Mahr & Csibra (M&C) dismiss future-oriented mental time travel as a central function of episodic memory (sect. 2.1.). In this commentary, we seek to challenge two central arguments to their conclusion: (1) their argument that deficient episodic memory does not impair future-oriented decisions and (2) their argument that a functional account of episodic memory based on future-directed mental time travel struggles to explain why it is possible to veridically recall past events.

There has been tremendous interest in mental time travel over the past decade, with a growing number of studies examining our capacity to imagine hypothetical episodes that either may take place in the future or that – counterfactually – could have happened in the past (Schacter et al. Reference Schacter, Benoit, De Brigard and Szpunar2015). These studies have revealed striking similarities between such episodic simulation and episodic memory. For example, as mentioned by M&C, the two capacities are supported by the same core network of brain regions (Addis et al. Reference Addis, Wong and Schacter2007; Benoit & Schacter Reference Benoit and Schacter2015; Hassabis et al. Reference Hassabis, Kumaran and Maguire2007a; Szpunar et al. Reference Szpunar, Watson and McDermott2007), are similarly deficient in amnesic patients (Hassabis et al. Reference Hassabis, Kumaran, Vann and Maguire2007b; Klein et al. Reference Klein, Loftus and Kihlstrom2002b; Race et al. Reference Race, Keane and Verfaellie2011, but see also Squire et al. Reference Squire, van der Horst, McDuff, Frascino, Hopkins and Mauldin2010), and also seem to exhibit parallel life span developmental trajectories (Addis et al. Reference Addis, Wong and Schacter2008; Busby & Suddendorf Reference Busby and Suddendorf2005). These observations have been taken to suggest that episodic simulation is based on an episodic memory system that provides stored details and constructive processes to recombine such details into novel events (Schacter et al. Reference Schacter, Addis and Buckner2007).

Central to our commentary is a common feature of episodic memory and episodic simulation that is grounded in their autonoetic format: The two capacities allow for the experience of “what it felt like” in past events, “what it could have felt like” in counterfactual events, and “what it would feel like” in prospective events. In the following, we describe how these mental experiences can facilitate farsighted decisions. We thereby seek to challenge the aforementioned two arguments.

First, as M&C rightly point out, despite their inability to imagine coherent future episodes, amnesic patients are not generally blind to the future (Craver et al. Reference Craver, Kwan, Steindam and Rosenbaum2014b). A paradigmatic example for future-oriented decisions consists of situations in which we have to choose between a smaller reward that we can receive immediately and a larger reward that we would receive only at a later time. People often make myopic decisions for the smaller reward in such situations, because they tend to discount the value of delayed rewards as a function of the time they would have to wait. M&C cite important evidence that amnesic patients don't discount future rewards more strongly than healthy people, despite their deficiency in simulating future episodes (Kwan et al. Reference Kwan, Craver, Green, Myerson, Boyer and Rosenbaum2012; see also Kwan et al. Reference Kwan, Craver, Green, Myerson and Rosenbaum2013). These data thus seem to indicate that mental time travel does not contribute to future-oriented decisions. However, although amnesic patients typically don't show exaggerated discounting, we suggest, as detailed in the following, that they lack a particular mechanism that can adaptively attenuate such impulsive tendencies.

A possible reason for temporal discounting is that we generally don't experience the anticipated emotional impact of a future reward when making a decision (e.g., Rick & Loewenstein Reference Rick and Loewenstein2008). However, by simulating the future moment of consuming the reward, we can mentally create this experience. This simulated experience, in turn, has been hypothesized to increase the valuation of the delayed reward and, consequently, to attenuate discounting (Benoit et al. Reference Benoit, Gilbert and Burgess2011; Boyer Reference Boyer2008). A growing number of studies have provided support for this hypothesis (e.g., Benoit et al. Reference Benoit, Gilbert and Burgess2011; Liu et al. Reference Liu, Feng, Chen and Li2013; O'Donnell et al. Reference O'Donnell, Oluyomi Daniel and Epstein2017; Palombo et al. Reference Palombo, Keane and Verfaellie2015; Peters & Büchel Reference Peters and Büchel2010). Critically, there is evidence that patients with hippocampal damage or atrophy following Alzheimer's disease, who are impaired at imagining future events, do not show reduced discounting in situations nominally fostering episodic simulation (Lebreton et al. Reference Lebreton, Bertoux, Boutet, Lehericy, Dubois, Fossati and Pessiglione2013; Palombo et al. Reference Palombo, Keane and Verfaellie2015; but see Kwan et al. Reference Kwan, Craver, Green, Myerson, Gao, Black and Rosenbaum2015, where participants may have benefited from semantic, rather than episodic, future simulation; for further discussion, see Schacter et al. Reference Schacter, Benoit and Szpunar2017). Although episodic memory is not necessary for all future-oriented decisions, we thus argue that – due to its autonoetic format – it conveys prospective emotions that can render such decisions more farsighted, thereby supporting a fitness-relevant mechanism that maximizes future benefits.

Second, M&C argue that a future-directed account of episodic memory struggles to explain why memories can be veridically recalled. Specifically, they suggest that re-experiencing a past episode does not contribute to future planning beyond what can be provided by semantic memory. In their example, one can infer that there is going to be a long line at the swimming pool without re-experiencing the extended wait during one's last visit. By contrast, we suggest that the emotions elicited by episodically remembering a past event (D'Argembeau et al. Reference D'Argembeau, Comblain and Van der Linden2003) can further serve to motivate future plans (e.g., avoiding the pool to prevent repeated boredom). Critically, this is only the case to the degree that our memories are veridical (e.g., if the line was indeed that long).

Moreover, memories of actual experiences can be contrasted with simulations of counterfactual events (e.g., “If instead we had taken a trip to the lake …”). These counterfactuals induce emotional responses, such as regret, that further intensify with repeated simulations (Stanley et al. Reference Stanley, Parikh, Stewart and De Brigard2017) and that have been shown to influence prospective choices (Camille et al. Reference Camille, Coricelli, Sallet, Pradat-Diehl, Duhamel and Sirigu2004). Counterfactual simulations of reliably remembered events can thus motivate future-oriented decisions, which we argue constitutes an adaptive role for the veridical recall of the past (see also Schacter et al. Reference Schacter, Benoit, De Brigard and Szpunar2015).

To conclude, we propose that episodic simulation – due to its autonoetic format – allows for the experience of prospective and counterfactual emotions that convey motivational incentives for farsighted decisions. We don't assume that there is necessarily only one central function to episodic memory, nor do we challenge a possible contribution to human communication. However, we suggest that future-oriented mental time travel – with its outlined adaptive value – remains a candidate that should not readily be dismissed.