Memory and Imagination: Perspectives on Constructive Episodic Simulation

Abstract

Questions concerning the relation between memory and imagination have occupied a prominent place in cognitive psychology and cognitive neuroscience during the past several decades. Johnson and Raye’s (1981) reality monitoring framework brought the issue into prominence by portraying the relation as constituting a fundamental problem that must be addressed by any cognitive model. Johnson and Raye’s framework provided the inspiration for subsequent cognitive and neuroimaging studies concerning the mnemonic qualities of perceived and imagined events, and the underlying psychological and neural processes that allow us to distinguish between them (for an update, see Simons, Garrison, and Johnson, 2017). A little more than a decade ago, a related line of research emerged that focused on striking neural and cognitive similarities that occur when people remember past experiences and imagine or simulate future and hypothetical experiences. The foundation of this work was set in the 1980s and 1990s by clinical observations that an amnesic patient with a severe deficit in recalling past experiences also had comparable difficulties imagining future experiences (Tulving, 1985; see also, Klein, Loftus, and Kihlstrom, 2002) and that suicidally depressed patients showed similar reductions in the specificity of autobiographical memories and future imaginings (Williams et al., 1996). Observations that remembering past experiences and imagining future experiences show similar neural correlates (Okuda et al., 2003), phenomenological characteristics (D’Argembeau and van der Linden, 2004) and individual differences (D’Argembeau and van der Linden, 2006) heightened interest in the relation between the two. This growing interest was galvanized by the publication in 2007 of several papers that strengthened the case for a link between remembering and imagining by revealing strong overlap in the brain regions engaged when people remembered past experiences and either imagined future experiences (Addis, Wong, and Schacter, 2007; Szpunar, Watson, and McDermott, 2007) or imagined novel scenes (i.e. scene construction: Hassabis, Kumaran, and Maguire, 2007a). This evidence from functional magnetic resonance imaging (fMRI) was complemented by neuropsychological evidence showing that several amnesic patients with medial temporal lobe damage exhibit scene construction deficits (Hassabis et al., 2007b). Based on these and related observations, Schacter, Addis, and Buckner (2007; see also, Buckner and Carroll, 2007) pointed to the existence of a core neural network,