Context-dependent repetition effects on recognition memory
Introduction
One key function of declarative memory, the conscious memory for facts and events (Cohen and Eichenbaum, 1993, Tulving, 1972), is to support recognition of stimuli that were previously encountered and to discriminate such stimuli from those that are novel. According to dual-process models of recognition memory, the ability to discriminate between items encountered at study and items only presented at test is supported by two independent processes, recollection and familiarity (Mecklinger, 2000, Wilding and Rugg, 1996, Yonelinas, 2002). Recollection refers to conscious retrieval of contextual details of the original study episode in which an item occurred. Thus recollection provides information both about the prior occurrence of an item and the context of that occurrence. By contrast familiarity based recognition is not accompanied by information from specific study episodes and therefore provides no means for making discriminations on the basis of contextual information. A critical assumption of most dual-process models is that recollection and familiarity are independent retrieval processes, i.e. recognition can be based solely on recollection without evoking recognition based on familiarity and vice versa (Rugg and Yonelinas, 2003, Yonelinas, 2001).
Dual-process accounts of recognition memory receive support from a number of different sources. In addition to findings from a body of behavioral and patient studies (for a comprehensive review, see Yonelinas, 2002) there is also an extensive amount of research demonstrating that event-related potentials (ERP) are sensitive to dissociate the contribution of familiarity and recollection to recognition memory (e.g., Curran, 2000, Friedman and Johnson, 2000, Rugg et al., 1998a). The ERP old/new effect is estimated from the difference between the ERPs associated with correct responses to old and new test items and comprise relatively more positive-going ERPs for old than for new test items. Based on their spatio-temporal characteristics and sensitivity to experimental manipulations this difference can be subdivided into at least two subcomponents. For the present purposes, the most important effects are an early frontal old/new effect (300–500 ms) and a later effect (400–800 ms) maximal over (left) parietal regions. Importantly, evidence suggests that the two ERP effects are dissociable on both topographic (e.g. see Mecklinger, 2000) and functional grounds (Jäger et al., 2006, Rhodes and Donaldson, 2007). Within dual-process accounts of recognition memory these ERP effects are taken to dissociate the contribution of familiarity and recollection (e.g., Curran, 2000, Friedman and Johnson, 2000, Opitz and Cornell, 2006), with the early frontal old/new effect reflecting familiarity and the late parietal old/new effect linked to recollection.
A common finding of many studies of recognition is a memory improvement when information is repeatedly presented for studying (Baddeley et al., 2001, Curran et al., 2006, van Strien et al., 2005). As an example, using a continuous recognition paradigm van Strien et al. (2005) have shown that subjects discriminated old from new stimuli faster and more accurate the more often old stimuli were repeated. It has been assumed that augmented recognition memory is achieved by strengthening the item-context bindings through repeated item presentations during the study episode. Norman and O’Reilly, 2003, O’Reilly and Norman, 2002 have developed a biologically plausible neuro-computational model1 suggesting that sparse neural coding within the hippocampus leads to distinct (pattern-separated) representations of arbitrary item-context bindings irrespective of their contextual similarity. In contrast, the medial temporal lobe cortex (MTLC) assigns similar representations to similar input using overlapping representations to code for the shared structure of events. By this means item representations become sharper over repeated exposures across different contexts. That is, the first presentation of an item weakly activates a large number of MTLC units, whereas repeated and thus familiar stimuli strongly activate a smaller number of units. At test, the presentation of a studied test probe initiates a set of processes that may be described as a comparison between the short-lived representation of the actual stimulus and the sharpened representation in the MTLC. Consequently, a scalar signal is provided that tracks the global similarity between the test probe and the studied items (Hintzman, 2001).
Repeated item presentations within the context of the same study episode will add contextual detail to the representation of this study episode, thereby strengthening the item-context bindings. This in turn will lead to a highly distinct hippocampal representation, whereas in MTLC units, a blurred representation will emerge as a large number of units is weakly activated by the item and its study context. During test, when the item is presented as a test probe the hippocampus is able to reconstruct the entire studied pattern, i.e. the item bound to its context (pattern completion, O’Reilly & Rudy, 2001), thereby enabling the retrieval of contextual information. From the dual process perspective, this binding of an item to its study context entails enhanced recollection (cf. Yonelinas & Parks, 2007). In contrast, given the blurred representation in the MTLC, i.e. the inability to sufficiently differentiate the representations of different events owing to the relatively low learning rate, only a weak familiarity signal should be elicited by items repeated within the same context. This proposed increase in recollection is corroborated by a recent study asking the participants to memorize words that were presented either once (weak words) or three times (strong words) during a study phase (Finnigan, Humphreys, Dennis, & Geffen, 2002). At test, correct old decisions after the presentation of strong words elicited a stronger left parietal old/new effect, indexing recollection, as compared to correct old decisions after the presentation of weak words.
Despite converging evidence that recognition of items repeatedly presented within the same study context is mainly based on recollection it is still unknown how repetition affects recognition memory if an item is repeated across different study contexts. This is often the case in real life where a particular event (e.g. the final goal of the soccer championship) is repeatedly encountered in different contexts, e.g. seen on the television and beeing told by a friend. Moreover, it has been speculated that this form of repetition leads to decontextualized factual knowledge about the world (Craik, 2006, Eichenbaum, 2006). According to the model proposed by Norman and O’Reilly (2003) the repetition of items across different contexts gives rise to a small overlap of contextual features resulting in separate but weak hippocampal representations. However, the sharpening mechanism in the MTLC operates faster and much more efficiently as compared to a situation when items are repeated within the same context in that only the item without contextual features is represented by MTLC units. At test, the test probe is compared against this sharpened representation of the item alone in the MTLC and a larger familiarity signal is provided, whereas the recollection of studied contexts is less likely because of the use of a shared structure to represent similar events preventing the binding of an item to an arbitrary context. While the MTLC model cannot support recollection of details from specific events owing to its relatively low learning rate it well supports familiarity judgments based on the sharpness of representations in MTLC. Consistent with this neuro-computational model I argue that item representations become sharper over repeated exposures across different contexts, thereby supporting recognition based on familiarity. In contrast, repetition within the same context should foster item-context bindings which lead to recollection.
The present experiment explored these hypotheses employing visual objects that were studied either once or three times. Crucially, for half of the repeated objects the study context remained the same (within context repetition) while for the other half the contextual features changed across repetitions (across context repetition). Beside the independent remember/know-procedure the putative ERP correlates of familiarity and recollection were used. Thus, an early frontal old/new effect and a late parietal old/new effect should be observed during the retrieval of items repeated across and within context, respectively.
Section snippets
Subjects
A total of 30 students from Saarland University participated in this study and were paid for their participation. The data from two participants were discarded due to malfunction of the recording equipment. The data from further two participants were excluded from all analyses because they exhibited excessive EOG artifacts. Of the remaining 26 participants (aged 21–29 years, 14 male) all were right handed as assessed by the Edinburgh Handedness Inventory (Oldfield, 1971).
Materials
Stimuli consisted of
Behavioral data
Table 1 shows the mean reaction times of correct responses and the probability of an old judgment to old and new items and the proportion of subsequent R- and K-responses. The ANOVA performed on these measures revealed significantly more old responses for old as compared to new items (F1,25 = 566.3, p < .001). In addition, it was revealed that repeated items were better recognized than items presented once during study (F1,25 = 95.1, p < .001) and that the study context also had an effect on
Discussion
Behavioral and electrophysiological measures were used to examine the subprocesses mediating recognition memory for items repeated within the same context or across different contexts. Consistent with the hypothesis the behavioral results indicated more accurate and faster recognition of repeated items as compared to items presented once during study. This is in agreement with a number of previous results demonstrating superior memory performance after repeated presentation during study (
Acknowledgments
I am grateful to Lea Meyer and Kristina Lemke for their valuable help during data acquisition. This work was supported by Grant FOR448 of the German Research Foundation.
References (63)
- et al.
Interleaving brain systems for episodic and recognition memory
Trends in Cognitive Sciences
(2006) Effects of attention and confidence on the hypothesized erp correlates of recollection and familiarity
Neuropsychologia
(2004)- et al.
The fn400 indexes familiarity-based recognition of faces
Neuroimage
(2007) Item, context and relational episodic encoding in humans
Current Opinion in Neurobiology
(2006)- et al.
Erp ‘old/new’ effects: Memory strength and decisional factor(s)
Neuropsychologia
(2002) - et al.
A new method for off-line removal of ocular artifact
Electroencephalography and Clinical Neurophysiology
(1983) - et al.
Intra- and inter-item associations doubly dissociate the electrophysiological correlates of familiarity and recollection
Neuron
(2006) - et al.
A spatio-temporal analysis of recognition-related event-related brain potentials
International Journal of Psychophysiology
(1998) The assessment and analysis of handedness: The edinburgh inventory
Neuropsychologia
(1971)- et al.
Hippocampal and neocortical contributions to memory: Advances in the complementary learning systems approach
Trends in Cognitive Sciences
(2002)
Functional aspects of recollective experience in face recognition
Consciousness and Cognition
Spherical splines for scalp potential and current density mapping
Electroencephalography and Clinical Neurophysiology
Spacing and repetition effects in human memory: Application of the sam model
Cognitive Science: A Multidisciplinary Journal
Dissociable correlates of recollection and familiarity within the medial temporal lobes
Neuropsychology
Electrophysiological evidence for the influence of unitization on the processes engaged during episodic retrieval: Enhancing familiarity based remembering
Neuropsychologia
Event-related potentials and recognition memory
Trends in Cognitive Sciences
An electrophysiological comparison of two indices of recollection
Journal of Memory and Language
Human recognition memory: A cognitive neuroscience perspective
Trends in Cognitive Sciences
Changes in brain electrical activity during extended continuous word recognition
Neuroimage
Rote rehearsal and spacing effects in the free recall of pure and mixed lists
Journal of Memory and Language
The relationship between electrophysiological correlates of recollection and amount of information retrieved
Brain Research
Electrophysiological dissociation of the neural correlates of recollection and familiarity
Brain Research
The nature of recollection and familiarity: A review of 30 years of research
Journal of Memory and Language
The relation between remembering and knowing as bases for recognition: Effects of size congruency
Journal of Memory and Language
The influence of criterion shifts on electrophysiological correlates of recognition memory
Journal of Cognitive Neuroscience
Preserved recognition in a case of developmental amnesia: Implications for the acquisition of semantic memory?
Journal of Cognitive Neuroscience
Memory, amnesia, an the hippocampal system
The role of recollection and familiarity in the context variability mirror effect
Memory and Cognition
Remembering items and their contexts: Effects of aging and divided attention
Brain potentials of recollection and familiarity
Memory and Cognition
Event-related potential explorations of dual processes in recognition memory
Cited by (18)
Sometimes we have to intentionally focus on the details: Incidental encoding and perceptual change decrease recognition memory performance and the ERP correlate of recollection
2015, Brain and CognitionCitation Excerpt :Both effects have been reported in earlier studies: Larger parietal old/new effects have been found when associative or contextual information is retrieved (Curran & Doyle, 2011; Donaldson & Rugg, 1998; Van Petten, Luka, Rubin, & Ryan, 2002; Wilding, 2000; Wilding & Rugg, 1997; for a similar result, see Nyhus & Curran, 2012). Additionally, an increase in the magnitude of this correlate has been associated with higher memory performance (e.g., Czernochowski et al., 2009; Evans, Herron, & Wilding, 2012; Friedman, de Chastelaine, Nessler, & Malcolm, 2010; Opitz, 2010; Van Petten et al., 2002; Van Strien, Hagenbeek, Stam, Rombouts, & Barkhof, 2005). In sum, we observed modulations of the ERP component of recollection that can be explained by higher memory performance and the additional retrieval of perceptual details.
Associative recognition processes are modulated by the semantic unitizability of memoranda
2014, Brain and CognitionCitation Excerpt :Aside from the expected LPC modulation, elicited by associative recognition judgments (i.e., intact > recombined), an unexpected pattern of intact = new > recombined for unrelated pairs emerged. This pattern is inconsistent with previous reports of decreased positivity for new materials (e.g., Bader et al., 2010; Greve et al., 2007; Kriukova et al., 2013; Mollison & Curran, 2012; Opitz, 2010; Rhodes & Donaldson, 2007, 2008); the reason for this discrepancy is unclear. A comparable finding is noted by Addante, Ranganath, and Yonelinas (2012), who describe an ERP in this time window elicited by correct but low-confidence item memory accompanied by correct source judgment that is less positive than the ERP elicited by correct rejection of foils.
Is a novel conceptual unit more than the sum of its parts?: FMRI evidence from an associative recognition memory study
2014, NeuropsychologiaCitation Excerpt :Note that some ERP studies did not find evidence for reduced recollection for unitized associations when study conditions promoted the contribution of recollection. For example in the Wiegand et al. (2010) study, all word pairs were studied twice, a condition which is known to increase recollection (Jacoby, Jones, & Dolan, 1998; Opitz, 2010a) while in the Rhodes and Donaldson studies (2007, 2008) multiple short study-test-cycles were used. The exact boundary conditions of when unitization attenuates recollection still have to be determined (see below).
Temporal texture of associative encoding modulates recall processes
2014, Brain and CognitionCitation Excerpt :This recall success effect, also observed in our prior study, invites comparison with a similar component found in recognition studies, the FN400 or the early midfrontal old–new effect, analogously characterized by a more negative deflection at anterior electrodes for probes judged to be new (e.g., Mecklinger, 2000; Rugg & Curran, 2007). As we have noted in our earlier study (Tibon & Levy, 2013), while in studies of retrieval of verbal materials retrieval success effects of this type emerge at a latency of 300–500 ms post cue, when picture cues are employed as memoranda the effect begins earlier (e.g., Ally & Budson, 2007; Ecker, Zimmer, Groh-Bordin, & Mecklinger, 2007; Herzmann, Jin, Cordes, & Curran, 2012; Iidaka, Matsumoto, Nogawa, Yamamoto, & Sadato, 2006; Jäger, Mecklinger, & Kipp, 2006; Opitz, 2010). Furthermore, in studies of picture recognition, not only the onset latency but the duration of the FN400 component is reported to be the similar to the cued recall effect in the present study (e.g., Iidaka et al., 2006, and Opitz, 2010; in Yu & Rugg, 2010, the effect was stable over the 200–500 ms range).
The effects of variable encoding contexts on item and source recognition
2023, Memory and Cognition