We consider Walker's thorough review in the context of thinking about future research on the relation between sleep and memory. We first address methodological issues including type of memory and sleep-stage dependency. We suggest a broader investigation of potential signaling molecules that may be critical to sleep-related consolidation. A brief review of the importance of the stress hormone cortisol illustrates this point.

Walker proposes that procedural memory formation involves two specific stages of consolidation: wake-dependent stabilization, followed by sleep-dependent enhancement. If sleep-based enhancement of procedural memory formation is now well supported by evidence obtained at different levels of cognitive and neurophysiological organization, wake-dependent mechanisms for stabilization have not been demonstrated as convincingly, and still require more systematic characterization.

Matthew Walker's article has prompted us to consider neuropsychiatric disorders and pharmacological effects associated with sleep alterations, and aspects of memory affected. Not all disorders involving insomnia show memory impairment, and hypersomnias can be associated with memory deficits. The use of cholinergic medication in dementia indicates that consideration of the link between sleep and memory is more than academic.

This commentary implicates the neostriatum in the production of the EEG sleep spindle and in the processing of motor procedural learning in sleep. Whether the sleep spindle may implement not only the consolidation-based enhancement of procedural learning, but also its initial consolidation, is considered; as is the fit between (1) corticostriatal anatomy and physiology, and (2) the physical properties of the sleep spindle.

In this commentary the “incredibly robust” evidence for the relationship between sleep and procedural memory is questioned; inconsistencies in the existing data are pointed out. In addition, some suggestions about extending research are made, for example, studying REM sleep augmentation or memory consolidation in patients with sleep disorders. Last, the possibility of a relationship between dreaming and memory processes is discussed.

Does sleep restore brain function or does it consolidate memory? I argue that memory consolidation during sleep is an offshoot of restitution. Continual learning causes local synapse-specific neural fatigue, which then masks expression of that learning, especially on time-limited tests of procedural skills. Sleep serves to restore the fatigued synapses, revealing the consolidation-based enhancement observed as a “latent” overnight improvement in learning.

Initial claims that REM sleep is important in the consolidation of all memories have been revised and reduced to the claim that sleep has a role only in the consolidation of procedural learning. Now, Walker hypothesizes that sleep has no role in the “stabilization phase of consolidation” but only in the “enhanced learning” phase of procedural learning. Evidence for this vague, truncated hypothesis remains as inconsistent as that for prior claims.

The Walker model raises a number of questions, particularly about the nature of the sleep states involved in consolidation enhancement. While REM sleep, Stage 2 sleep, and Stage 3/4 sleep have been implicated in procedural learning, we still do not understand which types of learning are involved with specific sleep states. Several possible ideas for future research are suggested.

Cellular mechanisms hypothesized to underlie sleep-dependent memory consolidation are expressed throughout the brain during sleep. Use of sleep deprivation to evaluate the functional importance of these mechanisms is confounded by degradation in waking performance resulting from impaired vigilance. There is a need for methods that will permit disruption of specific mechanisms during sleep only in the neuronal circuits most critically involved in learning. This should be accomplished without global sleep disruption and with preservation of the restorative aspects of sleep.

We propose that sleep is linked to synaptic homeostasis. Specifically, we propose that: (1) Wakefulness is associated with synaptic potentiation in cortical circuits; (2) synaptic potentiation is tied to the homeostatic regulation of slow wave activity; (3) slow wave activity is associated with synaptic downscaling; and (4) synaptic downscaling is tied to several beneficial effects of sleep, including performance enhancement.

Although considerable attention has been paid to the possible involvement of sleep in memory processing, there is no substantial evidence for it. Walker describes a phenomenon of consolidation-based enhancement (CBE), whereby performance on select procedural tasks improves with overnight sleep; that is, without additional practice on the tasks. CBE, however, appears restricted to a few tasks, and even with these tasks CBE is not confined to sleep but also occurs during wakefulness. Sleep serves no unique role in this process. At best, CBE is a slow, time-dependent process of consolidation that begins with task acquisition in waking and can under some circumstances extend to sleep.

Following on from the target article, which presented a new model of procedural skill memory development, in this response I will reflect on issues raised by invited commentators and further expound attributes of the model. Discussion will focus on: evidence against sleep-dependent memory processing, definitions of memory stages and memory systems, and relationships between memory enhancement, sleep-stages, dreaming, circadian time, and sleep-disorders.