Primacy versus recency in a quantitative model: activity is the critical distinction

Behavioral and neurobiological evidence shows that primacy and recency are subserved by memory systems for intermediate- and short-term memory, respectively. A widely accepted explanation of recency is that in short-term memory, new learning overwrites old learning. Primacy is not as well understood, but many hypotheses contend that initial items are better encoded into long-term memory because they have had more opportunity to be rehearsed. A simple, biologically motivated neural network model supports an alternative hypothesis of the distinct processing requirements for primacy and recency given single-trial learning without rehearsal. Simulations of the model exhibit either primacy or recency, but not both simultaneously. The incompatibility of primacy and recency clarifies possible reasons for two neurologically distinct systems. Inhibition, and its control of activity, determines those list items that are acquired and retained. Activity levels that are too low do not provide sufficient connections for learning to occur, while higher activity diminishes capacity. High recurrent inhibition, and progressively diminishing activity, allows acquisition and retention of early items, while later items are never acquired. Conversely, low recurrent inhibition, and the resulting high activity, allows continuous acquisition such that acquisition of later items eventually interferes with the retention of early items.     

Serial position effects in recognition memory for odors: A reexamination

Seven experiments examined recognition memory for sequentially presented odors. Following Reed (2000), participants were presented with a sequence of odors and then required to identify an odor from the sequence in a test probe comprising 2 odors. The pattern of results obtained by Reed (2000, although statistically marginal) demonstrated enhanced recognition for odors presented at the start (primacy) and end (recency) of the sequence: a result that we failed to replicate in any of the experiments reported here. Experiments 1 and 3 were designed to replicate Reed (2000), employing five-item and seven-item sequences, respectively, and each demonstrated significant recency, with evidence of primacy in Experiment 3 only. Experiment 2 replicated Experiment 1, with reduced interstimulus intervals, and produced a null effect of serial position. The ease with which the odors could be verbally labeled was manipulated in Experiments 4 and 5. Nameable odors produced a null effect of serial position (Experiment 4), and hard-to-name odors produced a pronounced recency effect (Experiment 5); nevertheless, overall rates of recognition were remarkably similar for the two experiments at around 70%. Articulatory suppression reduced recognition accuracy (Experiment 6), but recency was again present in the absence of primacy. Odor recognition performance was immune to the effects of an interleaved odor (Experiment 7), and, again, both primacy and recency effects were absent. There was no evidence of olfactory fatigue: Recognition accuracy improved across trials (Experiment 1). It is argued that the results of the experiments reported here are generally consistent with that body of work employing hard-to-name visual stimuli, where recency is obtained in the absence of primacy when the retention interval is short.     

Auditory List Memory in Rhesus Monkeys

Auditory memory of 2 rhesus monkeys was tested in a serial probe recognition task. Lists of four environmental or natural sounds were followed by a retention interval and a test. The test matched one of the list items on half of the trials. The retention interval was varied across sessions. Six experiments showed similar results and changes in the serial position function. At short retention intervals, there was good memory for first list items (primacy effect) and poor memory for last list items. At intermediate retention intervals, memory improved for last list items (recency effect). At long retention intervals (20 s and 30 s), the recency effect was strong, and the primacy effect had dissipated. These auditory primacy and recency effects and their changes with retention interval were opposite to those for visual memory. Implications for processes and mechanisms of memory are discussed.     

Rats' Memory for Serially Presented Novel Flavours: Evidence for Non-spatial Primacy Effects

Four experiments examined the effects of serially presenting a number of novel flavours to rats on their subsequent consumption of those flavours. In Experiments 1-4, rats were orally infused with 0.5 ml of flavour over 30 sec for each of five flavours in the exposure phase of the experiment. In these studies, primacy and recency effects emerged, the size of the primacy effect being related to the length of the retention interval, which varied from zero to twenty-four hours. Thus, both primacy and recency effects can be generated using nonspatial stimuli with rats.     

Primacy effects induced by temporal or physical context shifts are attenuated by a preshift test trial

When subjects are sequentially trained with a cue (A) paired separately with two outcomes (B and C) in different phases (i.e., A-B pairings followed by A-C pairings) testing in the training context after short retention intervals often reveals recency effects (i.e., stronger influence by A-C). In contrast, testing after long retention intervals or testing in a context different from that of training sometimes reveals primacy effects (A-B). Three experiments were conducted using rats in a Pavlovian conditioned bar-press suppression preparation to ascertain whether a nonreinforced test trial in the training context soon after training can attenuate this shift to primacy. Experiment 1 demonstrated that exposure to A shortly after both phases of training, but prior to a long retention interval, can attenuate shifts from recency to primacy otherwise observed with a long retention interval. Experiment 2 showed that exposure to A in the training context can also eliminate the shift from recency to primacy otherwise produced by shifting the physical context between training and test. Experiment 3 discredited a potential account of the results of Experiments 1 and 2. The effects observed in Experiment 1 and 2 are interpreted as early testing in the training context serving to initiate rehearsal of the A-C association due to the temporal proximity of A-C training.     

Primacy effects induced by temporal or physical context shifts are attenuated by a preshift test trial

When subjects are sequentially trained with a cue (A) paired separately with two outcomes (B and C) in different phases (i.e., A–B pairings followed by A–C pairings) testing in the training context after short retention intervals often reveals recency effects (i.e., stronger influence by A–C). In contrast, testing after long retention intervals or testing in a context different from that of training sometimes reveals primacy effects (A–B). Three experiments were conducted using rats in a Pavlovian conditioned bar-press suppression preparation to ascertain whether a nonreinforced test trial in the training context soon after training can attenuate this shift to primacy. Experiment 1 demonstrated that exposure to A shortly after both phases of training, but prior to a long retention interval, can attenuate shifts from recency to primacy otherwise observed with a long retention interval. Experiment 2 showed that exposure to A in the training context can also eliminate the shift from recency to primacy otherwise produced by shifting the physical context between training and test. Experiment 3 discredited a potential account of the results of Experiments 1 and 2. The effects observed in Experiment 1 and 2 are interpreted as early testing in the training context serving to initiate rehearsal of the A–C association due to the temporal proximity of A–C training.     

Auditory list memory and interference processes in monkeys.

Memory of 2 rhesus monkeys (Macaca mulatta) was tested in a serial probe recognition task with lists of 4 natural or environmental sounds, different retention intervals, and different manipulations of interference. At short retention intervals, increasing the separation of list items reduced the primacy effect and produced a recency effect. Similar results were shown by increasing interference across lists through item repetitions or making the first 2 list items high-interference items. These results indicated that decreasing first-item performance reduced proactive interference on memory of the last list items. At long (20 s) retention intervals, making the last list items of high interference reduced the recency effect, reduced retroactive interference, and produced a primacy effect. Taken together, interference plays a role in determining the primacy and recency effects of the serial-position function.     

Visual list memory in capuchin monkeys (Cebus apella)

Memory of 3 capuchin monkeys, Cebus apella, was tested with lists of 4 travel-slide pictures and different retention intervals. They touched different areas of a video monitor to indicate whether a test picture was in a list. At short retention intervals (0 s, 1 s, 2 s), memory was good for the last list items (recency effect). At a 10-s retention interval, memory improved for 1st list items (primacy effect). At long retention intervals (20 s and 30 s), primacy effects were strong and recency effects had dissipated. The pattern of retention-interval changes was similar to rhesus monkeys, humans, and pigeons. The time course of recency dissipation was similar to rhesus monkeys. The capuchin's superior tool-use ability was discussed in relation to whether it reflects a superior general cognitive ability, such as memory. In terms of visual memory, capuchin monkeys were not shown to be superior to rhesus monkeys.     

Outcome pre- and postexposure effects: retention interval interacts with primacy and recency

Effects of outcome-alone pretraining and posttraining exposure were investigated in conditioned suppression experiments conducted within a sensory preconditioning preparation with rats. Experiment 1 found that interference by outcome postexposure was stronger than that by outcome preexposure, suggesting a recency effect. Experiment 2 found that after a long retention interval, outcome preexposure produced more interference than outcome postexposure, suggesting a shift from recency to primacy with increasing retention interval. Experiment 3 showed that presentation of a priming stimulus that had been embedded within the earlier phase of treatment also caused a shift from recency to primacy. These results suggest that, at least in a sensory preconditioning paradigm, retrievability of outcome-alone exposure memory is an important determinant of any outcome-alone exposure effect.     

Recency judgments as a function of word frequency: A framing effect and frequency misattributions

The effects of word frequency on judgments of recency of item presentation were examined in two experiments. Subjects in Experiment 1 were presented two mixed lists of high- and low-frequency words followed by a list assignment task for recognized items. It was found that subjects were biased toward assigning low-frequency words to the more recently presented list. Subjects in Experiment 2 were presented a single mixed list of high- and low-frequency words followed by either a relative recency of presentation judgment task or a relative primacy of presentation judgment task. Each word pair on the tests contained one high-frequency word and one low-frequency word. It was found that, for the recency judgment task, subjects were biased to select the low-frequency item as having been presented more recently. However, on the parallel primacy judgment task, there were no effects of word frequency; moreover, overall accuracy levels were higher with primacy than with recency instructions. We interpret the effects of word frequency on recency judgments in Experiments 1 and 2 in terms of a misattribution of frequency-related differences in recollection-based recognition. The finding that recency and primacy instructions produced different patterns of results provides further evidence (Flexser & Bower, 1974) for an effect on performance of the way in which the temporal judgment task was framed.     

Primacy and recency effects in nonhuman primates.

The reports of primacy and recency memory effects in nonhuman primates have been criticized because they have all used an initiating response. That is, the presentation of the to-be-remembered list of items was always contingent on a response being initiated by the nonhuman primate. It has been argued that this initiating response improves performance for early items in the list, resulting in the occurrence of the primacy effect, independent of any memory processing mechanism. This criticism was addressed in the present study by not using an initiating response prior to the presentation of the list. Nevertheless, both a primacy and a recency effect were observed in all 6 rhesus monkeys evaluated using a serial probe recognition task. Thus, the results are similar to those for humans, in that both primacy and recency effects can be obtained in nonhuman primates. A brief literature review is included, and it is proposed that the primacy and recency effects observed in humans, nonhuman primates, and infraprimates can be explained within the context of the configural-association theory.     

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.     

Trial order and retention interval in human predictive judgment

The influences of order of trial type and retention interval on human predictive judgments were assessed for a cue that was reinforced on half of its training presentations. Subjects observed 10 cue-outcome presentations (i.e., reinforced trials) and 10 cue-alone presentations (i.e., nonreinforced trials) in one of three different orders: all nonreinforced trials followed by all reinforced trials(latent inhibition), reinforced and nonreinforced trials interspersed (partial reinforcement), or al lreinforced trials followed by all nonreinforced trials (extinction). Ratings were based mainly on the most recent event type (i.e., a recency effect) when the test occurred immediately after training but were based mainly on initial event types (i.e., a primacy effect) when the test occurred after a 48-h delay. The subjects tested both immediately and with a long retention interval did not exhibit this shift to primacy (i.e., the recency effect persisted). These results demonstrate noncatastrophic forgetting and the flexible use of trial order information in predictive judgments.     

A recency-based account of the primacy effect in free recall

Seven experiments investigated the role of rehearsal in free recall to determine whether accounts of recency effects based on the ratio rule could be extended to provide an account of primacy effects based on the number, distribution, and recency of the rehearsals of the study items. Primacy items were rehearsed more often and further toward the end of the list than middle items, particularly with a slow presentation rate (Experiment 1) and with high-frequency words (Experiment 2). Recency, but not primacy, was reduced by a filled delay (Experiment 3), although significant recency survived a filled retention interval when a fixed-rehearsal strategy was used (Experiment 4). Experimenter-presented schedules of rehearsals resulted in similar serial position curves to those observed with participant-generated rehearsals (Experiment 5) and were used to confirm the main findings in Experiments 6 and 7.     

Amnesia,rehearsal, and temporal distinctiveness models of recall

Classical amnesia involves selective memory impairment for temporally distant items in free recall (impaired primacy) together with relative preservation of memory for recency items. This abnormal serial position curve is traditionally taken as evidence for a distinction between different memory processes, with amnesia being associated with selectively impaired long-term memory. However recent accounts of normal serial position curves have emphasized the importance of rehearsal processes in giving rise to primacy effects and have suggested that a single temporal distinctiveness mechanism can account for both primacy and recency effects when rehearsal is considered. Here we explore the pattern of strategic rehearsal in a patient with very severe amnesia. When the patient’s rehearsal pattern is taken into account, a temporal distinctiveness model can account for the serial position curve in both amnesic and control free recall. The results are taken as consistent with temporal distinctiveness models of free recall, and they motivate an emphasis on rehearsal patterns in understanding amnesic deficits in free recall.     

Auditory and visual serial position functions obey different laws

Rhesus monkeys were tested in serial probe recognition tasks with either travel slide pictures or natural sounds. Tests with four-item lists produced serial position functions that were essentially opposite in shape for the two modalities and changed in opposite ways with retention interval. For visual memory, the primacy effect grew and the recency effect dissipated with retention interval. Capuchin monkeys, humans, and pigeons showed similar results. For auditory memory with rhesus monkeys, the recency effect grew and the primacy effect dissipated with retention interval. These results taken together, along with results from rehearsal tests of monkeys and humans, implicate two passive memory processes with different time courses. Interference among items within auditory lists was manipulated by varying the time between items and categories of items. Interference across lists was manipulated by varying the item pool size and, hence, item repetitions. Changes in the auditory serial position functions indicated that proactive and retroactive interference may have been instrumental in these dynamically changing serial position functions. Implications for theories and models of memory are discussed.     

Fluid and crystallized intelligence and primacy/recency components of short-term memory

Primacy and recency recall measures, obtained from both free and probed-recall tasks, were included in a battery of tests which also contained markers for fluid intelligence, crystallized intelligence, and cognitive speed factors. Our results confirmed previous findings that for subjects within the normal range of abilities, recency recall from probed-recall tasks is more closely linked to intelligence than is primacy recall. A new finding was also obtained: that primacy recall is correlated more with measures of cognitive speed than is recency recall. These results are discussed in terms of the concept of working memory and its relation to intelligence.     

Memory for lists of spatial events in the rat

Three experiments are reported in which rats were presented with lists of spatial events or responses in distinct visual contexts, and retention subsequently was measured for the events occuring at each serial position. In Experiments 1 and 2, lists contained different numbers of forced choices on successively presented T-mazes, and memory for each serial position was tested on each trial. In Experiment 3, lists were made up of different numbers of forced entries into randomly ordered arms on the eight-arm radial maze, and memory for one serial position was probed on each trial. All three experiments yielded serial position curves which demonstrated recency effects but no primacy effects. Comparisons of retention for different list lengths within each experiment indicated that retention of shorter lists was equivalent to retention of the final items of longer lists and superior to retention of the initial items on longer lists. This pattern of results indicated that there was no evidence of proactive inhibition in the serial position curves. Analysis of input and output interference effects and the results of control tests suggested that both retroactive inhibition and time-based forgetting processes were responsible for recency effects.     

Serial position curves in spatial memory of rats: Primacy and recency effects

Memory for lists of items was tested in rats (N = 18) in an 8-arm radial maze. In Experiment 1 trials consisted of a study phase, in which the rat could freely choose five arms to obtain a food reward, and a test phase in which the animal was presented with a choice between a novel and a previously visited arm. The rat received additional food reinforcement only when visiting the novel arm. The two phases of a trial were separated by a retention interval of 30 sec or of 4, 16 or 60 min. It was found that recall of the five free arm choices was related to the serial position of the previously visited arm. There was a significant recency effect at the 30-sec delay. With longer retention intervals this disappeared, and a significant primacy effect could be observed. In Experiment 2 the same animals were given forced arm entries during the study phase and delays of 30 sec or 4 or 16 min before the test phase. Again, there was a trend towards a recency effect after the shorter delays and a significant primacy effect after the 16-min interval. These results show that, in the recall of lists of spatial items, rats have serial position curves with primacy and recency effects, depending on the length of the retention interval.     

Comparisons of memory for nonverbal auditory and visual sequential stimuli

Properties of auditory and visual sensory memory were compared by examining subjects' recognition performance of randomly generated binary auditory sequential frequency patterns and binary visual sequential color patterns within a forced-choice paradigm. Experiment 1 demonstrated serial-position effects in auditory and visual modalities consisting of both primacy and recency effects. Experiment 2 found that retention of auditory and visual information was remarkably similar when assessed across a 10 s interval. Experiments 3 and 4, taken together, showed that the recency effect in sensory memory is affected more by the type of response required (recognition vs. reproduction) than by the sensory modality employed. These studies suggest that auditory and visual sensory memory stores for nonverbal stimuli share similar properties with respect to serial-position effects and persistence over time.