ABSTRACT:The role of the hippocampus in recent spatial memory has been well documented in patients with damage to this structure, but there is now evidence that the hippocampus may not be needed for the storage and recovery of a spatial layout that was experienced long before injury. Such preservation may rely, instead, on a network of dissociable, extra-hippocampal regions implicated in topographical orientation. Using functional magnetic resonance imaging (fMRI), we investigated this hypothesis in healthy individuals with extensive experience navigating in a large-scale urban environment (downtown Toronto). Participants were scanned as they performed mental navigation tasks that emphasized different types of spatial representations. Tasks included proximity judgments, distance judgments, landmark sequencing, and blocked-route problem-solving. The following regions were engaged to varying degrees depending on the processing demands of each task: retrosplenial cortex, believed to be involved in assigning directional significance to locales within a relatively allocentric framework; medial and posterior parietal cortex, concerned with processing space within egocentric coordinates during imagined movement; and regions of prefrontal cortex, present in tasks heavily dependent on working memory. In a second, eventrelated experiment, a distinct area of inferotemporal cortex was revealed during identification of familiar landmarks relative to unknown buildings in addition to activation of many of those regions identified in the navigation tasks. This result suggests that familiar landmarks are strongly integrated with the spatial context in which they were experienced. Importantly, right medial temporal lobe activity was observed, its magnitude equivalent across all tasks, though the core of the activated region was in the parahippocampal gyrus, barely touching the hippocampus proper.

Preserved remote spatial memory in amnesic people with bilateral hippocampal damage, including the well-studied case K.C., challenges spatial theories, which assume that the hippocampus is needed to support all allocentric spatial representations, old or new. It remains possible, however, that residual hippocampal tissue is functional and contributes to successful performance. Here, we examine brain activity with fMRI during the retrieval of spatial information in K.C. and in healthy controls using landmark and route stimuli from a premorbidly familiar neighborhood that K.C. can navigate normally. In all participants, activity was found in the parahippocampal cortex, but not in the hippocampus itself, during all navigational tasks on which K.C. performs well, even though part of his hippocampus remains viable. The opposite pattern was observed on a house recognition task, which is inconsequential to navigation, and on which K.C. performed poorly. On that task, K.C. recruited the right hippocampus presumably because even "familiar" houses were treated as novel by him, whereas controls recruited occipitotemporal cortex, including parahippocampal cortex. The distinction between recent and remote memory, therefore, may apply as much to spatial theories of hippocampal function as it does to theories emphasizing the role of the hippocampus in other types of explicit memory.

No abstract

Because emotional and neutral stimuli typically differ on non-emotional dimensions, it has been difficult to determine conclusively which factors underlie the ability of emotional stimuli to enhance immediate long-term memory. Here we induced arousal by varying participants' goals, a method that removes many potential confounds between emotional and non-emotional items. Hungry and sated participants encoded food and clothing images under divided attention conditions. Sated participants attended to and recalled food and clothing images equivalently. Hungry participants performed worse on the concurrent tone-discrimination task when they viewed food relative to clothing images, suggesting enhanced attention to food images, and they recalled more food than clothing images. A follow-up regression analysis of the factors predicting memory for individual pictures revealed that food images had parallel effects on attention and memory in hungry participants, so that enhanced attention to food images did not predict their enhanced memory. We suggest that immediate long-term memory for food is enhanced in the hungry state because hunger leads to more distinctive processing of food images rendering them more accessible during retrieval.

Summary. For over a hundred years, it has been accepted that remote memories are less vulnerable to disruption than are recent memories. The standard consolidation model posits that the hippocampus and related structures are temporary memory structures, necessary for acquisition, retention, and retrieval of all explicit (declarative) memories until they are consolidated elsewhere in the brain. We review lesion and neuroimaging evidence showing that important distinctions exist among different types of explicit memory and the structures that mediate them. We argue that retention and retrieval of detailed, vivid autobiographical memories depend on the hippocampal system no matter how long ago they were acquired. Semantic memories, on the other hand, benefit from hippocampal contribution for some time before they can be retrieved independently of the hippocampus. Even semantic memories, however, can have episodic elements associated with them which continue to depend on the hippocampus. In short, the evidence reviewed suggests strongly that the function of the 334 Moscovitch, Westmacott, Gilboa, et al. hippocampus (and possibly of related limbic structures) is to help encode, retain, and retrieve experiences, no matter how long ago the events comprising the experience occurred. We conclude that the evidence favors a multiple trace theory (MTT) of memory over the traditional model, and we indicate what future work is needed to resolve disputes.