How is the serial order of a verbal sequence coded? Some comparisons between models

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.     

Retrieval processes for serial order information

Three experiments sought to specify how list structure and rehearsal pattern influence the retrieval of well-learned serial order information. Subjects learned a serial list of 12 words followed by a probed recall task measuring response time. Adjacent list items served as retrieval cues to permit probing of an item by cues which maintained or crossed semantic or rehearsal boundaries. Evidence for structure in serial recall was inferred from the large cue format effects on response time. Such effects were found to be consistent with the semantic relationships in categorized lists and the acquisition rehearsal pattern in unrelated lists. When rehearsal grouping and semantic relatedness were in conflict. the cue format effects conformed mainly to the rehearsal pattern. Extended practice over five sessions did not eliminate these effects for many of the serial items. These results suggest that the structure of the serial list. whether based on previous associations or present rehearsal patterns. can provide a basis for retrieval. A hierarchical search model based on item and order information provided good fits of the data. The model suggested that response time varies with cue formats because cues differ in their efficiency at directing search to the correct response in the list structure. The structure. which is acquired at the time of learning. determines cue efficiency and. hence. the subsequent effects upon response time.     

KNOWLEDGE OF THE ORDINAL POSITION OF LIST ITEMS IN RHESUS MONKEYS

Abstract —What is learned during mastery of a serial task: associations between adjacent and remote items, associations between an item and its ordinal position, or both? A dear answer to this question is lacking in the literature on human serial memory because it is difficult to control for a "naive" subject's linguistic competence and extensive experience with serial tasks. In this article, we present evidence that rhesus monkeys encode the ordinal positions of items of an arbitrary list when there is no requirement to do so. First, monkeys learned four nonverbal lists (1–4). each containing four novel items (photographs of natural objects). The monkeys then learned four 4-item lists that were derived exclusively and exhaustively from Lists 1 through 4, one item from each list. On two derived lists, each item s original ordinal position was maintained. Those lists were acquired with virtually no errors. The two remaining derived lists, on which the original ordinal position of each item was changed, were as difficult to learn as novel lists. The immediate acquisition of lists on which ordinal position was maintained shows that knowledge of ordinal position can develop without the benefit of language, extensive list-learning experience, or explicit instruction to encode ordinal information.     

Initial codes for order in memory

The roles of positional and associative information in preserving the order of short sequences of items were investigated either by presenting an item in the same serial position on successive trials (position transfer), or by presenting the same pair of items in adjacent but shifted positions on successive trials (association transfer). While Experiment I found that positive transfer of position information occurred after just one repetition of an item in the same serial position, Experiment II found no evidence for association transfer after a single trial. If an item is presented in the same position, or if a pair of items are presented in adjacent positions on several successive trials, position transfer effects built up more rapidly than association transfer effects (Experiments III and IV). However, the rate at which association transfer effects are established depends on the type of item used, with lists of concrete words showing rapid establishment of associations (Experiment V) but lists of digits showing only slow development of associations (Experiment IV). It was concluded that serial order may be preserved both by positional and associative codes, but the establishment of an associative code is linked to the extraction of information about the items themselves.     

Positional cues in serial learning: The spin-list technique

To test the hypothesis that serial learning depends largely on the encoding and retrieval of position-to-item associations, we examined whether people can learn spin lists on which starting position is randomly varied across successive learning trials. By turning positional information from a reliable cue into a source of intertrial interference, we expected learning to be greatly impaired. Contrary to this hypothesis, we found that participants were only slightly worse at serial learning under spin conditions and that this impairment reflects a substantial increase in initiation errors coupled with a small increase in intertrial forgetting. These data show that participants can effectively use nonpositional cues when positional cues are unreliable.     

The effect of repetition on search through active memory

Short lists of digits, some of which contained repeated items, were searched in memory. Repeated items were recognized faster than non-repeated items. Search times for non-repeated items in lists with a repeat were slightly faster than those for lists of the same length in which no items were repeated. A serial position effect was found such that items at either the beginning or end of the memory list were responded to faster than items in the middle of the list. Neither the repeated digit effect nor the serial position effect support serial exhaustive search. An empirical equation was found by which response time is predicted by the log proportion of items in a memory list that match the test stimulus. These results are compatible with the notions of neural pathway activation and trace strength.     

Linking of serially ordered lists by macaque monkeys (Macaca mulatta): list position influences

Five monkeys (Macaca mulatta) were trained on 2 sets of 3 5-item serially ordered lists. Then, each set was either linked or not in a counterbalanced, within-subject design. Linking entailed training on the 2 pairs that ordered the 3 5-item lists into a single overall 15-item series. Choices on novel pairings after linking conditions attempted to define the unique contributions of knowledge of within-list ordinal position and between-lists link training. With linkage, the series was immediately treated as a 15-item ordered list. Without linkage, choices reflected list positions from initial learning, but continued testing with directional reward yielded gradual ordering into a 15-item list. Apparently, monkeys remembered and used initial list-position information, but linkage allowed inference of an integrated serial relationship among items. Results supported primate list memory as an organizational process.     

Serial position phenomena: Training and retrieval effects

With successive free recall lists primacy items are usually among the first to be reported on the initial list. Recency items will then take over and be first reported on later lists. This retrieval shift was studied under varied list conditions designed to counteract ordinary position effects, and proved to be a stable and resistent effect. One experiment had subjects practice on five different free recall lists over three trials, and the results agreed with the hypothesis that primacy reduction is caused by proactive interference rather than by the primacy to recency report shift. Experiments of the usual one-session laboratory type have consistently failed to show practice effects on serial lists with incompatible spatial and temporal order cues. In a case study of five subjects examined over a period of three months only slight improvement on single-trial lists was observed. However, when naive subjects were given four successive trials with the same type of cue-conflict list, prominent practice effects were easily demonstrated. Observations confirmed the assumption that repeated presentations of items located in middle list positions may counteract the privilege of primacy and recency positions. When repetitions were made within contracting and expanding lists, the results proved that active anchoring, making itself visible as primacy effects, is feasible with contracting lists but difficult with expanding lists. Active performance of list learning strategies generally results in primacy effects; whereas passive shortcut procedures, in learning and retrieval of information, produce recency effects. Predominant recency effects are symptoms of difficult task situations only partially mastered by the learner; primacy effects point to more successful elaboration. Overall, serial position effects do not seem to be due to structural memory stores so much as to the working of cognitive strategy factors. A problem-solving theory was presented as an alternative to information-processing models of serial learning and memory.     

Implicit learning of what comes when and where within a sequence: The time-course of acquiring serial position-item and item-item associations to represent serial order

Much research has been conducted aimed at the representations and mechanisms that enable learning of sequential structures. A central debate concerns the question whether item-item associations (i.e., in the sequence A-B-C-D, B comes after A) or associations of item and serial list position (i.e., B is the second item in the list) are used to represent serial order. Previously, we showed that in a variant of the implicit serial reaction time task, the sequence representation contains associations between serial position and item information (Schuck, Gaschler, Keisler, & Frensch, 2011). Here, we applied models and research methods from working memory research to implicit serial learning to replicate and extend our findings. The experiment involved three sessions of sequence learning. Results support the view that participants acquire knowledge about order structure (item-item associations) and about ordinal structure (serial position-item associations). Analyses suggest that only the simultaneous use of the two types of knowledge acquisition can explain learning-related performance increases. Additionally, our results indicate that serial list position information plays a role very early in learning and that inter-item associations increasingly control behavior in later stages.     

Serial list combination by monkeys (Macaca mulatta): test cues and linking

This investigation assessed prospective bases of non-human primate cognitive operations that support serial list memory. Four macaques learned 3-, 5-item ordered lists of objects (as two-choice problems) and then either did or did not (in a within-subject design) receive training on pairs that linked the three original lists into a 15-item serial order. Next, subjects experienced selective exposure trials on object pairs that either maintained or contrasted to the serial position relationships seen during original learning. Subsequent comprehensive tests assessed the interactive effects of linking and exposure conditions on choosing in accord with a combined 15-item serial order. Linking readily induced monkeys to merge lists into a 15-item order, but restricting early exposure to pairs with the same positional relationships as original training slowed, but did not prevent, list combination. Exposure to positional relationships congruent with the combined (15-item) list and different from those of original 5-item training aided both expression of the linking effect and acquisition after no link training. Thus, list linking facilitated serial reorganization by inducing release from error derived from memory for prior learned positional relationships. The task was recommended as a prospective evaluator of continuity of cognitive processes among species.     

Conditional recall and the frequency effect in the serial recall task: an examination of item-to-item associativity

The frequency effect in short-term serial recall is influenced by the composition of lists. In pure lists, a robust advantage in the recall of high-frequency (HF) words is observed, yet in alternating mixed lists, HF and low-frequency (LF) words are recalled equally well. It has been argued that the preexisting associations between all list items determine a single, global level of supportive activation that assists item recall. Preexisting associations between items are assumed to be a function of language co-occurrence; HF?CHF associations are high, LF?CLF associations are low, and mixed associations are intermediate in activation strength. This account, however, is based on results when alternating lists with equal numbers of HF and LF words were used. It is possible that directional association between adjacent list items is responsible for the recall patterns reported. In the present experiment, the recall of three forms of mixed lists??those with equal numbers of HF and LF items and pure lists??was examined to test the extent to which item-to-item associations are present in serial recall. Furthermore, conditional probabilities were used to examine more closely the evidence for a contribution, since correct-in-position scoring may mask recall that is dependent on the recall of prior items. The results suggest that an item-to-item effect is clearly present for early but not late list items, and they implicate an additional factor, perhaps the availability of resources at output, in the recall of late list items.     

Temporal aspects of retrieval in short-term serial retention

In three experiments, temporal properties of memory retrieval are studied. Latency of positionally probed recall of a single item is measured as a function of its serial position (SP) in a serial list. The requested SP is indicated by a positional probe, presented immediately after the list. For two organizational structures of the list (i.e., grouped vs. ungrouped lists) the experiments study the effects of precueing the probe position briefly before probe presentation, the cue signal indicating the SPs relevant for recall. The problem of confounded effects on latency of requested SP and of probe-signal location is also investigated. A bow-shaped SP-curve is found for ungrouped lists; precueing reduces RT but does not affect the bowed shape. This argues against selective confounding effects of spatial probe position and also suggests that precueing preactivates memory access. Grouping the items into two sublists eliminates precueing effects for items around the group boundary. Furthermore, precueing is detrimental when SPs at opposite ends of the list are precued. This suggests that the available points of memory access can only be utilized one at a time, so that precueing is not beneficial when the item cannot be reached via the preactivated point. Although the notion of single access to memory is in accord with both positional cueing theory and non-associative hierarchical theories (e.g., Estes 1972), various details of the results are in favor of the first type of explanation.     

Serial list linking by macaque monkeys (Macaca mulatta): list property limitations

Four sophisticated macaque monkeys (Macaca mulatta) learned 6 different, 15-item ordinal lists (via conditional, 2-choice discriminations) as part of a study assessing some properties of serial list memory in monkeys. After assuring that the first 3 lists were well retained, the researchers attempted to link these by training only the 2 end-item pairs that ordered all 45 items into an inclusive series. A 20-day test of possible pairings among these 45 items was then conducted. Subsequently, the other 3 lists were trained and tested for retention, but no link training was provided. Then, a test like the one that had followed linking was administered. Unlike previous outcomes with three 5-item lists, linking did not yield immediate merger of three 15-item lists into a 45-item list, although 45-item lists were acquired after sufficient exposure to the testing/training regimen under both linking and nonlinking conditions. List length as a limiting factor in linking suggested processing restrictions analogous to those observed in human list memory. Results supported further investigation of list-linking characteristics.     

Testing a positional model of the Hebb effect

In two experiments, we investigated the hypothesis that a strengthening of position-item associations underlies the improvement seen in performance on an immediate serial recall task, for a list that is surreptitiously repeated every third trial. Having established a strong effect of repetition, performance was tested on transfer lists in which half the items held the same position as in the repeated list (S-items) and the remainder moved (D-items). In Experiment 1, S-items showed a small advantage in order errors over control and D-items. A second experiment tested whether a design feature in Experiment 1 underlay this advantage. When the experimental design was better controlled, no improvement was shown for either S- or D-items over controls. These data were shown to be inconsistent with the results of computer simulations of a positional model. An alternative model is outlined.     

Interference processes in monkey auditory list memory

A rhesus monkey’s memory was tested for single items and four-item lists of natural and environmental sounds. Memory items were presented from a center speaker, followed by a retention delay and then a choice response to a test sound presented simultaneously from two side speakers. Recognition of the last item of four-item lists was much poorer than that of single items at 0-, 1-, and 2-sec delays, despite there being the same temporal relations between study and test. This result showed that the first three items proactively interfered with memory of the last list item. Proactive interference dissipated after 2 sec, revealing a recency effect that eventually equaled single-item performance. Recognition of the first item of four-item lists was much poorer than single items at 20- and 30-sec delays, showing that the last three items retroactively interfered with memory of the first list item. The results point to the critical nature of interference processes in the understanding of serial position functions.     

Position effects in comparative judgments of serial order: List structure vs. differential strength

Two experiments are reported that test the hypothesis that the serial position effect in comparative judgment of ordinal position in arbitrary serial lists results from differential memory or associative strength among list items. The serial position effect in comparative judgment is typically a pattern in which pairs that contain a term from one of the two extremes of the list are processed faster and more accurately than pairs that contain no end terms. The experiments show that a new term added to either the end or the middle of a well-practiced fourterm series behaves almost immediately like the end or central term, respectively, of a well-practiced five-term series. Furthermore, when the added term is removed, the list reverts immediately to the position effect obtained in a four-term series. Theories that explain the position effect by differential build-up of item strength or of interitem associative strength over practice cannot explain these effects. We propose instead that learning of a serial list is accomplished by assigning list members to positions in a general-purpose linear order schema and that subjects can make these assignments rapidly and flexibly.     

Generalization of position association in rote serial learning

Summary Thirty-six Ss practiced two serial lists of identical length in haphazard alternation. At each position in a list, a maximum number of intruding responses arrived from the same position in the other list, and the frequency of intrusion from a non-identical position in the other list declined with increase of positional disparity in either direction. The reliability of this gradient-like phenomenon, which could not be verified by the method of trend analysis owing to insufficient data, is partially indicated by the occurrence of interlist intrusions at a significantly higher rate between identical than between non-identical positions. These results are interpreted as an evidence in support of the hypothesis that not only associations develop between the items of a rote series and their serial positions but also these associations generalize between positions in inverse proportion to the intervening distance.     

Associative isolation: Unifying associative and list memory

Rather than treating paired associate and serial learning as involving the acquisition of distinct types of information [e.g. Murdock (1974). Human memory: Theory and data. New York: Wiley] I propose an Isolation Principle which treats the two as ends of a continuum. According to this principle, consecutive pairs of items are relatively isolated from other studied items in paired associates learning, but not isolated in serial list learning. The consequence is that variability that dominates forward and backward probed recall is highly correlated in pairs but less so, due to differential interference, in lists. This can explain an important dissociation: whereas forward and backward probes of pairs are nearly perfectly correlated, the correlation is only moderate for serial lists. I demonstrate this in a chaining model by varying item-to-item associative strengths and in a positional coding model by varying the similarity structure of item positions. This enables a range of models to account for data on pairs and lists, as well as potential intermediate or hybrid paradigms, within a single theoretical framework.     

Short-term memory in autism spectrum disorder

Three experiments examined verbal short-term memory in comparison and autism spectrum disorder (ASD) participants. Experiment 1 involved forward and backward digit recall. Experiment 2 used a standard immediate serial recall task where, contrary to the digit-span task, items (words) were not repeated from list to list. Hence, this task called more heavily on item memory. Experiment 3 tested short-term order memory with an order recognition test: Each word list was repeated with or without the position of 2 adjacent items swapped. The ASD group showed poorer performance in all 3 experiments. Experiments 1 and 2 showed that group differences were due to memory for the order of the items, not to memory for the items themselves. Confirming these findings, the results of Experiment 3 showed that the ASD group had more difficulty detecting a change in the temporal sequence of the items.