Memory & Cognition最新文献

Crossing a spatial boundary, such as a doorway, often signals the ending of one episode and the beginning of another, segmenting ongoing experience into events. When conducted in real environments, this 'doorway effect' differentially affects memory for objects encountered within and between events. The evidence for this occurring in immersive virtual reality (VR) environments is mixed. The present study investigated the 'doorway effect' in a VR environment and also examined how event memory is affected when segmentation processes are disrupted by interference tasks. Ninety participants explored a five-room VR building containing interactable objects. During exploration, some participants were presented with virtual distractor tasks that required either visuospatial working memory or simple rapid-reaction responses. These tasks occurred either at doorways or in the middle of rooms. Later recall for the temporal order and contextual location of objects was examined and compared with controls, who explored the building without any distractions. The results showed that a doorway effect was evident in the control, no-distractor condition. The visuospatial distractor task impaired memory for objects, but only when it occurred in the middle of a room, possibly because it separated the representation of the encapsulating room into different events. When this task was performed in the doorway, however, its effects overlapped with, and did not add to, the spatial boundary effect. Together these findings show that both spatial boundaries and spatial distractor tasks can segment memory for experience in an immersive VR environment.

Navigating changes is fundamental to everyday life and requires updating existing memories to incorporate new details. This study examined mechanisms underlying how reinstating an earlier event's context during a later event influences memory for both events. Two theories predict opposite outcomes. Interference theory holds that reinstating context from an existing memory while experiencing a new, overlapping event produces response competition and impairs memory for both. In contrast, integration theory predicts that context reinstatement cues retrieval of earlier memories, enabling associative encoding of past and present events that enhances memory. Prior work favors the latter, showing that reinstatement improves memory. Three experiments extended this work by directly testing roles for study-phase retrievals and change awareness during study and test. Word pairs with shared cues but changed responses (A-B, A-D) were presented with background contexts that either repeated or changed. Repeating contexts increased detection of changes and recall of earlier responses during study, both indexes of study-phase retrievals, as well as later cued recall of earlier (B) and changed (D) responses. The recall benefit was proportional to the extent of study-phase retrievals, implicating retrieval practice. Moreover, the effect was enhanced when participants remembered that changes had occurred, highlighting the role of recollecting integrated representations that included change attributes. These findings align with integration theory, suggesting that memory updating is most effective when current events cue retrieval of prior memories and engender associative encoding of past and present events, establishing elaborate representations that support subsequent recall.

Multiword expressions-also called multiword chunks, fixed expressions, lexical bundles, or formulaic sequences-are familiar sequences of words that occur with high frequency in language. Recent focus on multiword expressions, as distinct units of language with distinct processing ramifications, raises the question of how they are used during second language processing: Although processing of multiword expressions appears to grow more native-like as proficiency increases, even highly proficient L2 speakers may not process multiword expressions in entirely native-like ways. We conducted a visual world eye tracking study of idiom comprehension to examine how the balance between compositional and whole-phrase processing changes in second language comprehension. L1 and advanced L2 English speakers listened to sentences consisting of a literal or figurative biasing context clause and a target clause with the final word missing, while simultaneously viewing images representing possible literal or figurative continuations of the clause. Growth curve analysis of eye movement data provided suggestive evidence that L1 and L2 speakers were biased towards different processing strategies during idiom comprehension, with L2 speakers, in particular, taking a more compositional approach towards computing figurative meaning. We also found evidence of qualitative processing differences between literal and figurative expressions, with literal processing being strongly driven by anticipatory mechanisms and figurative processing being driven by within-phrase associations for both groups.

Future and past events can come to mind spontaneously. Previous research has focused on the characteristics of future and past event simulations, but few studies have investigated the types of input that initiate these simulations. To gain insight into the underlying mechanisms, we used event and location cues to elicit spontaneous future and past simulations. We investigated two novel questions. How do event and location cues influence the spontaneous events reported? Do future events, like spontaneous autobiographical memories, occur in chained-event sequences consisting of multiple consecutively generated related events? In a vigilance task, 127 participants located a left-facing arrow among right-facing arrows. On 49 of 350 trials, participants encountered an event or location cue. On 10 occasions, participants audio-recorded off-task thoughts they felt comfortable sharing. We contribute four novel insights. Participants produced a greater number of future events for event than for location cues, whereas they produced a greater number of past events for location cues. For event cues, hypothetical future events were more frequent than planning. Spontaneous future simulations, like past event simulations, occurred in chained event sequences. Finally, the distributions of connections within future and past chains were influenced by the original cue. We integrate ideas from the Constructive Episodic Simulation Hypothesis and the dual-process account of deliberate and spontaneous future thoughts by treating constructive memory in terms of pattern completion.

Consensus among sources often signals a claim's trustworthiness. However, when all sources simply echo information from a single origin, an "illusion of consensus" can arise, leading individuals to feel equally convinced by a dependent consensus (a single, repeatedly cited source) as by an independent consensus (corroborated information from multiple sources). This effect may occur because repetition can be perceived as a cue for credibility. We conducted four experiments to test whether a negative explanation (e.g., repetition was intended to sway opinion) or a positive explanation (repetition occurred because a source was especially reliable) for repetition of a claim could reduce or enhance the illusion of consensus. Participants engaged in a political poll-reporting task to assess confidence in claims supported by an independent consensus (news outlets citing different polling companies), a dependent consensus without explanation (news outlets citing the same polling company), or a dependent consensus accompanied by explanations intended to either boost or undermine source credibility. Independent consensus was more persuasive than dependent consensus. However, the persuasiveness of dependent consensus was increased or decreased, respectively, by positive or negative explanations for repetition. These effects were found when the explanations were supplied by the experimenter (Experiments 1-2a, 2b) and when participants generated their own explanations (Experiments 3, 4). These findings suggest that social inference processes contribute to the illusion of consensus, with implications for interventions designed to mitigate the impact of misleading claims repeated in media.

The nature of forgetting has long been a critical issue in the study of serial order memory, with output interference recognized as a significant factor in shaping serial position curves. However, it remains unclear whether this interference arises from active representations following retrieval (such as capacity consumption, overwriting of other items, blocking retrieval of other information, increased competitions) or if it stems directly from the act of retrieval itself. The present study sought to clarify these possibilities by utilizing retrieval-induced forgetting paradigms, which have been primarily developed within the domain of single-item memory. Through four preregistered experiments, we found that engaging in retrieval practice for part of an ordered list led to reduced order accuracy for the remaining list items. However, the effect size of this retrieval practice was comparable to that of merely re-presenting order information in working memory without engaging in retrieval, suggesting that the observed impairment was not retrieval-specific. This retrieval unspecificity was replicated in an additional experiment. This indicates that output interference in serial order memory is more likely attributable to active representations maintained in working memory, rather than being a direct consequence of retrieval. These findings are discussed within the framework of the two-factor account of retrieval-induced forgetting in the domain of item memory, which posits that both inhibition and competition play roles in the forgetting process.

Event-files are a central concept in human action control as they integrate perception and action. The basic idea is that event-files comprise internal representations of stimuli, motor programs, and sensory effects of actions that are bound together, maintained for some time, and then ultimately decay. Yet the exact nature of the decay of event-files remains elusive. Most previous studies analyzed the decay only on a very coarse-grained scale typically with only two or three different response-stimulus intervals (RSI). Because available data are inconsistent, the decay function could as yet not be delineated, particularly whether it is linear or resembles a typical forgetting curve. In the present study, we analyzed the decay of event-files using nine different RSIs between 400 ms and 2,000 ms in a large sample (N = 156) of neurotypical participants. If event-file decay mimics the typical forgetting curve, the concept of event-files can be better connected to the memory literature. Using unsupervised cluster analysis, the results showed, however, a more complex pattern. In fact, there were different decay functions across participants with only a subset showing a forgetting curve suggesting that (i) event-file decay functions are modulated by interindividual differences and (ii) that the previously observed power function might be a statistical artifact due to averaging. At least in the time range investigated in this study, event-files decay only for a subset of participants. Possible underlying processes are discussed.

In the real world, memory and navigation operate through the interplay between visual, proprioceptive, and vestibular systems to give rise to multimodal spatial representations of the environment. However, it remains unclear in what situations the absence of ambulatory cues might affect spatial memory, particularly given the widespread use of desktop virtual reality (VR) in studies of human spatial memory and navigation. In the present study, we employ ambulatory and immersive VR technology to probe people's memory for spatial layouts across the same or rotated perspectives. Critically, we explored whether cognitive processes underlying spatial memory for perspective change differ depending on whether self-motion cues are available to support performance (i.e., when perspective changes due to observer movement as opposed to display rotation). Leveraging the wealth of behavioral data that can be obtained from immersive VR technology, we demonstrate that self-motion cues enhance spatial memory accuracy by shaping the temporal dynamics of decision making, pupillary responses, and eye-movement behaviors. This study highlights the importance of integrating naturalistic movement cues, particularly those involving rotation, into ecologically valid paradigms which will enable further advancements of theories on real-world navigation.

Familiarity is the sense that something has been encountered before, either prior to or without being able to recall specifics of the previous encounter. Familiarity is thought to be driven by an overlap in features between stimuli recently activated in memory and the current stimulus. Research on word familiarity suggests that letters are a significant contributing feature. An unexplored question is the extent to which letter overlap needs to occur in the same positions between the initial encounter and the later encounter. Research on priming and lexical decision suggests that letters do not need to be in the exact correct locations for lexical activation to occur. Based on this research, two experiments are reported investigating how letter position matching might operate in word familiarity using a variant of recognition without cued recall. The results suggested that neither absolute nor relative position overlap were needed for some degree of letter matching to occur. Letters that were out of position that also did not maintain relative positions still contributed to word familiarity. Implications for the mechanism behind letter matching, as well as for other kinds of stimuli, are discussed.

Output interference in recognition refers to a decrease in performance over the course of a test. The goal of the current study was to determine whether experimentally shifting the decision criterion changes the form of output interference and to identify a process account of any interaction. In two experiments, we manipulated the decision criterion via changes in the base rate of the old items at test (80%, 50%, 20%). Experiment 1 implemented this manipulation within-subjects and failed to induce criterion shifts. In contrast, when the base rate was manipulated between-subjects in Experiment 2, decision criteria differed across conditions. Qualitative patterns suggested that liberal criteria attenuated the hit rate (HR) decline and increased the false alarm rate (FAR) across blocks, whereas conservative criteria yielded steeper HR declines with relatively stable FAR. To further examine this effect, Experiment 3 employed longer test lists and a larger sample. The criterion was manipulated only via prior information about the base rates, while the actual base rate was 50% in all conditions. Experiment 3 revealed a significant interaction between response bias and output interference in HR and FAR. When we used an independent data set (Layher et al., Journal of Experimental Psychology: Learning, Memory, and Cognition, 46(11), 2075-2105, 2020), we demonstrated the same patterns. To account for these findings, we conducted simulations with the Retrieving Effectively from Memory (Shiffrin & Steyvers, Psychonomic Bulletin & Review, 4(2), 145-166, 1997) model. The results were best captured by a learning-during-test mechanism in which every test item is encoded as a new memory trace.