Sleep deprivation influences performance on a variety of cognitive tasks [7
]. This outcome does not appear, however, to be explained simply by fatigue or boredom, but by direct effects of sleep deprivation on frontal lobe function, and in turn, cognitive processes that depend on the frontal lobes [7
]. It is principally the loss of slow-wave sleep that occurs with sleep deprivation that is thought to affect the frontal lobes and underlie the impact of sleep deprivation on executive control [89
Significant decreases in frontal lobe metabolism have been demonstrated in studies of sleep deprivation [89
]. This may be driven particularly by the lack of slow-wave sleep [98
]. Moreover, this decrease in metabolism is not fully restored with a full night of recovery sleep [98
], suggesting that while a recovery sleep may increase alertness [100
], underlying effects on the frontal lobes may persist.
It is hypothesized that sleep-deprived young adults may serve as an experimental model for age-related cognitive decline [80
]. Similarities in patterns of cognitive performance and brain activity between older adults and sleep-deprived young adults support this view [101
]. Similar to the pattern of “nonselective” recruitment of PFC found in older adults [2
], young adults overrecruit PFC during cognitive tasks following sleep deprivation [101
] but see [103
]. In one of these studies, Drummond et al. [102
] had subjects take part in a verbal learning task. Subjects that were sleep-deprived for 36 hours showed increased activation of several “control” regions including the dorsolateral PFC relative to control subjects. As noted by the authors, this pattern of activation was similar to the pattern found in older adults. Similarly, Chee and Choo [101
] found with a working memory paradigm that young subjects sleep-deprived for 24 hours showed a pattern of activation and deactivation in parts of frontal and parietal cortex that closely resembled the pattern typically observed in healthy older subjects. They found that while anterior medial frontal and posterior cingulate cortex showed significant deactivation, the left dorsolateral PFC showed an increase with sleep deprivation. This increase in left PFC activity in sleep-deprived subjects certainly parallels the increases found in left PFC in older adults during memory retrieval [53
]. These similarities in PFC overrecruitment among healthy older adults and sleep-deprived young adults suggest a common mechanism between sleep disruption and cognitive decline [80
]. This common mechanism may be decline in executive control, or broadly PFC function.
Cognitive impairments that arise from sleep deprivation are often found for executive control tasks [7
]. Neuroimaging studies have also provided evidence to suggest that sleep deprivation affects executive control dependent on the frontal lobes [18
]. In contrast to some studies that have failed to demonstrate significant effects of sleep deprivation on cognition using nonexecutive tasks [113
], studies using “executive tasks” have demonstrated a specific influence of sleep on PFC function. Some of these studies have investigated executive processes that are important for controlled memory retrieval, such as inhibition. For example, Breimhorst et al. [107
] and Schapkin et al. [110
] used a Go-NoGo paradigm to test the hypothesis that inhibitory processing is impaired with sleep disruption. This task requires inhibition on NoGo trials when subjects are instructed to inhibit their response to distracters. Using noise to disrupt sleep in young individuals, Schapkin et al. [110
] examined ERPs associated with Go and NoGo trials. They showed that the fronto-central P3 amplitude (a positive wave with a 300
ms peak latency) elicited by NoGo trials was reduced in the sleep disruption condition. However, the P3 elicited by Go trials was not affected by sleep disruption. The authors concluded based on these results that the decision process associated with Go trials was not influenced by sleep disruption. However, inhibitory processing associated with NoGo trials was negatively affected. This finding suggests that inhibitory control is impaired with sleep disruption.
Also testing a Go-NoGo paradigm in young adults, Breimhorst et al. [107
] examined Go-NoGo ERP effects in good and poor sleepers based on an objective sleep disturbance index using polysomnography. Breimhorst et al. [107
] also found that the NoGo P3 latency was longer in poor sleepers relative to good sleepers, reflecting deficient inhibitory processing. However, in contrast to the Schapkin et al. [110
] study, Breimhorst et al. [107
] also found decreased Go P3 amplitude in poor sleepers. This suggests that poor sleep also affected task-relevant processing, not just inhibition. Despite these differences, these studies collectively suggest that inhibitory processes are negatively impacted by sleep disruption.
Sleep deprivation also appears to influence task-switching processes. Task-switching involves cognitive flexibility and inhibition of irrelevant task-sets and is often considered a model paradigm of executive control [114
]. Couyoumdjian et al. [116
] found a significant increase in switch-costs with sleep deprivation in young adults. Importantly, this effect was driven by an increase in response times on switch trials. There was no change in response times on repeat trials, suggesting that sleep deprivation does not globally influence response time. Instead, it specifically affects subjects' ability to switch between task-sets. Also using a task-switching paradigm, Heuer et al. [117
] found deficits with task-switching following sleep deprivation. In this study, the task-switching costs were influenced by sleep deprivation only when subjects switched between two tasks as opposed to two stimulus-response mappings. This dissociation may support the view that sleep-deprivation influences inhibition of competing task-sets—a more internal implementation of inhibitory control, important for controlled memory retrieval.
Harrison and Horne [118
] revealed marked impairments on a short and entertaining test of inhibition shown to have a PFC-focus (the Haylings test [119
]) following 36 hours of sleep deprivation. This study demonstrated that it was not the tedium of the task that brought about sleep-related deficits, but rather the putative impairments in PFC-mediated inhibition.
Although multiple aspects of sleep deprivation, including a lack of all sleep stages, and increased stress and fatigue on the part of the subject, may contribute to these impairments, the lack of slow-wave sleep affecting PFC restoration is a possible mechanism by which these impairments in executive control occur [98
]. Together, these behavioral and neuroimaging investigations suggest that sleep deprivation has a considerable impact on executive control. We can next ask whether executive control impairments caused by sleep deprivation impact memory retrieval.
The role of sleep in the offline strengthening of memories through consolidation and integration is a well-established phenomenon in both humans and animals [120
], and this process may also be vulnerable to age-related decline [81
]. Sleep-dependent consolidation, however, will not be discussed in depth here because the present paper is focused on strategic memory processing following sleep (during retrieval) as opposed to memory processing during sleep (consolidation), and because a number of thorough reviews already exist in the literature on sleep and consolidation [81
In terms of episodic memory, there are very few studies that have investigated how different retrieval strategies and stages are influenced by sleep. Nonetheless, these studies have demonstrated specific impairments in retrieval processes that depend on the PFC (). Harrison and Horne [129
] examined both recognition memory and temporal order memory judgments. Following a period of sleep or sleep deprivation, subjects were asked to identify whether faces were presented at study or not (recognition) and to make a recency judgment by identifying on which of two study lists the face appeared (temporal order). This temporal order task was posited to depend on the PFC. They found that while recognition memory was left intact, temporal order memory was significantly impaired following sleep deprivation. Using a verbal learning task, Drummond et al. [102
] found that recall, but not recognition performance decreased with sleep deprivation. FMRI data collected in this study found increased PFC recruitment during encoding following sleep deprivation. However, the retrieval phase was not scanned in this particular paradigm. So it is unclear whether PFC overrecruitment occurred during retrieval as well following sleep deprivation. These findings suggest that sleep affects controlled retrieval tasks like recall, but leaves more automatic retrieval processes intact.
Table 2 Retrieval processes influenced by advanced aging, PFC damage, and sleep deprivation. Simple item recognition is thought not to depend on the PFC except under conditions in which subjects are required to distinguish between studied items and unstudied (more ...)
Supporting the notion that memory deficits in older adults and sleep-deprived young adults are similar, Nilsson et al. [130
] found similarities in memory performance between older, young alcohol intoxicated, and young sleep-deprived subjects. In a recall test using weakly and strongly related word pairs, they found that all experimental groups (older, intoxicated, and sleep-deprived) demonstrated the same pattern of deficits in which recall of weakly related word pairs was significantly lower than that of the control young adult group. Recall of strongly related word pairs, however was not affected. In this study, recall of weakly related word pairs should require greater reliance on executive control than recall of strongly related word pairs. The authors attributed this finding to both deficient encoding and retrieval and suggested a functional similarity between sleep deprivation, intoxication, and normal aging in terms of controlled memory processes.
Recognition memory is primarily uninfluenced by sleep deprivation, however similar to frontal lobe patients [28
] and older adults [132
], false recognition to semantically related lures has been shown to increase with sleep deprivation (). Diekelmann et al. [134
] used a false memory paradigm [135
] to test false recognition in young subjects sleep deprived during memory retrieval. Subjects sleep deprived during memory retrieval were more likely to incorrectly judge new words semantically related to studied words as “old.” This suggests that forms of recognition memory that depend on the PFC (distinguishing semantically-related lures from studied items) are influenced by sleep deprivation. To further support the view that this effect was not a result of less consistent memory consolidation, this study found that manipulations in sleep within the study-test interval did not influence false recognition. It was specifically the effect of sleep deprivation on retrieval that brought about an increase in false recognition.
Mograss et al. [104
] investigated how ERP old/new effects were influenced by sleep deprivation. They found that the late frontal effect was diminished following total sleep deprivation. This effect is attributable to insufficient PFC function, leading to insufficient retrieval of details, supporting the view that sleep deprivation interferes with PFC function during controlled memory retrieval. If retrieval impairments involve inappropriate allocation of PFC resources, sleep deprivation may have resulted in the misallocation of PFC resources manifested by an under-recruitment
Overall these sleep-deprivation studies of memory retrieval suggest that more controlled retrieval processes are impaired with sleep deprivation compared with more automatic ones and this dissociation may be driven by a breakdown in PFC function.
These results point to the possibility that age-related decreases in sleep contribute to executive control deficits. Conversely, because there is more to age-related sleep changes than a mere overall decrease in sleep, the impact of age-related sleep changes on cognition may not be completely comparable to sleep deprivation in young adults. Further, while many older adults exhibit marked impairments in cognitive performance and changes in brain activity, some older adults show little cognitive decline. These individual differences in cognitive decline may be explained, at least partially, by individual differences in sleep. Determining the way in which sleep plays a role in age-related cognitive decline may shed light on why some older adults but not others exhibit impairments.