Item Interference and Time Delays in Working Memory: Immediate Serial Recall

Limitations on immediate recall span have been related by some to temporal limitations of an articulatory rehearsal loop (Baddeley, 1986), and by others to retroactive interference from subsequent items (Lewandowsky & Murdock, 1989; Neath & Nairne, 1995). In the current study, the number of intervening items and retention delays were partially decoupled by varying presentation rate and materials sets. Recently, memory span effects often attributed to temporal limits in rehearsal were shown to be directly related to the output delays during the act of recall (Dosher & Ma, 1998). A functional model of working memory provides a framework within which to evaluate the relative contributions of item interference and retention delays in the accuracy of recall at each serial position. Output delays during recall and interference from items during study were the primary factors in limiting recall, with small additional effects of time delays during study.     

Distinctiveness and serial position effects in tonal sequences.

The proportion-of-the-total-duration rule (Kidd & Watson, 1992) states that the detectability of a change in a component of a tonal sequence can be predicted by the proportional duration of the changed component relative to the length of the sequence as a whole. A similar viewpoint relies on temporal distinctiveness to account for primacy, recency, and other serial position effects in memory (Murdock, 1960; Neath, 1993a, 1993b). Such distinctiveness models predict that an item will be remembered if it is more distinctive along some dimension relative to possible competitors. Three experiments explored the relation between distinctiveness and proportional duration by examining the effects of the proportion of the total duration of a tone in a sequence, serial position, and interstimulus interval (ISI) on the detection of a change in one component of a tonal sequence. Experiment 1 replicated the basic effect with relatively untrained subjects and a fixed frequency difference. Experiment 2 showed that distinctiveness holds for tonal sequences and a same/different task. Experiment 3 combined the two to show that proportional duration, ISI, and position of the changed tone all contribute to discrimination performance. The present research combines theories that have been proposed in the psychophysics and memory fields and suggests that a comprehensive principle based on relative distinctiveness may be able to account for both perceptual and memory effects.     

Distinctiveness and serial position effects in tonal sequences

The proportion-of-the-total-duration rule (Kidd & Watson, 1992) states that the detectability of a change in a component of a tonal sequence can be predicted by the proportional duration of the changed component relative to the length of the sequence as a whole. A similar viewpoint relies on temporal distinctiveness to account for primacy, recency, and other serial position effects in memory (Murdock, 1960; Neath, 1993a, 1993b). Such distinctiveness models predict that an item will be remembered if it is more distinctive along some dimension relative to possible competitors. Three experiments explored the relation between distinctiveness and proportional duration by examining the effects of the proportion of the total duration of a tone in a sequence, serial position, and interstimulus interval (ISI) on the detection of a change in one component of a tonal sequence. Experiment 1 replicated the basic effect with relatively untrained subjects and a fixed frequency difference. Experiment 2 showed that distinctiveness holds for tonal sequences and a same/different task. Experiment 3 combined the two to show that proportional duration, ISI, and position of the changed tone all contribute to discrimination performance. The present research combines theories that have been proposed in the psychophysics and memory fields and suggests that a comprehensive principle based on relative distinctiveness may be able to account for both perceptual and memory effects.     

Learning representations of wordforms with recurrent networks: comment on sibley, kello, plaut, & elman (2008)

Sibley et al. (2008) report a recurrent neural network model designed to learn wordform representations suitable for written and spoken word identification. The authors claim that their sequence encoder network overcomes a key limitation associated with models that code letters by position (e.g., CAT might be coded as C‐in‐position‐1, A‐in‐position‐2, T‐in‐position‐3). The problem with coding letters by position (slot‐coding) is that it is difficult to generalize knowledge across positions; for example, the overlap between CAT and TOMCAT is lost. Although we agree this is a critical problem with many slot‐coding schemes, we question whether the sequence encoder model addresses this limitation, and we highlight another deficiency of the model. We conclude that alternative theories are more promising.     

Evidence for similar principles in episodic and semantic memory: The presidential serial position function

When people recall a list of items that they have just experienced (an episodic memory task), the resulting serial position function looks strikingly similar to that observed when people are asked to recall the presidents of the United States (a semantic memory task). Despite the similarity in appearance, there is disagreement about whether the two functions arise from the same processes. A local distinctiveness model of memory, SIMPLE, successfully fit the presidential data using two underlying dimensions: one corresponding to item (or presidential) distinctiveness and the other to order (or positional) distinctiveness. According to the model, presidential primacy and recency are due to the same mechanisms that give rise to primacy and recency effects in both shortand long-term episodic memory. All of these primacy and recency effects reflect the relative distinctiveness principle (Surprenant & Neath, 2009): Items will be well remembered to the extent that they are more distinct than competing items at the time of retrieval.     

Pauses and durations exhibit a serial position effect

This article reports evidence of two kinds of serial position effects in immediate serial recall: One involves interresponse pauses, and the other response durations. In forward and backward recall, responding was faster at initial and final positions than at center positions, exhibiting a bow-shaped function relative to serial position. These data were obtained in a spoken recall study in which ungrouped lists of four to six words and postcuing of recall direction were used. The pause pattern is consistent with several models of serial memory, including a distinctiveness model (Brown, Neath, & Chater, 2002) and a version of the ACT—R model augmented with a spontaneous grouping strategy (Maybery, Parmentier, & Jones, 2002). The duration pattern suggests that response articulation depends on the processing context, rather than being modular.     

Positional information in short-term memory: Relative or absolute?

Evidence suggests that short-term memory for serial order includes information about the positions of items in a sequence. This information is necessary to explain why substitution errors between sequences tend to maintain their position within a sequence. Previous demonstrations of such errors, however, have always used sequences of equal length. With sequences of different length, both transpositions between groups (Experiment 1) and intrusions between trials (Experiment 2) are shown to respect position relative to the end as well as to the start of a sequence. These results support models in which position is coded by start and end markers, but not models in which position is coded in temporal or absolute terms. Possible interpretations of an end marker are discussed.     

Distinctiveness in serial memory for spatial information

Several studies have shown that recall performance depends on the extent to which an item differs from other items in a sequence (the distinctiveness effect; see, e.g., Kelley & Nairne, 2001). Distinctiveness effects, however, have been demonstrated mainly in the verbal domain. The present study extends distinctiveness effects to the spatial domain. In two experiments, participants recalled the order in which series of spatially located dots had been presented. Item discriminability was varied within the sequence by manipulating the duration of the interval inserted between the presentation of the dots (Experiment 1) and the perceptual characteristics of the stimuli (Experiment 2). The results showed that these manipulations in the spatial domain produce distinctiveness effects similar to those observed with verbal material (see, e.g., Neath & Crowder, 1990) and suggest that distinctiveness models of memory should take into account the processing of spatial information.     

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.     

Grouping in short-term verbal memory: Is position coded temporally?

The nature of the mechanisms that code item position in serial short-term verbal recall was investigated with reference to temporal grouping phenomena--effects that arise when additional pauses are inserted in a presented list to form groups of items. Several recent models attempt to explain these phenomena by assuming that positional information is retained by associating items with contextual information. According to two of the models--the Phonological Loop model (Hitch, Burgess, Towse, & Culpin, 1996) and the OSCAR model (Brown, Preece, & Hulme, 2000)--contextual information depends critically on the timing of item presentation with reference to group onset. By contrast, according to the Start-End model (Henson, 1998) and a development from it, which we label the Oscillator-Revised Start-End model (Henson & Burgess, 1997), contextual information is independent of time from group onset. Three experiments examined whether coding of position is time dependent. The critical manipulation was to vary stimulus-onset asynchrony from one group to the next in the same list. Lists of consonants were presented visually, but with vocalization in Experiment 1, auditorily in Experiment 2, and auditorily with articulatory suppression in Experiment 3. The pattern of order errors consistently favoured the predictions of the time-independent models over those of the time-dependent models in that across-group transpositions reflected within-group serial position rather than time from group onset. Errors involving intrusions from previous lists also reflected within-group serial position, thereby extending support for the time-independent models.     

Modelling serial position curves with temporal distinctiveness

We offer a critique of the temporal distinctiveness model of serial position effects (Nairne, Neath, Serra, & Byun, 1997). The temporal distinctiveness model combines a precise definition of stimulus distinctiveness with a memory perturbation process. The critique is empirically motivated—we show that with a more complete analysis, the temporal distinctiveness model does not adequately account for Nairne et al.' experimental data. To better account for the data, we independently modified two components of Nairne et al.' model: the mathematical form of the definition of temporal distinctiveness and the mathematical form of the mapping from distinctiveness to free-recall probabilities. Both of these modifications provided for better fits. Yet both Nairne et al.' definition and our modified definition are fairly arbitrary. We show that a significant challenge to this approach is to find theoretically motivated constraints of the temporal distinctiveness model while providing for adequate fits to data.     

Position of code and code for position: From isomorphism to a sensorimotor account of space perception

The paper starts with a discussion of the assumption that positions of the outer world are coded by the anatomical locations of firing neurons within retinotopic maps (“coding position by position”). This “code position theory” explains space perception by some kind of structural isomorphism since it implies that the perceptual space is based on a spatial structure within the brain. The axiom of structural isomorphism is rejected. Subsequently, a sensorimotor account of space perception is outlined according to which the spatial structure of the outer world is coded by the temporal structure of cortical processing. The basis is that action changes the perceiver's relationship to the outer world and, therefore, changes the representation of the outer world coded by the sensory responses of the brain. According to this view the code for position is not a spatial but a temporal structure resulting from action (“coding position by action”). The sensorimotor account offers a possible solution to the binding problem. The paper ends with some remarks on the possible origin and function of retinotopic representations.     

Deciding about decision models of remember and know judgments: a reply to Murdock (2006)

B. B. Murdock (2006) has interpreted remember-know data within a decision space defined by item and associative information, the fundamental variables in his general recognition memory model TODAM (B. B. Murdock, 1982). He has related parameters of this extended model to stimulus characteristics for several classic remember-know data sets. The authors show that this accomplishment is shared by both one- and two-dimensional signal-detection-based remember-know models (J. C. Dunn, 2004; C. M. Rotello, N. A. Macmillan, & J. A. Reeder, 2004; J. T. Wixted & V. Stretch, 2004), which can be represented in this same decision space and can be related to stimulus characteristics with similar success. Murdock claims that his model, unlike its competitors, is a process model; however, the process aspects of TODAM are not used in his application, and the decision aspects are identical to a previously proposed model. Murdock's claim that one competing model (STREAK; C. M. Rotello et al., 2004) is not fully specified is shown to be false. The new model is not superior to existing ones, but comparisons among the models to date are not definitive. The authors describe several strategies that might be applied to distinguish among them.     

Modelling serial position curves with temporal distinctiveness

We offer a critique of the temporal distinctiveness model of serial position effects (Nairne, Neath, Serra, & Byun, 1997). The temporal distinctiveness model combines a precise definition of stimulus distinctiveness with a memory perturbation process. The critique is empirically motivated-we show that with a more complete analysis, the temporal distinctiveness model does not adequately account for Nairne et al.'s experimental data. To better account for the data, we independently modified two components of Nairne et al.'s model: the mathematical form of the definition of temporal distinctiveness and the mathematical form of the mapping from distinctiveness to free-recall probabilities. Both of these modifications provided for better fits. Yet both Nairne et al.'s definition and our modified definition are fairly arbitrary. We show that a significant challenge to this approach is to find theoretically motivated constraints of the temporal distinctiveness model while providing for adequate fits to data.     

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.     

Remembering plurals: Unit of coding and form of coding during serial recall

Subjects were required to recall lists of six words which had been presented visually in sequence. One or two of the words might be plural nouns. A substantial number of errors occurred in which the plural form became detached from its original root. This is taken as evidence for a morpheme-based code as opposed to a unitary word code. A significantly high proportion of these errors maintained the form of the plural ( /S/, /Z/ or (ЪZ/) which is considered evidence for a primarily phonological coding of the plural morpheme. There is however a suggestion that on some occasions the plural is coded morphemically—i.e., in a way which does not distinguish between the various plural endings.     

Short- and long-term memory contributions to immediate serial recognition: Evidence from serial position effects

A long-standing body of research supports the existence of separable short- and long-term memory systems, relying on phonological and semantic codes, respectively. The aim of the current study was to measure the contribution of long-term knowledge to short-term memory performance by looking for evidence of phonologically and semantically coded storage within a short-term recognition task, among developmental samples. Each experimental trial presented 4-item lists. In Experiment 1 typically developing children aged 5 to 6 years old showed evidence of phonologically coded storage across all 4 serial positions, but evidence of semantically coded storage at Serial Positions 1 and 2. In a further experiment, a group of individuals with Down syndrome was investigated as a test case that might be expected to use semantic coding to support short-term storage, but these participants showed no evidence of semantically coded storage and evidenced phonologically coded storage only at Serial Position 4, suggesting that individuals with Down syndrome have a verbal short-term memory capacity of 1 item. Our results suggest that previous evidence of semantic effects on “short-term memory performance” does not reflect semantic coding in short-term memory itself, and provide an experimental method for researchers wishing to take a relatively pure measure of verbal short-term memory capacity, in cases where rehearsal is unlikely.     

Verb and auxiliary movement in agrammatic Broca's aphasia

Verb production in agrammatic Broca's aphasia has repeatedly been shown to be impaired by a number of investigators. Not only is the number of verbs produced often significantly reduced, but verb inflections and auxiliaries are often omitted as well (e.g., Bastiaanse, Jonkers, & Moltmaker-Osinga, 1996; Saffran, Berndt, & Schwartz, 1989; Thompson, Shapiro, Li, &Schendel, 1994, 1997). It has been suggested that these problems are, in part, caused by the fact that finite verbs need to be moved from their base-generated position to inflectional nodes in the syntactic tree (e.g., Bastiaanse & Van Zonneveld, 1998). Others have suggested that production deficits in agrammatism can be predicted based on the position that certain structures take in the syntactic tree (Friedmann & Grodzinsky, 1997; Hagiwara, 1995). If the former theory is correct, several predictions can be made. First of all, the discrepancy between production of finite verbs in the matrix and embedded clause that has been found for Dutch (Bastiaanse & Van Zonneveld, 1998) should not be observed in English, since the word order of the matrix and embedded clause are the same in the latter language. Second, if verb movement (including movement of auxiliaries) is problematic for speakers with agrammatic aphasia, then a hierarchy in the production of auxiliaries in yes/no questions, auxiliaries, and finite verbs in declarative sentences in English would be expected, since the former has been moved and the two latter are in base-generated position. In the present paper, these hypotheses were tested in a cross-linguistic study of Dutch and English. Results showed the position in the syntactic tree does not predict deficit patterns; rather the critical factor appears to relate to whether or not verb or auxiliary movement is required.     

Visual representation of eye gaze is coded by a nonopponent multichannel system

To date, there is no functional account of the visual perception of gaze in humans. Previous work has demonstrated that left gaze and right gaze are represented by separate mechanisms. However, these data are consistent with either a multichannel system comprising separate channels for distinct gaze directions (e.g., left, direct, and right) or an opponent-coding system in which all gaze directions are coded by just 2 pools of cells, one coding left gaze and the other right, with direct gaze represented as a neutral point reflecting equal activation of both left and right pools. In 2 experiments, the authors used adaptation procedures to investigate which of these models provides the optimal account. Both experiments supported multichannel coding. Previous research has shown that facial identity is coded by an opponent-coding system; hence, these results also demonstrate that gaze is coded by a different representational system to facial identity.     

Neuroanatomical correlates of category-specific semantic disorders: A critical survey

Current models of verbal short‐term memory (STM) propose various mechanisms for serial order. These include a gradient of activation over items, associations between items, and associations between items and their positions relative to the start or end of a sequence. We compared models using a variant of Hebb's procedure in which immediate serial recall of a sequence improves if the sequence is presented more than once. However, instead of repeating a complete sequence, we repeated different aspects of serial order information common to training lists and a subsequent test list. In Experiment 1, training lists repeated all the item–item pairings in the test list, with or without the position–item pairings in the test list. Substantial learning relative to a control condition was observed only when training lists repeated item–item pairs with position–item pairs, and position was defined relative to the start rather than end of a sequence. Experiment 2 attempted to analyse the basis of this learning effect further by repeating fragments of the test list during training, where fragments consisted of either isolated position–item pairings or clusters of both position–item and item–item pairings. Repetition of sequence fragments led to only weak learning effects. However, where learning was observed it was for specific position–item pairings. We conclude that positional cues play an important role in the coding of serial order in memory but that the information required to learn a sequence goes beyond position–item associations. We suggest that whereas STM for a novel sequence is based on positional cues, learning a sequence involves the development of some additional representation of the sequence as a whole.