A Distributed Representation of Temporal Context

The principles of recency and contiguity are two cornerstones of the theoretical and empirical analysis of human memory. Recency has been alternatively explained by mechanisms of decay, displacement, and retroactive interference. Another account of recency is based on the idea of variable context (Estes, 1955; Mensink & Raaijmakers, 1989). Such notions are typically cast in terms of a randomly fluctuating population of elements reflective of subtle changes in the environment or in the subjects' mental state. This random context view has recently been incorporated into distributed and neural network memory models (Murdock, 1997; Murdock, Smith, & Bai, 2001). Here we propose an alternative model. Rather than being driven by random fluctuations, this formulation, the temporal context model (TCM), uses retrieval of prior contextual states to drive contextual drift. In TCM, retrieved context is an inherently asymmetric retrieval cue. This allows the model to provide a principled explanation of the widespread advantage for forward recalls in free and serial recall. Modeling data from single-trial free recall, we demonstrate that TCM can simultaneously explain recency and contiguity effects across time scales.     

Associative processes in immediate recency

Some theorists have argued that the immediate recency effect observed in free recall reflects the emptying out of the contents of a short-term memory buffer (Davelaar, Goshen-Gottstein, Ashkenazi, Haarmann, & Usher, 2005). Others have argued that immediate recency reflects the properties of temporal context used to cue free recall (Howard & Kahana, 2002). We examined immediate free recall of lists with an item from the middle of the list repeated at or near the end. If associative processes contribute to immediate recency, as predicted by the temporal context account, the neighbors of the initial presentation of the repeated item should show enhanced recall at the initial stages of immediate recall. Recall transitions early in output--and even the initiation of recall itself-showed evidence for temporally defined associations that resemble those observed in recall from long-term memory. These results have strong implications for models of the immediate recency effect in free recall.     

Reply to Farrell and Lewandowsky: Recency—contiguity interactions predicted by the temporal context model

Farrell and Lewandowsky (2008) argued that the temporal context model (TCM; Howard & Kahana, 2002) cannot explain nonmonotonicities in the contiguity effect seen at extreme lags. However, TCM actually predicts these nonmonotonicities to the extent that end-of-list context persists as a retrieval cue during recall and to the extent that end-of-list context generates a recency effect. We show that the observed nonmonotonicity in the contiguity effect interacts with the recency effect, as predicted by TCM. In conditions, such as immediate and continualdistractor free recall, that exhibit strong recency, one observes more prominent nonmonotonicities in the contiguity effect than in conditions, such as delayed free recall, that attenuate recency. The nonmonotonicities in the contiguity effect at extreme lags, and the interactions between recency and contiguity, result from the role of end-of-list context as a retrieval cue in TCM. Results of an additional simulation based on the Howard and Kahana (2002) version of TCM may be downloaded from http://pbr.psychonomic-journals.org/content/supplemental.     

Contextual variability and serial position effects in free recall.

In immediate free recall, words recalled successively tend to come from nearby serial positions. M. J. Kahana (1996) documented this effect and showed that this tendency, which the authors refer to as the lag recency effect, is well described by a variant of the search of associative memory (SAM) model (J. G. W. Raaijmakers & R. M. Shiffrin, 1980, 1981). In 2 experiments, participants performed immediate, delayed, and continuous distractor free recall under conditions designed to minimize rehearsal. The lag recency effect, previously observed in immediate free recall, was also observed in delayed and continuous distractor free recall. Although two-store memory models, such as SAM, readily account for the end-of-list recency effect in immediate free recall, and its attenuation in delayed free recall, these models fail to account for the long-term recency effect. By means of analytic simulations, the authors show that both the end of list recency effect and the lag recency effect, across all distractor conditions, can be explained by a single-store model in which context, retrieved with each recalled item, serves as a cue for subsequent recalls.     

Aging and contextual binding: Modeling recency and lag recency effects with the temporal context model

Normal aging has been shown to spare recency effects in the initiation of free recall while disrupting temporally defined associations. The temporal context model (TCM) explains recency and temporally defined associations as consequences of a gradually changing context signal and recovery of those contextual states, respectively. Here we extend TCM to account for the dissociation between recency and temporally defined associations in younger and older adults. Modeling results suggested that the effect of aging was restricted to a decrement in the ability of items to recover the temporal contexts in which they were presented, a function that has been hypothesized to depend on the hippocampus.     

A context-based theory of recency and contiguity in free recall

The authors present a new model of free recall on the basis of M. W. Howard and M. J. Kahana's temporal context model and M. Usher and J. L. McClelland's leaky-accumulator decision model. In this model, contextual drift gives rise to both short-term and long-term recency effects, and contextual retrieval gives rise to short-term and long-term contiguity effects. Recall decisions are controlled by a race between competitive leaky accumulators. The model captures the dynamics of immediate, delayed, and continual distractor free recall, demonstrating that dissociations between short- and long-term recency can naturally arise from a model in which an internal contextual state is used as the sole cue for retrieval across time scales.     

Effects of aging and IQ on item and associative memory

The effects of aging and IQ on performance were examined in 4 memory tasks: item recognition, associative recognition, cued recall, and free recall. For item and associative recognition, accuracy and the response time (RT) distributions for correct and error responses were explained by Ratcliff's (1978) diffusion model at the level of individual participants. The values of the components of processing identified by the model for the recognition tasks, as well as accuracy for cued and free recall, were compared across levels of IQ (ranging from 85 to 140) and age (college age, 60-74 years old, and 75-90 years old). IQ had large effects on drift rate in recognition and recall performance, except for the oldest participants with some measures near floor. Drift rates in the recognition tasks, accuracy in recall, and IQ all correlated strongly. However, there was a small decline in drift rates for item recognition and a large decline for associative recognition and cued recall accuracy (70%). In contrast, there were large effects of age on boundary separation and nondecision time (which correlated across tasks) but small effects of IQ. The implications of these results for single- and dual-process models of item recognition are discussed, and it is concluded that models that deal with both RTs and accuracy are subject to many more constraints than are models that deal with only one of these measures. Overall, the results of the study show a complicated but interpretable pattern of interactions that present important targets for modeling.     

Age dissociates recency and lag recency effects in free recall

The temporal relations among word-list items exert a powerful influence on episodic memory retrieval. Two experiments were conducted with younger and older adults in which the age-related recall deficit was examined by using a decomposition method to the serial position curve, partitioning performance into (a) the probability of first recall, illustrating the recency effect, and (b) the conditional response probability, illustrating the lag recency effect (M. W. Howard & M. J. Kahana, 1999). Although the older adults initiated recall in the same manner in both immediate and delayed free recall, temporal proximity of study items (contiguity) exerted a much weaker influence on recall transitions in older adults. This finding suggests that an associative deficit may be an important contributor to older adults' well-known impairment in free recall.     

The relationship between memory performance and the number of rehearsals in free recall

The purpose of the present study was to examine the relationship between the number of rehearsals and long-term recall performance by means of an overt rehearsal procedure. Subjects were induced to concentrate their rehearsal activity on specific items in free recall tasks. In Experiment 1, both the primacy and the recency items that had received additional rehearsals were recalled with higher probability than were the ordinarily rehearsed items in immediate recall but not in final recall. Experiment 2 was designed to extend the occasions for subjects’ rehearsal by manipulating the rate of item presentation. The overall pattern of resulting data shewed that for neither recency items nor primacy items does the additional overt rehearsal reliably lead to facilitative effects on long-term recall performance. The possibility of a qualitative change in rehearsal is discussed.     

Temporal associations and prior-list intrusions in free recall

When asked to recall the words from a just-presented target list, subjects occasionally recall words that were not on the list. These intrusions either appeared on earlier lists (prior-list intrusions, or PLIs) or had not appeared over the course of the experiment (extra-list intrusions). The authors examined the factors that elicit PLIs in free recall. A reanalysis of earlier studies revealed that PLIs tend to come from semantic associates as well as from recently studied lists, with the rate of PLIs decreasing sharply with list recency. The authors report 3 new experiments in which some items in a given list also appeared on earlier lists. Although repetition enhanced recall of list items, subjects were significantly more likely to make PLIs following the recall of repeated items, suggesting that temporal associations formed in earlier lists can induce recall errors. The authors interpret this finding as evidence for the interacting roles of associative and contextual retrieval processes in recall. Although contextual information helps to focus recall on words in the target list, it does not form an impermeable boundary between current- and prior-list experiences.     

The word-length effect in probed and serial recall

The word-length effect in immediate serial recall has been explained as the possible consequence of rehearsal processes or of output processes. In the first experiment adult subjects heard lists of five long or short words while engaging in articulatory suppression during presentation. Full serial recall or probed recall for a single item followed the list either immediately or after a 5-second delay to encourage rehearsal. The word-length effect was not influenced by recall delay, but was much smaller in probed than in serial recall. Examination of the serial position curves suggested that this might be due to a recency component operating in probed recall. Experiment 2 confirmed a word-length-insensitive recency effect in probed recall and showed that this was resistant to an auditory suffix, unlike the small recency effect found in serial recall. Experiment 3 used visual presentation without concurrent articulation. Under these conditions there was no recency effect for either recall method, but the word-length effect was again much smaller in probed than in serial recall. This was confirmed in Experiment 4, in which the presentation of serial and probed recall was randomized across trials, showing that the differences between recall methods could not be due to encoding strategies. We conclude that for visual presentation, at least part of the word-length effect originates in output processes. For auditory presentation the position is less clear, as serial and probed recall appear to draw on different resources. The nature of the output processes that may give rise to word-length effects is discussed.     

Going beyond a single list: Modeling the effects of prior experience on episodic free recall

We present an extension of the search of associative memory (SAM) model that simulates the effects of both prior semantic knowledge and prior episodic experience on episodic free recall. The model incorporates a memory store for preexisting semantic associations, a contextual drift mechanism, a memory search mechanism that uses both episodic and semantic associations, and a large lexicon including both words from prior lists and unpresented words. These features enabled the model to successfully account for the effects of prior semantic knowledge and prior episodic learning on the pattern of correct recalls and intrusions observed in free recall experiments.     

Associative retrieval processes in free recall

I present a new method for analyzing associative processes in free recall. While previous research has emphasized the prominence of semantic organization, the present method illustrates the importance of association by contiguity. This is done by examining conditional response probabilities in the output sequence. For a given item recalled, I examine the probability and latency that it follows an item from a nearby or distant input position. These conditional probabilities and latencies, plotted as a function of the lag between studied items, reveal several regularities about output order in free recall. First, subjects tend to recall items more often and more rapidly from adjacent input positions than from remote input positions. Second, subjects are about twice as likely to recall adjacent pairs in the forward than in the backward direction and are significantly faster in doing so. These effects are observed at all positions in the output sequence. The asymmetry effect is theoretically significant because, in cued recall, nearly symmetric retrieval is found at all serial positions (Kahana, 1995; Murdock, 1962). An attempt is made to fit the search of associative memory model (Raaijmakers & Shiffrin, 1980, 1981) with and without symmetric interitem associations to these data. Other models of free recall are also discussed.     

Recency-sensitive retrieval processes in long-term free recall

In several experiments, each presentation of a to-be-remembered item in a free-recall list was both preceded and followed by a distracting activity and recall was delayed by an additional period of distracting activity. Pronounced long-term effects of recency were obtained, the standard short-term memory interpretation of recency effects in free recall notwithstanding. The results are interpreted as reflecting retrieval processes that are obscured by procedural characteristics of typical free-recall experiments.     

Primacy and recency effects in immediate free recall of sequences of spatial positions

This study evaluated the serial position curve based on free recall of spatial position sequences. To evaluate the memory processes underlying spatial recall, some manipulations were introduced by varying the length of spatial sequences (Exp. 1) and modifying the presentation rate of individual positions (Exp. 2). A primacy effect emerged for all sequence lengths, while a recency effect was evident only in the longer sequences. Moreover, slowing the presentation rate increased the magnitude of the primacy effect and abolished the recency effect. The main novelty of the present results is represented by the finding that better recall of early items in a sequence of spatial positions does not depend on the task requirement of an ordered recall but it can also be observed in a free recall paradigm.     

Evidence against a short-term-store account of long-term recency effects

When subjects perform a distractor task before and after every item on a list, recall of the last itemis much higher than recall of items from the middle of the list. Koppenaal and Glanzer (1990) have shown that this long-term recency effect can be eliminated by using, after the last item, a distractor task different from that used elsewhere on the list. They interpreted this finding as evidence in favor of a short-term-store account of long-term recency effects. This account is challenged by the results reported here. Practice either on the task or on time-sharing between the task and list items had little impact on the recency effect. Also, substantial recency effects were found when a different distractor task occurred after every list position. Thus, it is not true that long-term recency effects are found only when subjects have an opportunity to adapt to the distractor task. Our results are not consistent with a short-term-store account of recency effects.     

Recency and suffix effects in serial recall of musical stimuli

Auditory presentation of verbal items leads to larger recency effects in recall than visual presentation. This enhanced recency can be eliminated if a stimulus suffix (an irrelevant sound) follows the last item. Four experiments tested the hypothesis that recency and suffix effects in serial recall result from a speech-specific process. It was demonstrated that serial recall of musical notes played on a piano exhibited substantial recency effects. These recency effects were reduced when the list items were followed by either a piano chord or the word start. However, a white-noise suffix had no effect on recency. This pattern of data is consistent with current work on auditory perception and places constraints on theories of recency and suffix effects.     

Integrating spatial context learning over contradictory signals: Recency effects in contextual cueing

Contextual cueing is a visual search phenomenon in which memory of global visual context guides spatial attention towards task-relevant portions of the search display. Recent work has shown that the learning processes underlying contextual cueing exhibit primacy effects; they are more sensitive to early experience than to later experience. These results appear to pose difficulties for associative accounts which typically predict recency effects; behaviour being most strongly influenced by recent experience. The current study utilizes trial sequences that consist of two contradictory sets of regularities. In contrast to previous results, robust recency effects were observed. In a second study it is demonstrated that this recency effect can be minimized, but not reversed by systematically manipulating task-irrelevant features of the search display. These results provide additional support for an associative account of contextual cueing and suggest that contextual cueing may, under some circumstances, be more sensitive to recent experience.     

Children's cued recall: Developmental differences in retrieval operations

Second and fifth grade children were aurally presented a study list containing eight target items plus primacy and recency buffers, followed first by free recall, then by cued recall for items not produced in free recall. Two additional buffer items were used at Grade 5 in an effort to equate the two age groups in free recall of targets. The cues were either of high or of moderate associative strength, and equal for the two age groups. Free recall of targets did not differ with age, suggesting equivalent storage of targets. At Grade 2 the cued recall level matched expectations based on a simple associative account of cuing effects. However, cued recall at Grade 5 was much higher than the normative association values, implicating strategic retrieval operations, the possible nature of which was discussed.     

Recency effects in delayed recall of mouthed stimuli

Auditory presentation leads to greater recency effects in recall than does visual presentation. This phenomenon (the modality effect) is found in both free and serial recall and in both immediate and delayed recall. Silent mouthing of visually presented stimuli also leads to enhanced recency effects in immediate serial recall. Two experiments reported here extend the generality of the mouthing effect by demonstrating that enhanced recency effects of mouthed stimuli occur in delayed serial and free recall. These results are inconsistent with theories that attribute the modality effect to a purely auditory sensory memory.