Feature binding in perceptual priming and in episodic object recognition: evidence from event-related brain potentials

It is argued that explicit remembering is based on so-called episodic tokens binding together all perceptual features of a visual object. In episodic recognition, these features are collectively reactivated. In support of this view, it has been shown that changing sensory features of a stimulus from study to test decreases subject's performance in an episodic recognition task, even though the changed features are irrelevant for the recognition judgment. On the other hand, repetition priming is unaffected by such manipulations of perceptual specificity. Implicit memory performance is therefore thought to depend on structural representations, so-called types, comprising only invariant perceptual features, but no exemplar-specific details. Event-related potentials (ERPs) in our study revealed electrophysiological evidence for the differential involvement of these perceptual memory traces in explicit and implicit memory tasks. Participants attended either a living-nonliving task or an episodic recognition task with visually presented objects. During test both groups of participants processed new objects and old objects, which were repeated either identically or in a mirror-reversed version. In the implicit task ERPs showed an occipitoparietal repetition effect, which was the same for identically repeated items and mirror reversals. In contrast, in the explicit task an early mid-frontal old/new effect for identical but not for mirror-reversed old objects was observed indicating involuntary access to perceptual information during episodic retrieval. A later portion of the old/new effect solely differentiated both types of old items from new ones.     

Repetition and brain potentials when recognizing natural scenes: task and emotion differences

Repetition has long been known to facilitate memory performance, but its effects on event-related potentials (ERPs), measured as an index of recognition memory, are less well characterized. In Experiment 1, effects of both massed and distributed repetition on old–new ERPs were assessed during an immediate recognition test that followed incidental encoding of natural scenes that also varied in emotionality. Distributed repetition at encoding enhanced both memory performance and the amplitude of an old–new ERP difference over centro-parietal sensors. To assess whether these repetition effects reflect encoding or retrieval differences, the recognition task was replaced with passive viewing of old and new pictures in Experiment 2. In the absence of an explicit recognition task, ERPs were completely unaffected by repetition at encoding, and only emotional pictures prompted a modestly enhanced old–new difference. Taken together, the data suggest that repetition facilitates retrieval processes and that, in the absence of an explicit recognition task, differences in old–new ERPs are only apparent for affective cues.     

A comparison of implicit memory tasks in Korsakoff and alcoholic patients

Korsakoff patients generally perform at a normal level on implicit memory tasks in contrast to explicit memory tasks. While this difference is sometimes explained in terms of different memory systems underlying the tasks, the different roles of perceptual and conceptual/semantic processes in these tasks have also been emphasized: explicit tasks require mainly conceptual/semantic processes and implicit tasks are based principally on perceptual processes. However, it has been suggested recently that conceptual/semantic processes may also be involved in some implicit memory tasks (e.g. a Free Association task). Therefore, the performance of Korsakoff and alcoholic patients is here compared in three implicit memory tasks (Stem Completion, Word Identification and Free Association) and one explicit memory task (Cued Recall), allowing us to disentangle what really matters: the nature of the task (implicit or explicit), or the underlying processes (perceptual or conceptual/semantic). The results show only semantic priming in the Free Association and Cued Recall tasks of the alcoholic patients, suggesting that Korsakoff patients have problems particularly with conceptually-driven processing. The implicit or explicit nature of the memory task is not critical.     

Preserved priming but insensitivity to perceptual fluency on recognition judgments in Alzheimer's disease

Background: Previous research on Alzheimer's disease (AD) has not yielded a consensus regarding the preservation of automatic memory processes, although there is a consensus that conscious recollection processes are impaired in AD. Methods: In the present study, we examined perceptual specificity effects (PSEs) in word recognition judgments (explicit memory task; Experiment 1) and word fragment completion (implicit memory task; Experiment 2) performed by individuals with mild AD and elderly adults without dementia (controls). Results: In recognition judgments, control subjects, but not individuals with AD, demonstrated PSEs (Experiment 1). In contrast, neither group showed PSEs on word fragment completion and their priming magnitudes were comparable (Experiment 2). Conclusions: The findings suggest that perceptually automatic processes in explicit memory judgments and implicit memory processes are different and that the former are specifically impaired in AD.     

ERP recordings during a picture fragment completion task: effects of memory instructions.

Sequential implicit (picture fragment completion) and explicit (free recall and recognition) memory tasks were performed by two groups of young adults. Subjects assigned to the intentional condition were asked to memorize the stimuli presented during the picture fragment completion task, whereas subjects assigned to the incidental condition were not so instructed. In the picture fragment completion task, all subjects showed savings (priming), i.e., their identification thresholds for previously seen fragments were lower than those for fragments not previously seen, and no between-group performance difference was found. Free recall performance was better for the intentional than the incidental group, but recognition performance did not differ between the groups. Despite similar between-group performance, event-related brain potentials (ERPs) recorded during the picture fragment completion and recognition tasks revealed between-group differences in information processing. The ERP data suggest that similar behavioral performance during memory tasks may have been associated with distinct neuronal mechanisms subserving priming and recollection.     

Perceptual fluency, semantic familiarity and recognition-related familiarity: an electrophysiological exploration

Scalp recorded event-related potentials (ERPs) were used to examine the neuronal activity associated with perceptual fluency, semantic familiarity and recognition-related familiarity. We assume that ERP differences between first and second presentations of non-famous faces in an implicit memory condition reflect perceptual fluency, ERP differences between first presentations of famous and non-famous faces reflect semantic familiarity (i.e., familiarity arising from semantic memory retrieval), and early ERP differences between first and second presentations of non-famous and famous faces in an explicit recognition memory task reflect recognition-related familiarity. Semantic familiarity elicited a broadly distributed effect between 200 and 300 ms after stimulus onset, possibly representing the activation of face recognition units. Between 300 and 450 ms, frontal effects were observed for semantic familiarity and recognition-related familiarity, while perceptual fluency was associated with a centro-parietally focused effect. Thus, familiarity arising from the retrieval of semantic information and recognition-related familiarity depend at least partly on the same neuronal circuits, while these are dissociable from those mediating perceptual fluency.     

Memory impairment in schizophrenia: a study using event-related potentials in implicit and explicit tasks.

Although memory impairment is recognized as a major fact of schizophrenia, only a few studies have investigated memory impairments with specifically designed event-related potential (ERP) protocols. In this study, ERPs were recorded from 15 schizophrenia patients and 15 matched control subjects during implicit and explicit memory tasks for unfamiliar faces. The results showed that patients have a reduced modulation of an N400-like component in both the implicit and explicit tasks that suggests a deficient integration of incoming information with personal knowledge. Patients also displayed an enhanced frontally distributed activity in the explicit task that may represent an impairment in the integration of intrinsic contextual information, a disturbance in the ability to inhibit proactive interference or a combination of both processes. Finally, the modulation of the late positive component did not differ from that in control subjects in both implicit and explicit tasks, suggesting that the impairment in mnemonic binding processes suggested in schizophrenia is more qualitative, i.e. incomplete or inappropriate, due to the anomalies in antecedent processes. The correlations observed between impairments of ERP modulation and symptoms further support these interpretations.     

Emotional memory is perceptual

In two experiments it was investigated which aspects of memory are influenced by emotion. Using a framework proposed by Roediger (American Psychologist 45 (1990) 1043-1056), two dimensions relevant for memory were distinguished the implicit-explicit distinction, and the perceptual versus conceptual distinction. In week 1, subjects viewed a series of slides accompanied with a spoken story in either of the two versions, a neutral version, or a version with an emotional mid-phase. In week 2, memory performance for the slides and story was assessed unexpectedly. A free recall test revealed superior memory in the emotional condition for the story's mid-phase stimuli as compared to the neutral condition, replicating earlier findings. Furthermore, memory performance was assessed using tests that systematically assessed all combinations of implicit versus explicit and perceptual versus conceptual memory. Subjects who had listened to the emotional story had superior perceptual memory, on both implicit and explicit level, compared to those who had listened to the neutral story. Conceptual memory was not superior in the emotional condition. The results suggest that emotion specifically promotes perceptual memory, probably by better encoding of perceptual aspects of emotional experiences. This might be related to the prominent position of perceptual memories in traumatic memory, manifest in intrusions, nightmares and reliving experiences.     

Repetition suppression of induced gamma band responses is eliminated by task switching

The formation of cortical object representations requires the activation of cell assemblies, correlated by induced oscillatory bursts of activity > 20 Hz (induced gamma band responses; iGBRs). One marker of the functional dynamics within such cell assemblies is the suppression of iGBRs elicited by repeated stimuli. This effect is commonly interpreted as a signature of 'sharpening' processes within cell-assemblies, which are behaviourally mirrored in repetition priming effects. The present study investigates whether the sharpening of primed objects is an automatic consequence of repeated stimulus processing, or whether it depends on task demands. Participants performed either a 'living/non-living' or a 'bigger/smaller than a shoebox' classification on repeated pictures of everyday objects. We contrasted repetition-related iGBR effects after the same task was used for initial and repeated presentations (no-switch condition) with repetitions after a task-switch occurred (switch condition). Furthermore, we complemented iGBR analysis by examining other brain responses known to be modulated by repetition-related memory processes (evoked gamma oscillations and event-related potentials; ERPs). The results obtained for the 'no-switch' condition replicated previous findings of repetition suppression of iGBRs at 200-300 ms after stimulus onset. Source modelling showed that this effect was distributed over widespread cortical areas. By contrast, after a task-switch no iGBR suppression was found. We concluded that iGBRs reflect the sharpening of a cell assembly only within the same task. After a task switch the complete object representation is reactivated. The ERP (220-380 ms) revealed suppression effects independent of task demands in bilateral posterior areas and might indicate correlates of repetition priming in perceptual structures.     

Cerebral evoked potentials and conscious and unconscious recognition of faces: application to the study of prosopagnosia

Twelve normal subjects and a prosopagnosic patient were tested in a classification task of a random display of well-known among unknown faces. Each face was presented several times. Event-related potentials (ERP) and reaction time (RT) were studied as a function of face repetition and familiarity. For normal subjects, the greater the repetition level, the more positive ERPs were on both hemispheres: between 250 and 600 msec. Moreover, the familiarity of faces modified ERPs between 350 and 600 msec. In contrast for the patient, the greater the repetition, the more negative the ERPs were. This "negative effect" was maximum on right parieto-temporal leads and was longer for unrecognized well-known than for unknown faces. These results support a differential processing of faces as a function of their memory representations for both normal subjects and patients. They further demonstrate the existence of covert face recognition processes in prosopagnosia.     

Human declarative memory formation: segregating rhinal and hippocampal contributions

The medial temporal lobe (MTL) is the core structure of the declarative memory system, but which specific operation is performed by anatomically defined MTL substructures? One hypothesis proposes that the hippocampus carries out an exclusively mnemonic operation during declarative memory formation that is insensitive to content, whereas the rhinal cortex carries out an operation supporting memory formation indirectly. To explore the interaction between a salient item feature and memory formation, we contrasted neural correlates of memory formation of high‐ and low‐frequency words. Event‐related potentials (ERPs) were recorded via depth electrodes from within the MTL in nine epilepsy patients while they memorized single words. To assess memory formation, ERPs to words subsequently recalled in a free recall test were contrasted with ERPs to forgotten words. More high‐ than low‐frequency words were remembered. High‐frequency words led to distinct ERP subsequent memory effects in rhinal cortex and hippocampus. Low‐frequency words, however, were only associated with the hippocampal ERP effect. The anatomically restricted interaction between word frequency and memory formation might indicate a semantically affected operation in the parahippocampal region supporting memory formation indirectly. By contrast, the missing interaction in hippocampal recordings might suggest a direct correlate of declarative memory formation that is insensitive to item properties. Hippocampus 2002;12:514–519. © 2002 Wiley‐Liss, Inc.     

The effect of encoding manipulation on word-stem cued recall: an event-related potential study

The purpose of the present study was to find out whether the neural correlates of explicit retrieval from episodic memory would vary according to conditions at encoding when the words were presented in separate study/test blocks. Event-related potentials (ERPs) were recorded while participants performed a word-stem cued-recall task. Deeply (semantically) studied words were associated with higher levels of recall and faster response times than shallowly (lexically) studied words. Robust ERP old/new effects were observed for each encoding condition. They varied in magnitude, being largest in the semantic condition. As expected, scalp distributions also differed: for deeply studied words, the old/new effect resembled that found in previous ERP studies of word-stem cued-recall tasks (parietal and right frontal effects, between 400-800 and 800-1100 ms post-stimulus), whereas for shallowly studied words, the parietal old/new effect was absent in the latter latency window. These results can be interpreted as reflecting access to different kinds of memory representation depending on the nature of the processing engaged during encoding. Furthermore, differences in the ERPs elicited by new items indicate that subjects adopted different processing strategies in the test blocks following each encoding condition.     

Event-related potential studies of associative recognition and recall: electrophysiological evidence for context dependent retrieval processes

To investigate the neural correlates of episodic recollection the ERP correlates of memory for new associations (recently studied novel word pairs) were investigated using two tasks, associative recognition and associative recall. For the recognition task subjects discriminated old from new word pairs and, for pairs judged old, reported whether the pairs were intact or recombined (compared to at study). For the recall task, subjects discriminated old from new words and, for each word judged old, reported its study associate. ERPs were recorded at test from 25 scalp electrodes, with a 1944-ms recording epoch. In Experiment 1, the tasks were randomly interleaved. Consistent with previous findings, relative to the ERPs for correctly classified new items, the ERP correlates of successful associative recognition consisted of a sustained left parietal positivity, and two frontal positivities, one early and bilateral, the other occurring later and showing a right-sided maximum. In contrast to previous findings, successful associative recall elicited similar effects to those found for recognition. Topographic analyses revealed that the distribution of these retrieval-related ERP effects were similar across the two tasks, suggesting that the recognition and recall of associative information gives rise to activity in overlapping, if not the same, neural populations. In Experiment 2 the tasks were blocked. In contrast to the findings of Experiment 1, successful associative recall elicited left parietal and late onsetting right frontal positivities, in the absence of the early bilateral frontal positivity. This finding suggests that frontally-distributed memory-related ERP effects are both neurally and functionally dissociable. Specifically, we argue that the functional significance of the early frontally distributed ERP effect cannot be accounted for by the 'post-retrieval processing' hypothesis that is taken to account for the late right frontal effect, suggesting that episodic recollection itself is neither neurally nor functionally homogenous.     

Modality-specific effects of immediate word repetition: electrophysiological evidence

Event-related potentials (ERPs) elicited by visually presented words were employed to investigate whether the neural correlates of repetition within- and across-modality differ when repetition is immediate, and the influence of explicit memory maximal. Relative to the ERPs elicited by first presentations, ERPs elicited by immediate, within-modality repetitions began to differ from approximately 200 ms post-stimulus onset. ERP repetition effects elicited by across-modality repetition did not onset until approximately 150 ms later. Within- and across-modality repetition effects were also dissociable neuroanatomically, exhibiting different scalp distributions. The findings support the proposal that the modality-sensitive component of visual word repetition effects operates at an early, pre-semantic processing stage.     

Perceptual priming versus explicit memory: dissociable neural correlates at encoding

We addressed the hypothesis that perceptual priming and explicit memory have distinct neural correlates at encoding. Event-related potentials (ERPs) were recorded while participants studied visually presented words at deep versus shallow levels of processing (LOPs). The ERPs were sorted by whether or not participants later used studied words as completions to three-letter word stems in an intentional memory test, and by whether or not they indicated that these completions were remembered from the study list. Study trials from which words were later used and not remembered (primed trials) and study trials from which words were later used and remembered (remembered trials) were compared to study trials from which words were later not used (forgotten trials), in order to measure the ERP difference associated with later memory (DM effect). Primed trials involved an early (200-450 msec) centroparietal negative-going DM effect. Remembered trials involved a late (900-1200 msec) right frontal, positive-going DM effect regardless of LOP, as well as an earlier (600-800 msec) central, positive-going DM effect during shallow study processing only. All three DM effects differed topographically, and, in terms of their onset or duration, from the extended (600-1200 msec) fronto-central, positive-going shift for deep compared with shallow study processing. The results provide the first clear evidence that perceptual priming and explicit memory have distinct neural correlates at encoding, consistent with Tulving and Schacter's (1990) distinction between brain systems concerned with perceptual representation versus semantic and episodic memory. They also shed additional light on encoding processes associated with later explicit memory, by suggesting that brain processes influenced by LOP set the stage for other, at least partially separable, brain processes that are more directly related to encoding success.     

Implicit, explicit, and semantic memory functions in Alzheimer's disease and Huntington's disease.

A priming task involving a word-stem completion paradigm was administered to patients with Alzheimer's disease (AD), patients with Huntington's disease (HD), and normal control subjects. The task was done under conditions of both implicit and explicit recall. Explicit and implicit recall were positively correlated in all three groups. After controlling for explicit recall ability through ANCOVA, AD patients were found to be normally susceptible to the effects of priming on implicit recall. HD patients, however, exhibited significantly increased susceptibility to priming, suggesting that they may have carried out the implicit task in a manner different from that of normals and AD patients. In a second experiment, AD patients were found to supply words of significantly lower association strength than the other two groups in a "free association" task using words from a published list of word association norms. This apparent degradation of semantic memory was found to be strongly correlated with explicit recall performance, suggesting that explicit, implicit, and semantic memory functions decline in parallel in AD. Results are discussed with respect to the difficulties inherent in attempts to demonstrate selective impairments of conceptually distinct forms of memory.     

Selective memory and memory deficits in depressed inpatients

We investigated memory impairment and mood-congruent memory bias in depression, using an explicit memory test and an implicit one. Thirty-six severely depressed inpatients that fulfilled DSM-IV criteria for major depressive disorder and 36 healthy controls matched for sex, age, and educational level participated in the study. Explicit memory was assessed with a free recall task and implicit memory with an anagram solution task. Results showed that depressed and controls differed in explicit memory performance, depending on the amount of cognitive distraction between incidental learning and testing. Implicit memory was not affected. In addition, severely depressed patients showed a mood-congruent memory bias in implicit memory but not in explicit memory. The complex pattern of results is discussed with regard to relevant theories of depression.     

Implicit and explicit memory in young, old, and demented adults.

Previous study of scopolamine and memory (Grober et al., 1989) showed that young adults given moderate or high doses of scopolamine maintained maximum cued recall in spite of a dose-dependent decrement in free recall when memory was assessed by cued selective reminding (CSR), a procedure which circumvents inattention and induces semantic processing. Intact recall by CSR indicates either that scopolamine impairs memory indirectly through effects on attention and information processing or that it impairs explicit memory but not implicit memory. In the present study which was done to determine if CSR reflects explicit or implicit memory, a free association test was used to estimate implicit memory after CSR was administered; explicit memory was estimated with a final trial of cued recall. Data from young, nondemented, and demented adults indicate that CSR reflects explicit memory supporting the interpretation of the previous study that scopolamine does not produce direct impairment of explicit memory.     

Implicit and explicit memory in patients with obsessive-compulsive disorder: an event-related potential study

The implicit and explicit memory in patients with obsessive-compulsive disorder (OCD) was investigated using the event-related potential (ERP). For the assessment of implicit memory, a lexical decision task was administered. Among a total of 320 words and 140 non-words, 200 words were repeated, while the remaining 120 words and the 140 non-words were not repeated. For explicit memory, a continuous recognition task was administered, in which 280 words were repeated and 100 were not repeated. On the recognition task, both the controls and OCD patients showed more positivity to the old words than to the new words during the 200-600 ms period post-stimulus. Both groups showed faster response time to the old words than to the new words. On the lexical decision task, the controls showed the old/new effect during the 200-500 ms period post-stimulus, while the OCD patients did not. However, OCD patient showed faster response time to the old words than to the new words, although OCD patients showed prolonged response times to the old words compared to the controls. These results indicate that OCD patients have preserved explicit and implicit memory. The absence of old/new effect on ERP in OCD patients was discussed in terms of dysfunction of frontostriatal system, which plays an important role in both OCD and implicit memory.