The Electrophysiological Correlates of the Perceptual Representation System

Abstract

Tulving and Schacter proposed that there is a specific memory system responsible for priming, which they termed the Perceptual Representation System (PRS). It is considered responsible for representing the form and structure of stimuli rather than information about meaning, it is highly inflexible in its representation of stimuli, and is thought to comprise three domain-specific subsystems: the visual word form subsystem, the auditory word form subsystem, and the structural description subsystem. The purpose of the present programme of studies was to examine the properties of subsystems of the PRS underlying priming. Evidence for the PRS comes from a number of sources: neuropsychological, experimental dissociations in non-clinical samples, and neuroimaging studies. Although informative with respect to the theory, each of these methodologies has its own limitations. A procedure developed by Rugg et al. (1998) overcame a number of previous limitations of studies of priming and provided a tool for examining the subsystems of the PRS. Rugg et al. (1998), using ERPs, observed two qualitatively different ERP old/new effects associated with visual word-form priming (i.e., N400 old/new effect) and recollection memory (i.e., LPC old/new effect), and on this basis reported the first dissociation of the visual word form subsystem from explicit memory. Importantly, this dissociation was achieved when the two types of memory were tested concurrently, rather than using different tasks, and in a way that minimised the possible confounding of priming by explicit memory. Studies 1a, 1b, and 2 reported in Chapter 3 were directed to replicating the findings of Rugg et al. (1998) using highly similar procedures. Visually presented words were used to identify neural correlates of priming and explicit memory. The substantial success of these studies made possible an extension of the experimental work to the study of objects as stimuli (Studies 3a and 3b reported in Chapter 4) and to words presented aurally (Studies 4a and 4b in Chapter 5). Thus the three subsystems of the PRS were examined. Prior to this research programme there had not been any replications of the findings of Rugg et al. (1998) reported in the literature and no attempts to use stimuli in domains other than the visual. Using visually presented words, objects, and aurally presented words, priming and explicit memory effects were identified. Across all modes of stimuli, the explicit memory effect was indexed between 500 and 800 ms post-stimulus, and was maximal at left parieto-central regions. The presence of the LPC old/new effect at the left parietal electrode site may be explained as reflecting the operations of verbal memory, which is localised to the left hemisphere. A N400 old/new effect that appeared to index priming was also present across each mode of stimuli. Spatially, this effect was present most consistently across tests at parietal sites and specifically the right parietal site. This may suggest that a common area that subserves the PRS is located in the posterior regions of the cortex. These outcomes support findings from neuroimaging studies that have identified priming effects associated with visual word, object, and auditory word priming tasks in secondary sensory and associative areas of the cortex. Each mode of priming displayed N400 old/new effects at the right hemisphere, a spatial distribution that was not generally evident for the explicit memory effect. This outcome (right lateralisation) supports the proposition that the right hemisphere is engaged more than the left hemisphere in perceptual priming. The right hemisphere advantage associated with perceptual priming, has been consistently identified in studies across each mode of priming and across neuropsychological, behavioural, and neuroimaging studies. Qualitative differences were observed between the three subsystems of the PRS and explicit memory which indicate independent operations. This dissociation was achieved when the two types of memory were tested concurrently, rather than using different tasks, and in a way that minimises the possible confounding of implicit by explicit memory. On this basis, convincing evidence is provided to support the Multiple Memory Systems (MMS) view that memory consists of independent explicit and implicit memory systems. Further, the methods of this research programme have been shown to be a sensitive tool for studying the properties of the PRS. The priming effects elicited by each of the three subsystems were characterised by pre-semantic and non-conscious processing of the perceptual features of stimuli. The priming effect elicited by objects and aural words were also found to be longer-lasting effects ( 30 min) which provided evidence to support the long-term nature of the PRS subsystems. These characteristics converge with the conclusions other research perspectives have similarly associated with the PRS. Taken together, the properties of the three subsystems of the PRS identified in this research provide evidence to support the PRS and MMS theories. A final outcome of this research programme relates to ERP methods and memory. This research programme has observed N400 old/new effects across three modes of priming. This outcome is important as there are no reports in the literature of N400 old/new effects associated with object and auditory priming. These results provide evidence to support the functional significance of the N400 old/new effect as an index of the PRS subsystems that support priming.