Perceptual priming enhances the creation of new episodic memories
Introduction
One of the most fundamental advances in the study of human memory in recent years has been the distinction between implicit and explicit memory phenomena (Graf and Schacter, 1985, Schacter, 1987). Implicit memory has been defined as the expression of past experiences occurring beyond the boundaries of consciousness and without any intentional recollection. Priming is one of the most well-known implicit memory phenomena and refers to a change in the speed or accuracy with which a stimulus is processed, following prior experience of the same or related stimulus. Different kinds of priming have been identified, such as perceptual priming, which is based on the stimuli’s physical properties (Tulving & Schacter, 1990). By contrast, explicit memory entails the conscious recollection of previously studied information. The Remember/Know procedure (Gardiner, 1988, Tulving, 1985) is intended to gauge states of consciousness associated with memory retrieval. As such, it constitutes a sound approach to autonoetic consciousness and noetic consciousness respectively (for a review, see Gardiner, 2001, Tulving, 2001).
Numerous studies of behavioral, neuropsychological and neuroimaging dissociations (for a review, see Gardiner, 2001, Roediger and McDermott, 1993, Schacter and Buckner, 1998, Schacter et al., 2004, Wheeler and Buckner, 2004, Yonelinas, 2002) have demonstrated that these three distinct memory expressions (perceptual priming, knowing and remembering) rely on three distinct memory systems, namely the perceptual (Tulving & Schacter, 1990; see also Schacter, 1990, Schacter, 1992, Schacter, 1994), semantic and episodic memory systems (Tulving, 1995, Tulving, 2001, Tulving, 2002, Tulving and Markowitsch, 1998). While these memory systems overlap to a certain extent, they differ in their processing characteristics, representations and neural substrates (Schacter & Tulving, 1994). However, rather than being the expression of two distinct processes or memory systems, it has been suggested that Remember/Know behavioral dissociations may actually reflect variations in a response criterion/decision process along the continuum of a single memory trace (one-dimensional model) that can be modeled on the basis of signal detection (SD) theory (e.g., Donaldson, 1996, Hirshman and Master, 1997). Even if some data on the receiver operating characteristics (ROCs) have questioned this model (e.g., Rotello, Macmillan, & Reeder, 2004), it may nonetheless be worthwhile assessing this one-dimensional SD account if only to discard a potential explanation based on decision processes (Hirshman et al., 2002, Hirshman and Master, 1997).
Tulving, 1995, Tulving, 2001 has given us a powerful model to explain the functional relations between memory systems, which he calls the serial-parallel-independent (SPI) model. This states that although the encoding of information occurs serially, moving from perceptual to semantic and then to episodic memory, distinct information is stored in each system and information can be retrieved from one system independently of the other systems. However, little is known about the role of perceptual priming and, consequently, about the contribution of perceptual representations to the creation of episodic memories. These considerations lead us to the unsolved question of whether episodic memory creation relies on perceptual representations stored in perceptual memory—representations which also subtend perceptual priming.
On the one hand, there is substantial evidence to show that perceptual priming and explicit memory rely on distinct neural substrates and representations. Whereas amnesic patients demonstrate intact perceptual priming despite explicit memory impairment (for a review, see Schacter & Buckner, 1998), Gabrieli, Fleischman, Keane, Reminger, and Morrell (1995) reported the case of a patient with a right occipital lobe lesion who presented the reverse dissociation. In an fMRI study using a priming paradigm adapted from the word-stem completion task, Schott et al. (2006) found that activity while encoding words in perceptual memory occurred in regions involved in perception and identification, whereas activity during explicit memory encoding occurred in the bilateral medial temporal lobe and left prefrontal cortex, but not in those regions involved in perceptual priming.
On the other hand, there is growing evidence to suggest that implicit and explicit memory share common resources and representations, and interact at encoding. Those regions involved in processing perceptual information that are believed to support perceptual priming are also thought to be involved in the formation and recollection of memories (e.g., Wheeler et al., 2000, Woodruff et al., 2005; for a review, see Buckner & Wheeler, 2001; see also Johnson and Chalfonte, 1994, Moscovitch, 1994). Using fMRI, Turk-Browne, Yi, and Chun (2006) recently found that subsequently remembered scenes were associated with greater neural and behavioral priming during encoding than subsequently forgotten ones (see also Kirchhoff, Wagner, Maril, & Stern, 2000). These authors posit that implicit and explicit memory may share factors and representations at encoding but rely on different retrieval mechanisms. However, in their study, explicit memory was assessed in the form of a high-/low-confidence judgment which does not necessarily imply the presence of episodic memory (Gardiner, 2001). Moreover, as all the items were repeated twice, it was impossible to assess a single episodic memory trace.
The present study sought to assess the impact of perceptual priming on the creation of new episodic memories. The idea was to prompt the encoding of unique events in episodic memory that had been perceptually primed to a greater or lesser degree during an incidental encoding phase. To achieve this, the experiment was divided into three distinct phases (see Fig. 1): (1) an initial study phase consisting of a perceptual phonological processing, during which verbally presented target words were heard either once or three times; (2) a priming test phase occurring 24 h later and consisting of a perceptual auditory lexical decision task where new control words (“no-priming” condition) and target words heard either once (“low-priming” condition) or three times (“high-priming” condition), were presented against the background of two different sounds; and (3) a subsequent Remember/Know/Guess (R/K/G) recognition task featuring words belonging to the three distinct priming conditions (“no-priming, “low-priming” and “high-priming”). The role of the sound context was twofold: (1) to enhance the likelihood of observing a perceptual priming effect by degrading the subjects’ perception of the words (a classic feature of auditory perceptual priming tasks, e.g., Schacter & Church, 1992); and (2) to provide a context (hence two different sound backgrounds) for encoding the words as unique events to be recalled in a recognition task. The second phase therefore corresponded to the incidental encoding phase of a subsequent recognition task.
Section snippets
Subjects
Thirty-six subjects were recruited for this experiment, aged between 18 and 35 years (M = 23.3 and SD = 5.3). There were 24 women and 12 men. All were right-handed, free from hearing disorders and had at least 12 years’ schooling.
Material
The material was recorded by a female voice and consisted of 374 words taken from the LEXIQUE French database (New, Pallier, Brysbaert, & Ferrand, 2004; www.lexique.org), 86 pseudo-words and 20 consonantal sounds used during the first, phonological study phase (see Section
Priming and recognition results
The RT analysis (see Fig. 2) revealed a significant main effect of the priming condition [F(2, 70) = 20.2, p < 10−4]. Planned comparisons revealed (1) that words in the “low-priming” condition were processed faster than in the “no-priming” condition [F(1, 35) = 15.18, p < .001] and (2) that words in the “high-priming” condition were processed faster than in the “low-priming” condition [F(1, 35) = 6.6, p < .05].
The analysis of corrected recognition proportions (Hits–FAs) showed a significant main effect of the
Discussion
The primary aim of this study was to gain a better understanding of the impact of perceptual priming on the creation of new episodic memories. Our results revealed a significant priming effect at test, as assessed by RTs, and showed that manipulating the number of repetitions during the first study phase increased the magnitude of auditory perceptual priming, even 24 h later. According to the SPI model of memory systems (Tulving, 1995), this result suggests that the encoding of perceptual
Acknowledgments
The authors thank Professor Endel Tulving and William P. Banks for their helpful comments regarding an earlier version of this manuscript. The authors also thank Elizabeth Portier for reviewing the English style and Berengère Guillery-Girard and Mickael Laisney for their very appreciate help.
References (47)
- et al.
Insights from semantic dementia on the relationship between episodic and semantic memory
Neuropsychologia
(2000) - et al.
Repetition and the brain: neural models of stimulus-specific effects
Trends in Cognitive Sciences
(2006) - et al.
Memory consolidation, retrograde amnesia and the hippocampal complex
Current Opinion in Neurobiology
(1997) - et al.
Priming and the brain
Neuron
(1998) - et al.
Linking implicit and explicit memory: Common encoding factors and shared representations
Neuron
(2006) - et al.
Functional anatomic correlates of remembering and knowing
Neuroimage
(2004) - et al.
Content-specificity of the neural correlates of recollection
Neuropsychologia
(2005) The nature of recollection and familiarity: A review of 30 years of research
Journal of Memory & Language
(2002)Recognition and source memory as multivariate decision processes
Psychological Science
(2000)- et al.
The cognitive neuroscience of remembering
Nature Review Neuroscience
(2001)