Magnitude of activation in the medial temporal lobe during encoding affects recollection and familiarity.

摘要

The goal of this thesis was to examine in humans how medial temporal lobe (MTL) regions support recognition memory. Recognition memory has been characterized by some researchers as being supported by two processes: recollection and familiarity. Recollection is believed to be dependent upon retrieval of contextual details related to the learning experience, whereas familiarity is thought to depend upon memory of an item's previous occurrence without contextual retrieval. Despite evidence supporting the existence of two processes, the functional neuroanatomy supporting them remains unclear. Three experiments were carried out to explore the neural correlates underlying recollection and familiarity. Experiment 1 used behavioral measures and computational modeling to predict recognition performance on a recognition task with complex pictures in control subjects (n = 8) and in subjects with decreased acetylcholine levels (n = 8). Recollection and familiarity were both shown to be affected by acetylcholine blockade using the drug scopolamine. Based on the behavioral results, an extension of Yonelinas' dual-process model was developed. This model suggests that it was the quality of recollection, but not the number of items remembered, that was affected by scopolamine. Experiment 2 used functional MRI (fMRI) in nine subjects to examine brain regions underlying both recollection and familiarity using behavioral methods designed in Experiment 1. The results revealed that the magnitude of activation during encoding in multiple regions within the MTL distinguished between items that were subsequently recollected and those that were not. However, the distribution of activated regions did not differ between recollection and familiarity. Experiment 3 used fMRI in nine additional subjects to examine how the MTL regions shown to predict recollection in Experiment 2 were affected by the administration of scopolamine. Activation patterns supporting recollection were significantly decreased in perirhinal and fusiform cortex, whereas hippocampal and parahippocampal activation was unaffected. Experiments 1 and 3 support the idea that acetylcholine is important in establishing the conditions during encoding that support recognition. Together the results of these experiments suggest that the strength of recognition might not be mediated by a particular brain region, but may depend upon the magnitude of activation in multiple regions during encoding.