Suddendorf & Corballis (S&C) could hardly have asked for a better moment to focus on mental time travel. In recent months there has been a virtual explosion of relevant papers, including neuroimaging studies (Addis et al. Reference Addis, Wong and Schacter2007; Szpunar et al. Reference Szpunar, Watson and McDermott2007), investigations of amnesic patients (Hassabis et al. Reference Hassabis, Kumaran, Vann and Maguire2007) and scrub jays (Raby et al. Reference Raby, Alexis, Dickinson and Clayton2007), a concept essay (Schacter & Addis Reference Schacter and Addis2007b), and two theoretical reviews (Buckner & Carroll Reference Buckner and Carroll2007; Schacter & Addis Reference Schacter and Addis2007a). The future of thinking about the future appears to be now.

We focus on issues raised by S&C that are addressed by our recent work on past and future events. In particular, we consider two key ideas contained in S&C's theater production metaphor: (1) memory is not a literal recording of experience; and (2) episodic future thinking – like episodic remembering – needs to be decomposed into constituent components.

S&C emphasize, rightly, that memory is constructive rather than reproductive. Though not a new idea, its links to future-related thinking have been overlooked (for exceptions, see Dudai & Carruthers Reference Dudai and Carruthers2005b; Suddendorf & Corballis Reference Suddendorf and Corballis1997). To strengthen the link, we have advanced a constructive episodic simulation hypothesis (Schacter & Addis Reference Schacter and Addis2007a; Reference Schacter and Addis2007b). According to this hypothesis, remembering past events and imagining or simulating future events draw on similar kinds of information in episodic memory and involve many shared processes. In particular, episodic remembering and future thinking both depend critically on relational processes that link or bind together distinct elements of an experience. This latter idea is especially important because our hypothesis holds that episodic simulation of future events requires a system that allows the flexible recombination of details from past events into novel scenarios. Episodic memory possesses exactly these characteristics, which makes the system highly adaptive for simulating novel future scenarios based on past experiences, but also makes the system, when used for remembering, prone to errors and distortions that arise from miscombining stored elements – a common form of memory distortion.

This hypothesis predicts considerable overlap in the processes that support remembering the past and imagining the future. S&C review such evidence, which is bolstered further by the recent studies noted earlier. For example, two new neuroimaging studies (Addis et al. Reference Addis, Wong and Schacter2007; Szpunar et al. Reference Szpunar, Watson and McDermott2007) build on earlier work from Okuda et al. (Reference Okuda, Fujii, Ohtake, Tsukiura, Tanji, Suzuki, Kawashima, Fukuda, Itoh and Yamadori2003) by demonstrating that a number of prefrontal, medial temporal, and parietal regions previously associated with episodic remembering show similar increases in activity, relative to control tasks, when imagining the future or recollecting the past. A key finding with important implications for the constructive episodic simulation hypothesis is that the hippocampus shows increased activity when people construct and elaborate on both future and past events; indeed, the right hippocampus shows greater activity during construction of future than past events (Addis et al. Reference Addis, Wong and Schacter2007). Given the role of the hippocampus in supporting relational processing and perhaps other aspects of episodic memory, these data suggest that episodic simulation of future events involves an even more intense constructive process than does episodic remembering of past events.

Such observations mesh nicely with the recent findings from Hassabis et al. (Reference Hassabis, Kumaran, Vann and Maguire2007) that four of five hippocampal amnesics exhibited deficits in the ability to imagine novel events. The fMRI findings also fit with a recent behavioral study in which we demonstrated that older adults, compared with younger adults, show reduced specificity both in their recall of past autobiographical episodes and their imaginings of possible future episodes (Addis et al., in press).

S&C's theater production metaphor highlights the need to decompose the complex activities of mental time travel into more basic components. Such componential approaches have been usefully applied to memory by numerous investigators, and should be similarly helpful when applied to future event simulation. Consider, for example, the finding from our fMRI study that several brain regions show greater activity during future versus past event construction (Addis et al. Reference Addis, Wong and Schacter2007). As noted earlier, one of those regions is the right hippocampus. Another such region is a medial part of right anterior prefrontal cortex (BA 10). If we think of these regions as components a network used for past and future event simulation, the next task is to characterize their respective contributions. In the context of S&C's theater metaphor, we suggest that medial BA 10, assisted by hippocampus, might serve as the “stage” on which the production unfolds.

S&C discuss the stage component with respect to Baddeley's well-known tripartite model of working memory, consisting of a central executive, phonological loop, and visuo-spatial sketchpad. Note, however, that Baddeley (Reference Baddeley2000) updated the model to include a fourth component, the episodic buffer. In the updated model, the central executive is associated with strategic control of processing, whereas the episodic buffer is “a limited capacity system that provides temporary storage of information held in a multimodal code, which is capable of binding information from the subsidiary systems, and from long-term memory, into a unitary episodic representation” (Baddeley Reference Baddeley2000, p. 417).

The function of binding multimodal information from diverse systems makes the episodic buffer well suited to serving as a stage for future event simulations. FMRI data from a working memory task that requires integration of spatial and verbal information indicate that medial BA 10 is a candidate site of the episodic buffer, because it is preferentially activated during maintenance of integrated versus unintegrated information (Prabhakaran et al. Reference Prabhakaran, Narayanan, Zhao and Gabrieli2000). Given the proximity of this region to the one in which we documented greater activity for future imagining than remembering, we are encouraged to postulate that medial BA 10 holds together the diverse elements compromising a future event simulation, and may work closely with the hippocampus in constructing the elements that support the full-blown theater production. Exploring the interrelations of these regions constitutes a critical next step for future-oriented research.