Periodontitis and all-cause mortality in 60- and 66-year-old individuals: a 23-year prospective cohort study.

In humans, normal memory for everyday facts and events (i.e., explicit or declarative memory) requires a system of anatomically related structures in the medial temporal lobe (MTL) that includes the hippocampus and laterally adjacent entorhinal, perirhinal and parahippocampal cortex (). At a circuit level of analysis, rhinal cortical elements of this system serve as bi-directional relays, conveying the multimodal associative input that the hippocampus operates on in support of memory, and returning the output of hippocampal processing to the distributed networks of neocortical sites that ultimately store memory (). Considerable attention has centered on defining the information processing functions mediated by the component structures of the MTL memory system. This interest derives in part from the translational perspective that the characteristics of impairment that result from MTL damage can provide a window on the underlying circuit dysfunction in disorders that affect memory. For example, the proposal that the hippocampus and entorhinal cortex are differentially impacted in the context of aging and Alzheimer’ disease (AD) might be linked to specific features of impaired memory observed in these conditions. Seminal descriptions of the intensively studied amnesic patient, H.M. () spawned an influential era of nonhuman primate research in neuroscience, aimed at establishing a precise neurology of memory, i.e., a comprehensive account of the MTL structures damaged in H.M. that are responsible for his profound inability to form new memories for ongoing events. The development of a simple, easily administered test of recognition memory for monkeys proved key to the success of that effort (; ). Indeed, the failure of early attempts to model H.M.’s amnesia in monkeys is now understood to be a consequence of using testing procedures that fail to engage the particular form of memory that requires the MTL system. Visual object recognition, tested by asking monkeys to make either matching or nonmatching selections between novel stimuli and items viewed prior to a delay, proved much more sensitive. Most often administered in a nonmatching format, with trial-unique object pairs, demands on memory can be parametrically manipulated in this task by increasing the retention interval over which sample items need to be recognized. Influential lesion studies in monkeys using DNMS revealed core features of MTL functional organization. For example, among the brain regions resected in patient H.M., the amygdala is not required for normal performance (), nor does amygdala damage exacerbate the DNMS impairment that results from lesions involving other MTL areas (). In contrast, the entorhinal, perirhinal and parahippocampal cortices adjacent to the hippocampus - originally considered sites of incidental damage in H.M. - are critically important components of the MTL system (; ; c.f., ; ). Indeed, substantial involvement of these rhinal cortical areas appears both necessary and sufficient for an MTL lesion to yield recognition memory impairment of a severity that approaches H.M.’s amnesia (; ; ). Based on the anatomical circuit organization of the MTL system, however, it might also be that the devastating memory effects of rhinal cortical damage are attributable in part to the downstream disruption of hippocampal processing caused by such lesions. Unfortunately, limited evidence is available, particularly in monkeys, that might distinguish processing capacities intrinsic to the rhinal cortex from the contribution of the hippocampus to broader MTL memory function. For an illuminating retrospective on this body of work, see . Alongside foundational contributions to understanding the brain organization of memory, lesion studies of MTL amnesia in nonhuman primates gave rise to one of the more enduring puzzles in systems and behavioral neuroscience. Using a wide variety of experimental approaches, the consensus from research in animal models and humans is that normal episodic/declarative memory requires the hippocampus proper, and that damage or dysfunction restricted to the hippocampus can result in modest, but clinically significant memory impairment (; ). Thus, it is striking that studies examining the effects of selective lesions of the hippocampus on DNMS performance in monkeys have reported essentially all possible outcomes, from no effect whatsoever (), to mild or moderate impairment under conditions of increased memory demand (; ), to substantial impairment in both initial task acquisition and recognition accuracy on the delay component of DNMS, independent of retention interval length (). Many factors have been speculated to account for these disparities, including lesion-related factors (e.g., variation in volume, extent, off-target involvement), procedural differences (e.g., DNMS stimulus set size or behavioral test history), and experimental design considerations (e.g., sample size and statistical power, and pre-operative