Recognition memory and the medial temporal lobe: a new perspective

Abstract

Recollection has been proposed to be especially dependent on the hippocampus, and familiarity on the adjacent perirhinal cortex. The authors instead suggest that the hippocampus and the perirhinal cortex both play a role in recollection and familiarity, and that these two regions of the medial temporal lobe differ mainly in the degree to which stimuli are encoded in a concrete (in the case of the perirhinal cortex) or abstract (in the case of the hippocampus) manner. The authors suggest that the methods that have traditionally been used to separate recollection from familiarity instead separate strong memories (strong in both recollection and familiarity) from weak memories (weak in both recollection and familiarity). A compelling test of the divided-labour account is provided by comparing the degree of impairment on tests of recall (based on recollection) and recognition (based on recollection and familiarity) in patients with hippocampal lesions. The divided-labour account requires that recall be differentially impaired, but group studies consistently show that recall and recognition are similarly impaired in patients with hippocampal lesions. The effects of selective hippocampal lesions have often been explored in animals using delayed non-matching to sample tasks and novel-object recognition tasks, in which single items must be remembered independently of any context, and where one might suppose that the task depends substantially on familiarity. The evidence suggests that the ability to remember a single item across a delay of more than just a few minutes depends substantially on the hippocampus, even when the task has no overtly associative or contextual component. Analyses of the receiver operating characteristic and of remember–know judgments have often been taken to support the divided-labour model, but a reinterpretation of the evidence in terms of traditional signal-detection theory suggests that many past studies have misconstrued weak memory in patients with hippocampal lesions as evidence for a selective recollection deficit. Different nonlinear relationships between fMRI activity and memory strength in the hippocampus and the perirhinal cortex have often been taken to support the divided-labour view, but those differences are more likely to reflect nonlinear properties of the measurement scale. In the hippocampus, elevated activity is often not detected when memory is weak, and this holds true even for recollection-based memory. Thus, a failure to detect elevated activity for weak memories is not evidence that the hippocampus plays no role in familiarity, but is instead indicative of the nonlinear properties of the measurement scale. In the perirhinal cortex, fMRI activity for strong memories often does not exceed that for memories of moderate strength, and this holds true even for recollection-based memory. Thus, a failure to detect further elevated activity for the strongest memories is not evidence that the perirhinal cortex plays no role in recollection. Single-unit recording studies in monkeys show that neurons in the perirhinal cortex (like neurons in the hippocampus) encode associative information and play a role in associative recollection. Single-unit recording data in humans show that neurons in the hippocampus (like neurons in the perirhinal cortex) encode familiarity and play a role in recognition decisions even when recollection fails.