The prevailing model of conscious memory holds that a group of loosely related structures, collectively called the medial temporal lobe (MTL), work together as a single functional unit. According to this current orthodoxy, the structures in the MTL are thought to store the memory of specific objects, facts and events. This model has many adherents despite the fact that it is contradicted by a large body of empirical data. We have developed a competing hypothesis, which is consistent with these data. Our model holds that different structures within the MTL have different functions, and that their collective functions extend beyond those assigned to the MTL by the orthodox model. To address the first goal, we have considered the evolutionary history of components of the MTL. One part of the MTL, the perirhinal cortex, has been implicated in visual recognition and association memory. Despite its small size, the perirhinal cortex plays a central role in understanding the cerebral cortex, vision and memory;it figures in discussions of cognitive capacities as diverse as object perception, semantic knowledge, feelings of familiarity and conscious recollection. Two conceptual constructs have encompassed perirhinal cortex. The current orthodoxy incorporates perirhinal cortex within the MTL as a memory area;an alternative considers it to be a sensory area with a role in both perception and memory. A historical perspective provides insight into both of these ideas. Perirhinal cortex came to be included in the MTL because of two accidents of history. In evolutionary history, the hippocampus migrated from its ancestral situation, as medial cortex, into the temporal lobe;in the history of neuropsychology, a memory system that originally consisted of the amygdala and hippocampus came to include perirhinal cortex. These two histories explain why a part of the sensory neocortex, perirhinal cortex, entered into the conceptual construct called the MTL. They also explain why some experimental results seem to exclude a perceptual function for this sensory area, while others embrace perception. The exclusion of perceptual functions results from a history of categorizing tasks as perceptual or mnemonic, often on inadequate grounds. By instead exploring the role of perirhinal cortex in encoding, representing and retrieving stimulus information, it can be understood as a part of the sensory neocortex, one that has much the same relationship with the hippocampus as do other parts of the neocortex that evolved at about the same time. As for the neural substrates of associative memory, our work has made use of two measures, both involving objects. One type of memory, known as paired associate learning (PAL), involves arbitrary, nonspatial associations between two different visual stimuli. A second measure of memory involves arbitrary associations between visual stimuli and spatially-directed responses, known as visuomotor learning (VML). Amnesic patients are notoriously poor in acquiring arbitrary associations;by tapping into these kinds of associative memory this work makes a direct link to human amnesia. Our past work has shown that transection of the fornix, like aspirations lesions of the hippocampal formation plus underlying cortex, disrupt VML. To move this work forward, we investigated what part of the hippocampal system (hippocampus proper, subicular complex, or entorhinal cortex) is essential for VML. At the same time, we compared performance on the conceptually related PAL. In this work we examined the contributions of brain regions to memory by using temporary inactivations rather than static lesions. This approach provides within-subject measures of memory that can be more sensitive than comparisons across groups. Recently, we found that bilateral infusions of a drug intended to temporarily inactivate the hippocampus, relative to saline infusions, failed to disrupt performance on either task. These findings strongly suggest that structures outside the hippocampus proper, either alone or together with the hippocampus, contribute to performance on these tasks. In the period under review we extended this work to investigate the effects of temporary inactivation of the entorhinal cortex, bilaterally, on the same two tasks. The entorhinal cortex is part of the MTL, contains neurons that give rise to axons passing through the fornix, and is included in aspiration lesions of the hippocampal formation. Thus, damage to entorhinal cortex may account for the aforementioned effects of lesions of the fornix and hippocampal formation on VML. Our preliminary data show that, unlike hippocampal inactivations, entorhinal cortex inactivations disrupt VML performance. We were unable to evaluate the effects of entorhinal cortex inactivation on PAL. Because the behavioral effects of the entorhinal cortex inactivations were tested using the same task, drug concentrations and infusion procedures as for the hippocampus, these data indicate that our negative result after hippocampal inactivations is not due to lack of efficacy of the infusate. Instead, the results suggest that the entorhinal cortex, but not the hippocampus, is essential for rapid acquisition of VML problems. A third goal of this project is to test hippocampal-prefrontal interactions in fast associative learning. We predict that both the fornical and nonfornical pathways from the entorhinal cortex to the prefrontal cortex are critical to these kinds of rapid learning. We currently have no plans to pursue this goal. In collaboration with another investigator, we have investigated the contribution of the hippocampus proper to memory using additional methods. Although extensive MTL damage in humans and animals causes robust anterograde amnesia, defining the contributions of the hippocampus proper has proved challenging. We tested the memory of adult animal subjects that received excitotoxin lesions (N-methyl-D- aspartic acid;NMDA) involving extensive, MRI-confirmed damage restricted to the hippocampus. Unoperated animals served as controls. Subjects were tested on a battery of tasks including a standardized version of the trial-unique, delayed nonmatching-to-sample (DNMS) task, and a novel DNMS variant that employed a small set of discriminative stimuli used repeatedly across test sessions. Assessment also included the classic delayed response task. Behavioral testing failed to reveal deficits on any task. These findings suggest that traditional assessments used in development of a model of MTL amnesia fail to capture the key operating characteristics of declarative/episodic memory mediated by the primate hippocampus.
