.001).
Aging could be a protective factor against the effects of extended wakefulness especially on sustained attention failures due to an attenuation of sleep pressure with duration of time awake.
The rhythms and demands of modern societies imply that many workers need to support optimal cognitive functioning throughout extended period including nighttimes while performing complex activities (e.g., health, security and transport). Moreover, extended work during the night is known to increase the risk of professional errors
Two major regulatory processes, the circadian system driven by the endogenous biological clock and the sleep-wake homeostatic process which is dependent on the duration of prior wakefulness (sleep pressure/sleep need), interact to regulate sleep and wakefulness according to nycthemeral variations. The circadian process regulates wake- and sleep-promoting mechanisms (timing, consolidation)
Aging is associated with marked changes in the timing, consolidation and structure of sleep. Specifically, marked changes appear in sleep timing, quality and duration, such as decreases in sleep depth (measured by arousal threshold), sleep intensity (measured by slow wave activity (SWA)), sleep continuity (measured by awakenings during the night), and sleep duration
Sleepiness and neurobehavioral functions have also been shown to depend on the interaction of homeostatic and circadian processes
Many studies have shown that extended wakefulness impairs neurobehavioral performance (i.e., sustained attention) as assessed by a basic test of simple reaction time
Two studies have shown that neither the homeostatic process
Inhibition of action is a major component of executive control (i.e., higher cognitive functions) to afford adapted behavioral responses
The experiments testing the effect of sleep deprivation on PFC-related executive functions show inconsistent results. Indeed, some studies report that sleep deprivation has adverse effects on decision making
To our knowledge, the contribution of circadian system and sleep pressure influences on motor inhibitory control as a function of age has never been studied. The aim of our study is to determine the age-related evolution of simple or executive performance under high or low sleep pressure conditions.
Twenty five healthy participants, 11 older [Age (±SD) = 68±1.4 years, range 66–70 years] and 14 young participants [Age (±SD) = 23±2.7 years, range 20–29 years], were recruited via advertisements (at Universities, organizations or hospitals of Bordeaux and Toulouse) or internet announcements.
Participants gave their written and informed consent to the study which was approved by the local ethics committee (committee for the protection of persons participating in biomedical research, Comité de Protection des Personnes (CPP) Sud-Ouest et Outre Mer III).
Exclusion criteria were medical, psychiatric, neurologic and sleep disorders as assessed by screening questionnaires. Volunteers with self-reported excessive daytime sleepiness (Epworth Sleepiness Scale, score ≥11)
A neuropsychological assessment ensured that older volunteers had no motor-, attention- or memory-related impairments. A neuropsychologist assesses a set of informant-based items describing performance of activities of daily living
Each participant was monitored for 7 days with actimeters (Actiwatch®, Cambridge Neurotechnology, United Kingdom) confirming normal sleep timing and sleep duration, and showing at least 85% mean sleep efficiency over a week to be recruited. Participants were instructed to maintain their usual-preferential sleep patterns (habitual sleep/wake timing and sleep duration) verified by actimetric recordings 3 days before each condition of the protocol.
They spent an adaptation night in the laboratory to familiarize them to sleep in a hospital environment with EEG recording.
After a baseline night, a 40-h Sleep Deprivation (SD) condition (top panel) and a 40-h NAP condition alternating short wake/sleep cycles (150/75 minutes) (lower panel) under constant routine protocol were carried out, followed by an 8-h recovery night.
Each participant underwent 2 conditions, SD and NAP conditions (2 days each), in a balanced crossover design with a washout period of at least 2 weeks.
After a baseline sleep night, a 40-h SD under constant routine protocol or a 40-h NAP condition, both under constant conditions (semi-recumbent posture during scheduled wakefulness and supine during scheduled sleep/nap episodes, isocaloric snacks at regular intervals), was carried out
In the SD condition, the volunteers were kept awake during a 40-h extended wakefulness period to obtain a “high sleep pressure condition” interacting with the circadian process.
In the NAP condition, the volunteers adopted 10 alternating wake/sleep cycles (150/75 minutes) during a 40-h multiple sleep nap period resulting in a “low sleep pressure condition” to counteract homeostatic pressure and to examine the circadian influence.
A constant dim light level (<10 lux) during wakefulness and complete darkness (0 lux) during scheduled sleep/nap episodes were set. The protocol ended with an 8-h recovery sleep night.
Prior to the experiment, the participants were invited to complete training sessions to be familiarized with simple (Simple reaction time task SRTT) and executive tasks (Go/Nogo task) of the protocol.
The tests were performed 11 times every 3H45 throughout each condition.
A 10-min simple reaction time test (SRTT) on a PALM personal organizer
The Go/Nogo task requires frequent automatic responding to stimuli interspersed with the need to suppress (i.e., to inhibit) a response from a specific, less frequently occurring stimulus.
The computerized Go/Nogo task is related to inhibitory functions and consists of 2 kinds of visual stimuli presented individually and in random order in the centre of the screen in white on a black background for 1250 ms preceded by a 250 ms fixation point and followed by a 500 ms interstimulus interval: 75% of Go stimuli (respond to a stimulus) and 25% of Nogo stimuli (refrain from responding to a stimulus). Thus a motor response had to be executed (Go) by pressing the space bar on the keyboard as quickly as possible, or inhibited (Nogo). The stimuli Go and Nogo (arrows to the left or to the right) were counter-balanced across participants. The experiment was programmed using E-Prime (v1.2, Psychology Software Tools, Inc., Pittsburgh, PA, USA, 2006). A total of 576 stimuli divided into 9 task blocks were shown during the 30 min task. This task was assessed every 3H45 (8H, 11H45, 15H50, 19H15, 23H, 2H45, 6H30, 10H15, 14H, 17H45 and 21H30).
Immediately before each test sessions, participants were asked how sleepy they were on a 100-mm visual analogue scale (VAS), with scores ranging from 0 (“not sleepy at all”) to 100 (“very sleepy”).
Mean of the 10% slowest
All variables were analyzed with three-way ANOVAs with repeated factors “condition” (SD vs. NAP), time (T 1–11) and the between subject-factor “age” (young vs. older). Planned comparisons were performed to localize statistical differences in significant main effect or interaction. Alpha criterion was set at P = .05. Statistica® (StatSoft Inc. 2010, Statistica for Windows, Maisons-Alfort, France, Version 9.1) was used.
No significant difference appears on total sleep time before SD condition and NAP condition (Mean ± SD = 482±49 versus 474±56, respectively; Wilcoxon test, Z = 1.183, NS). No significant difference appears on sleep efficiency before SD condition and NAP condition (Mean ± SD = 89±2.7 versus 89±3.3, respectively; Wilcoxon test, Z = 0.484, NS).
SD = Sleep deprivation SRTT = Simple reaction time task.
VAS | SRTT | Go/Nogo | |||||||||
Effect | d.f. | Subjective Sleepiness | 10% slowest RTs | Go RTs | % missed Go | % false positive Nogo | |||||
|
1, 23 | 1.5 | NS | 9.3 |
|
5.3 |
|
4.0 | = .056 | 3.6 | = .070 |
|
1, 23 | 5.7 |
|
11.6 |
|
5.9 |
|
24.9 |
|
6.6 |
|
|
10, 230 | 14.0 |
|
16.0 |
|
10.7 |
|
8.8 |
|
5.8 |
|
|
1, 23 | 0.4 | NS | 2.5 | NS | 1.0 | NS | 3.3 | = .082 | 0.4 | NS |
|
10, 230 | 2.6 |
|
4.2 |
|
4.7 |
|
2.2 |
|
1.3 | NS |
|
10, 230 | 5.4 |
|
4.1 |
|
5.3 |
|
8.7 |
|
2.6 |
|
|
10, 230 | 2.6 |
|
0.7 | NS | 0.9 | NS | 1.9 |
|
1.2 | NS |
d.f. = Degree of Freedom.
VAS = Visual analog scale.
SRTT = Simple Reaction Time Task.
The main effect “age” yielded significance for the 10% slowest RTs (F1,23 = 9.3, P<.01) with significantly slower reaction times in older than young participants. The main factor “condition” was significant (F1,23 = 11.6, P<.01) with significantly slower reaction times in SD condition than NAP condition. The main factor “time” was significant (F10,230 = 16.0, P<.001) with significantly slower reaction times during (P<.001) and after (P<.001) the biological night compared to the baseline day. The factor “time” significantly interacts with the factor “condition” (F10,230 = 4.1, P<.001) with a slowing of reaction times after the biological night more pronounced in the SD than in the NAP condition. The factor “age” did not significantly interact with the factor “condition” (F1,23 = 2.5, NS), but with the factor “time”, with young participants becoming as slow as older participants at the end of the biological night and during the subsequent day in the SD and NAP conditions (F10,230 = 4.2, P<.001), except in the late afternoon (17H20: P<.05 and 21H05: P<.05). The interaction “age”, “condition”, “time” did not yield any significance.
SD = Sleep deprivation.
The main effect “age” yielded significance for Go RTs (F1,23 = 5.3, P<.05) with significantly slower reaction times in older than young participants. The main factor “condition” was significant (F1,23 = 5.9, P<.05) with significantly slower reaction times in SD condition than NAP condition. The main factor “time” was significant (F10,230 = 10.7, P<.001) with significantly slower reaction times during (P<.001) and after (P<.05) the biological night compared to the baseline day. The factor “time” significantly interacts with the factor “condition” (F10,230 = 5.3, P<.001) with a slowing of reaction times after the biological night more pronounced in the SD than in the NAP condition. The factor “age” did not significantly interact with the factor “condition” (F1,23 = 1.0, NS), but with the factor “time”, with young participants becoming as slow as older participants at the end of the biological night and during the subsequent day in the SD and NAP conditions (F10,230 = 4.6, P<.001), except in the evening (21H30: P<.05). The interaction “age”, “condition”, “time” did not yield any significance.
SD = Sleep deprivation.
It is to note that two young participants out of the 25 participants did not miss any Go trial in the overall of the NAP condition.
The main effect “age” yielded significant tendency for the % missed Go (F1,23 = 4.0, P = .056) with higher % missed in young than older participants. The main factor “condition” was significant (F1,23 = 24.9, P<.001) with significantly higher % missed in SD condition than NAP condition. The main factor “time” was significant (F10,230 = 8.8, P<.001) with significantly higher % missed during (P<.01) and after (P<.001) the biological night compared to the baseline day. The factor “time” significantly interacts with the factor “condition” (F10,230 = 8.7, P<.001) with higher % missed during and after the biological night exclusively in the SD condition. The factor “age” did not significantly interact with the factor “condition” (F1,23 = 3.3, P = 0.08), but with the factor “time” (F10,230 = 2.2, P<0.05). The interaction “age”, “condition”, “time” yielded significance (F10,230 = 1.9, P<0.05). Planned comparisons show that young participants made higher % missed than older participants during the subsequent day after the biological night in the SD condition (14H: P<.05 and 17H45: P<.05) while age group difference was inexistent in the NAP condition.
SD = Sleep deprivation.
The main effect “age” did not yield significance for the % false positive Nogo (F1,23 = 3.6, P = 0.07). The main factor “condition” was significant (F1,23 = 6.6, P<0.05) with significantly higher % false positive Nogo in SD condition than NAP condition. The main factor “time” was significant (F10,230 = 5.8, P<0.001) with significantly higher % false positive Nogo during (P = .057) and after (P<.01) the biological night compared to the baseline day. The factor “time” significantly interacts with the factor “condition” (F10,230 = 2.6, P<.01) with higher % false positive Nogo during and after the biological night in the SD than in the NAP condition. The factor “age” did not significantly interact with the factor “condition” (F1,23 = 0.4, NS) nor with the factor “time” (F10,230 = 1.3, NS). The interaction “age”, “condition”, “time” did not yield any significance (F10,230 = 1.2, NS).
The main effect “age” did not yield significance for the VAS Sleepiness (F1,23 = 1.5, NS). The main factor “condition” was significant (F1,23 = 5.7, P<.05) with significantly higher subjective sleepiness scores in SD condition than NAP condition. The main factor “time” was significant (F10,230 = 14.0, P<.001) with significantly higher subjective sleepiness scores during (P<.001) and after (P<.001) the biological night compared to the baseline day. The factor “time” significantly interacts with the factor “condition” (F10,230 = 5.4, P<.001) with higher subjective sleepiness scores during (P<.05) and after (P<.001) the biological night compared to the baseline day, which were more pronounced in the SD than in the NAP condition. The factor “age” did not significantly interact with the factor “condition” (F1,23 = 0.4, NS), but with the factor “time” (F10,230 = 2.6, P<.01). The interaction “age”, “condition”, “time” yielded significance (F10,230 = 2.6, P<.01). Planned comparisons show that young participants estimate themselves less sleepy than older participants during the day following normal sleep while youngest become as sleepy as older participants during the biological night in the SD condition. No age group difference did exist in the NAP condition.
Our study confirms that normal aging leads to a cognitive slowing (i.e., increased reaction time) in simple and complex tasks
Regarding the influence of sleep deprivation on speed-related processing, we found a slowing of reaction time performance on simple and executive tasks during and after the biological night in the SD condition in both age groups, which was even more pronounced for young participants. The latter tend to become as slow as older participants at the end of the biological night and during the morning hours of the subsequent day. This could mean that the circadian process has a greater adverse effect on younger people than on older ones. Blatter et al. (2006)
In addition, we observe that the older people’s performance curve follows a flattened time course under low sleep pressure in the NAP condition compared to that of young participants. Inasmuch as the condition (high vs. low sleep pressure) does not influence this pattern (interaction age*time*condition not significant), our study confirms that age-related lower vulnerability to extended wakefulness seems predominantly due to an attenuated circadian regulation on reaction time performance in the older group
Regarding accuracy performance, actions errors during a Go/Nogo task can result either from sustained attention failure (i.e., omission errors) or from inhibition failure (i.e., commission errors). The percentages of omission and commission errors are stable across day and night when sleep pressure is low (i.e., in the multiple naps condition). Our study shows that, conversely to reaction time performance, the accuracy on executive task, which represents the success criterion of correctly achieving a task, is not modulated by the circadian component. We observe a deterioration of accuracy performance under high sleep pressure (i.e., sleep deprivation condition). Indeed, our results show that young and older individuals experience difficulty in ability to inhibit an inappropriate prepotent response (i.e., inhibition failure) and difficulty in responding accurately to appropriate stimuli (i.e., sustained attention failure) during and after a night of sleep deprivation. These results corroborate those of Drummond et al. (2006)
Here, we used a constant routine protocol that constitutes the gold standard to measure circadian modulation of neurobehavioral functions, as well as the effect of sleep pressure developing with duration of time awake. In addition, the condition of scheduled sleep at regular intervals during a 40-h episode makes it possible to maintain low sleep pressure conditions and thus reveals the circadian rhythm without the confounding effects of elevated sleep pressure. However, further studies using a forced desynchrony protocol are needed to identify the contribution of the homeostatic and circadian processes on performance.
Moreover, we evaluate the effects of age, circadian and homeostatic influences on behavioral inhibition (i.e., ability to suppress a prepotent response) through commission errors on a Go/Nogo task. Further studies will have to evaluate others aspects of response withholding as the ability to stop a response that has already been initiated (e.g., Stop signal paradigm)
Our study confirms the importance of circadian and homeostatic factors in the regulation of neurobehavioral function. However, in addition to the age factor, the characteristics of the tasks (simple or executive)
In conclusion, we show that inhibitory motor control (i.e., suppression of an inappropriate prepotent motor response) is fully preserved in no sleep-deprived aged people while equally impaired by extended wakefulness in young and older people. Our study reveals that error-related processing in a behavioral inhibition task does not seem to be regulated by circadian processes contrary to speed-related processing. Moreover, older people demonstrate not only an attenuation of the circadian influence on speed-related processing but also a reduction of sleep pressure with duration of time awake on sustained attention error-related processing. Therefore, aging could be a protective factor against the effects of extended wakefulness on sustained attention failures due to an attenuation of sleep pressure with duration of time awake. Strategies could be developed to prevent accidents according to the age of workers and their work schedule.
We thank the following clinical research assistants: M. Bacarisse, C. Valtat, A. Boiseau and V. Bibène for selecting participants, collecting and monitoring data and for administrative, technical, and logistic support.