Neural correlates of recognition memory for complex visual stimuli in the medial temporal lobe.

While a general role for the medial temporal lobe (MTL) in long-term memory is undisputed, the specific contributions made by MTL subregions remain contentious. For some accounts, separate but stimulus-independent regions support two different memory processes/types of mnemonic information (recollection and familiarity/items and contexts). According to another account, memory processes are not the primary organising principle of MTL function. Instead, sub-regions process different kinds of stimuli - for example, objects and scenes. The three experiments in this thesis were designed to compare competing models of MTL function by measuring, using functional magnetic resonance imaging, the neural correlates of successful recollection- and familiarity-based memory judgements for different types of complex visual stimuli. Experiment 1 was designed to explore differences between encoding- and retrieval-related neural activity for faces and scenes. There was some evidence for stimulus- specific memory processing within the perirhinal cortex, anterior hippocampus and parahippocampal gyms for faces, and the posterior hippocampus for scenes. The data from Experiment 1, however, offered limited insights into the processes that supported memory for these different stimulus types. In Experiment 2 encoding-related activity was assessed in a paradigm where participants had to indicate in which of two contexts/sources objects and scenes had been encountered in a study phase. In the hippocampus and parahippocampal gyrus there was activity predicting successful source encoding only for scenes. Activity in the perirhinal cortex, meanwhile, predicted item memory but not source memory for objects only. These findings are consistent with claims that stimulus-type is important for MTL function and aligns the hippocampus and perirhinal cortex with the processes of recollection and familiarity, respectively. Experiment 3 was designed to investigate MTL contributions to source memory for objects, incorporating changes to the experiment design that were implemented in light of consideration of reasons for the absence of encoding-related activity predicting source memory for objects in Experiment 2. The critical finding was neural activity in the perirhinal cortex that predicted accurate source memory for objects. Collectively these results are consistent with views that the perirhinal cortex and hippocampus are differentially involved in processing objects and scenes, rather than in supporting distinct kinds of memory process