Effects of partial and acute total sleep deprivation on performance across cognitive domains, individuals and circadian phase

Abstract

Background: Cognitive performance deteriorates during extended wakefulness and circadian phase misalignment, and some individuals are more affected than others. Whether performance is affected similarly across cognitive domains, or whether cognitive processes involving Executive Functions are more sensitive to sleep and circadian misalignment than Alertness and Sustained Attention, is a matter of debate.

Methodology/principal findings: We conducted a 2 × 12-day laboratory protocol to characterize the interaction of repeated partial and acute total sleep deprivation and circadian phase on performance across seven cognitive domains in 36 individuals (18 males; mean ± SD of age = 27.6 ± 4.0 years). The sample was stratified for the rs57875989 polymorphism in PER3, which confers cognitive susceptibility to total sleep deprivation. We observed a deterioration of performance during both repeated partial and acute total sleep deprivation. Furthermore, prior partial sleep deprivation led to poorer cognitive performance in a subsequent total sleep deprivation period, but its effect was modulated by circadian phase such that it was virtually absent in the evening wake maintenance zone, and most prominent during early morning hours. A significant effect of PER3 genotype was observed for Subjective Alertness during partial sleep deprivation and on n-back tasks with a high executive load when assessed in the morning hours during total sleep deprivation after partial sleep loss. Overall, however, Subjective Alertness and Sustained Attention were more affected by both partial and total sleep deprivation than other cognitive domains and tasks including n-back tasks of Working Memory, even when implemented with a high executive load.

Conclusions/significance: Sleep loss has a primary effect on Sleepiness and Sustained Attention with much smaller effects on challenging Working Memory tasks. These findings have implications for understanding how sleep debt and circadian rhythmicity interact to determine waking performance across cognitive domains and individuals.

Figure 1. Effect of sleep history on total sleep time and performance.

( A ) Protocol: The study (N = 36) consisted of two 12-day, i.e. 11-night, laboratory sessions as shown in the double raster plot, in which consecutive 24-h periods are plotted both next to and below each other. Following an adaptation night and a baseline night with 8-h Time In Bed (TIB; black bars), sleep opportunity in the subsequent seven nights was either 10 h in the Control condition, or 6 h in the Sleep Restriction (SR) condition. This was followed by a 39-h (Control) or 41-h (SR) acute total sleep deprivation period (grey bars), and a recovery sleep episode (TIB = 12 h). A cognitive performance test battery (blue bars) was administered on the baseline (B) day and each of the following six days (D1 to D6), and during total sleep deprivation (TSD; TD1 = first day of total sleep deprivation; TN1 = night of total sleep deprivation; TD2 = second day of total sleep deprivation). ( B ) Total Sleep Time in the Control (open circles) and SR condition (closed circles). ( C ) Subjective Alertness assessed by the Karolinska Sleepiness Scale (KSS). ( D ) Sustained Attention assessed by the Psychomotor Vigilance Task (speed of the slowest 10% responses) and the Sustained Attention Response Task (A’). ( E ) A’ of Working Memory tasks with increasing executive load (verbal 1- to 2- to 3-back). In all panels, the least square means and standard errors estimated with PROC MIXED in SAS are plotted. Asterisks indicate the significance of the contrast between conditions (***P<0.001, **P<0.01, and *P<0.05).

Figure 2. Comparison of effect sizes for Subjective Alertness, Sustained Attention, and Working Memory.

( A ) Effect size of repeated partial sleep deprivation. It was assessed by comparing performance during D5 and D6 between conditions. ( B ) Effect size of acute total sleep deprivation on performance during the circadian day. It was assessed by comparing performance on TD1 to performance on TD2 across conditions. ( C ) Effect size of acute total sleep deprivation on performance during the circadian night. It was assessed by comparing performance on TD1 to performance in TN1 across conditions. Horizontal lines indicate cut-offs for small, medium, and large effect sizes.

Figure 3. Effect of PER3 genotype on performance during repeated partial and subsequent acute total sleep deprivation.

( A ) Time course of Subjective Alertness (top panel) and Working Memory/Executive Functions (bottom panel) in PER3^4/4, PER3^4/5, and PER3^5/5 individuals. The least square means and standard errors estimated with PROC MIXED in SAS are plotted. Asterisks indicate the significance of the contrast between conditions (***P<0.001, **P<0.01, and *P<0.05). Open circles = Control condition; filled circles = Sleep Restriction (SR) condition. ( B, C, D ) Performance during the SR and the Control conditions averaged throughout the protocol (B - TD2) in the three genotypes. The interaction between Genotype and Condition was significant for ( B ) Subjective Alertness (P = 0.0039) and for ( D ) Working Memory, but only for verbal 3-back (P = 0.04). For neither of the ( C ) Sustained Attention measures was the Genotype × Condition interaction statistically significant (ns = not significant).

Figure 4. Comparison of the effect sizes for the Genotype × Sleep Deprivation interaction across cognitive domains.

( A ) During partial sleep deprivation. It was assessed by comparing performance during D5 and D6 between conditions. ( B ) During total sleep deprivation. It was assessed by comparing performance on TD1 to performance on TD2 across conditions. Horizontal lines indicate cut-offs for small effect sizes.

Figure 5. Effects of sleep history on circadian modulation of performance during total sleep deprivation.

Time course of ( A ) Subjective Alertness, ( B ) Sustained Attention, and ( C ) Working Memory during total sleep deprivation (TSD) in the Sleep Restriction (filled circles) or the Control condition (open circles) across different circadian phases relative to dim light melatonin onset (DLMO; grey filled area = the melatonin profile averaged between the two conditions). ( D ) Effects size of prior partial sleep deprivation on performance during TSD per 4-h circadian melatonin bins averaged across the six performance measures. Error bars represent the between performance measure standard error of the mean. Horizontal lines indicate cut-offs for small, medium, and large effect sizes.

Figure 6. Effects of PER3 genotypes on the circadian modulation of performance during total sleep deprivation following partial sleep deprivation.

( A ) Verbal 3-back performance during total sleep deprivation (TSD) in the Sleep Restriction (filled symbols) and the Control conditions (open circles) separately for the three PER3 genotypes. (DLMO: dashed grey vertical line and melatonin profile averaged between the two conditions shaded in grey). ( B ) Effect sizes for the Genotype × Sleep History Condition interaction during TSD for Subjective Alertness, Sustained Attention, and Working Memory, computed for each 4-h circadian melatonin bin. ( C ) Effect size averaged across the six performance measures. Error bars represent the between performance measure standard error of the mean. Horizontal lines indicate cut-offs for small and medium effect sizes.