Certain method of loci (MoL) prerequisites – familiar, coherently ordered locations – should appear during dreaming if the latter is, in fact, elaborative memory encoding as hypothesized by Llewellyn. A review of the literature suggests that dreamed locations are neither familiar nor coherently ordered and thus unsuitable for facilitating memory in this sense. This conclusion converges with other evidence that episodic memory is dependent upon non–rapid eye movement (NREM), rather than REM, sleep.

According to the tuning-for-criticality theory, the essential role of sleep is to protect the brain from super-critical behaviour. Here we argue that this protective role determines the content of dreams and any apparent relationship to the art of memory is secondary to this.

Toward illuminating the structure of Llewellyn's dream theory, I compare it in formal terms to Freud's dream theory. An alternative to both of these dream machines, grounded in the distribution of cholinergic activation in the central nervous system, is presented. It is suggested that neither “high” nor “low” dream theory is sufficient to account for the properties of dreams.

The hypothesis that dreaming is involved in off-line memory processing is difficult to test because major methodological issues have to be addressed, such as dream recall and the effect of remembered dreams on memory. It would be fruitful – in addition to studying the ancient art of memory (AAOM) in a scanner – to study the dreams of persons who use AAOM regularly.

Links with the Humanities are to be welcomed, but they cannot be exempted from normal scientific criteria. Any hypothesis regarding the function of dreams that is premised on rapid eye movement (REM)/dream isomorphism is unsupportable on empirical grounds. Llewellyn's hypothesis has the further problem of counter-evidence in respect of its claim that dreaming relies upon hippocampal functions. The hypothesis also lacks face validity.

Neuroimaging studies show that episodic memory encoding is associated with increased activity in hippocampus and lateral prefrontal cortex; however, the latter structure shows decreased activity in rapid eye movement (REM) sleep. Together with few episodic memory traces in REM sleep, and REM sleep deprivation affecting hippocampus-independent emotional processes, this argues for generic information processing in REM sleep rather than linking episodic memory traces.

Llewellyn has written a fascinating article about rapid eye movement (REM) dreams and how they promote the elaborative encoding of recent memories. The main message of her article is that hyperassociative and fluid cognitive processes during REM dreaming facilitate consolidation. We consider one potential implication of this analysis: the possibility that excessive or out-of-phase REM sleep fuels dissociative symptomatology. Further research is warranted to explore the psychopathological ramifications of Llewellyn's theory.

Rapid eye movement (REM) dreaming results in “emotionally intelligent encoding,” according to the target article. Building on this, we argue that elaborative encoding alters emotional processing of upcoming events and thereby functions as prospective emotion regulation. After elaborative encoding, future events are appraised differently and result in a redirected emotional response. Disturbed elaborative encoding might be relevant for emotional dysregulation in psychopathology.

Can ancient art of memory (AAOM) principles explain the function of dreaming? The analysis of self-experience in dreams suggests that the answer is no: The phenomenal dream self lacks certain dimensions that are crucial for the efficacy of AAOM in wakefulness. However, the comparison between dreams and AAOM may be fruitful by suggesting new perspectives for the study of lucid dreaming as well an altered perspective on the efficacy of AAOM itself.

I argued that rapid eye movement (REM) dreaming is elaborative emotional encoding for episodic memories, sharing many features with the ancient art of memory (AAOM). In this framework, during non–rapid eye movement (NREM), dream scenes enable junctions between episodic networks in the cortex and are retained by the hippocampus as indices for retrieval. The commentaries, which varied in tone from patent enthusiasm to edgy scepticism, fall into seven natural groups: debate over the contribution of the illustrative dream and disputes over the nature of dreaming (discussed in sect. R1); how the framework extends to creativity, psychopathology, and sleep disturbances (sect. R2); the compatibility of the REM dream encoding function with emotional de-potentiation (sect. R3); scepticism over similarities between REM dreaming and the AAOM (sect. R4); the function of NREM dreams in the sleep cycle (sect. R5); the fit of the junction hypothesis with current knowledge of cortical networks (sect. R6); and whether the hypothesis is falsifiable (including methodological challenges and evidence against the hypothesis) (sect. R7). Although the groups in sections R1–R6 appear quite disparate, I argue they all follow from the associative nature of dreaming.

Why does performing certain tasks cause the aversive experience of mental effort and concomitant deterioration in task performance? One explanation posits a physical resource that is depleted over time. We propose an alternative explanation that centers on mental representations of the costs and benefits associated with task performance. Specifically, certain computational mechanisms, especially those associated with executive function, can be deployed for only a limited number of simultaneous tasks at any given moment. Consequently, the deployment of these computational mechanisms carries an opportunity cost – that is, the next-best use to which these systems might be put. We argue that the phenomenology of effort can be understood as the felt output of these cost/benefit computations. In turn, the subjective experience of effort motivates reduced deployment of these computational mechanisms in the service of the present task. These opportunity cost representations, then, together with other cost/benefit calculations, determine effort expended and, everything else equal, result in performance reductions. In making our case for this position, we review alternative explanations for both the phenomenology of effort associated with these tasks and for performance reductions over time. Likewise, we review the broad range of relevant empirical results from across sub-disciplines, especially psychology and neuroscience. We hope that our proposal will help to build links among the diverse fields that have been addressing similar questions from different perspectives, and we emphasize ways in which alternative models might be empirically distinguished.

Self-control is a necessary component of subjective effort, but it depends only on farsighted motivation, with no additional, depletable resource. The aversiveness of boring tasks probably comes from their interference with endogenous reward, a new and potentially controversial concept. The self-control needed to stick with any kind of aversive experience increases as the values of the competing motives draw closer together.

The cognitive impairments shown by brain-damaged patients emphasize the role of task difficulty as a major determinant for performance. We discuss the proposal of Kurzban et al. in light of our findings on right-hemisphere–damaged patients, who show increasing awareness deficits for the contralesional hemispace when engaged with resource-consuming dual tasks. This phenomenon is readily explained by the assumption of unspecific depletable resources.

Though Kurzban et al.'s model explains a considerable set of empirical findings, it cannot accommodate other results without relying on extra assumptions. We offer an addition to the model, and suggest that cost-benefit analyses themselves depend on executive function, and therefore can be biased. The adapted model allows for explaining depletion effects, as well as their reversals, documented in the literature.

As a comment on Kurzban et al.'s opportunity cost model, we propose an alternative view of mental effort and the action selection mechanism in the brain. Our hypothesis utilizes local resource depletion within neuronal networks, which justifies from a neurophysiological perspective why mental fatigue diminishes after switching to a novel task and explains action selection by means of neural competition theory.

The depletion effect, a decreased capacity for self-control following previous acts of self-control, is thought to result from a lack of necessary psychological/physical resources (i.e., “ego depletion”). Kurzban et al. present an alternative explanation for depletion; but based on statistical techniques that evaluate and adjust for publication bias, we question whether depletion is a real phenomenon in need of explanation.

Kurzban et al. argue that the experiences of “effort,” “boredom,” and “fatigue” are indications that the costs of a task outweigh its benefits. Reducing the costs of tasks to “opportunity costs” has the effect of rendering tasks costless and of denying that they can be inherently boring or tedious, something that “vigilance tasks” were intentionally designed to be.

We argue that maximising utility does not promote survival. Hence, there is no reason to expect people to modulate effort according to a task's opportunity costs. There is also no reason why our evaluation of the marginal opportunity costs of tasks should predictably rise with repetition. Thus, the opportunity cost model cannot explain why tasks typically become harder over time.

A homeostatic energy model of awareness proposes that the anterior insular cortex engenders feelings that provide an amodal valuation of homeostatic energy utilization in an opponent, bivalent emotional control system. Feelings are the “common currency” which enable optimal utilization in the physical and mental behavior of a highly social primate. This model offers a different perspective.

We welcome the development of a new model on effort and performance and the critique on existing resource-based models. However, considering the vast evidence for the significant impact of experienced task demand on resource allocation, we conclude that Kurzban et al.'s opportunity cost model is only valid for one performance condition: if task demand is unknown or unspecified.