In experiments using simple stimuli, the active working memory system is often estimated to have a fixed capacity no matter how long participants are given to.
Active cognitive control of working memory is central in most human memory models, but behavioral evidence for such control in nonhuman primates is absent and neurophysiological evidence, while suggestive, is indirect. We present behavioral evidence that monkey memory for familiar images is under active cognitive control.
Concurrent cognitive demands during the memory delay impaired matching-to-sample performance for familiar images in a demand-dependent manner, indicating that maintaining these images in memory taxed limited cognitive resources. Performance with unfamiliar images was unaffected, dissociating active from passive memory processes. Active cognitive control of memory in monkeys demonstrates that language is unnecessary for active memory maintenance. Experiment 1: Primary findingsWe presented monkeys with visual matching-to-sample recognition tests on touchscreen computers and required them to complete one of three distractor tasks during the memory interval. The three tasks required the same motor response but varied in cognitive demand: 1) touch a blue square that appeared in a randomly-selected corner of the screen (motor only), 2) touch a photograph that appeared in a randomly-selected corner of the screen (motor + image perception), or 3) classify a photograph as depicting a bird, fish, flower, or person by touching the appropriate symbol in one of the four corners of the screen (motor + image perception + classification). Touching a uniform blue square should require the least cognitive processing. Viewing unfamiliar photographs may elicit more cognitive processing than viewing a blue square because the photograph is more visually complex and presumably more interesting.
Finally, classifying photographs should require the most cognitive processing because the monkeys had to accurately assign the images to one of four categories to proceed to the memory test. If remembering required active maintenance of the studied image during the memory interval, accuracy should be impaired least by the motor task and most by the classification task. Passive retention should be unaffected by these manipulations of concurrent cognitive demand. Memory tests with four levels of concurrent cognitive demandMonkeys were required to remember an image over a memory interval that was either empty, or filled by one of three tasks: 1) motor: touch a blue square, 2) image: touch a non-classifiable image, or 3) classify: classify a central image as a bird, fish, flower, or person by touching the corresponding symbol.
Motor and image stimuli could appear in any of the four screen corners. All three concurrent tasks required the same motor response.
2.1.2 StimuliStimuli were color photographs of exemplars from categories the monkeys had previously learned to classify as birds, fish, flowers, or people. The to-be-remembered stimuli consisted of two sets: a small set of four images, highly-familiar from previous testing, and a large set of 1400 relatively-unfamiliar images. Each category was equally represented within each set. For the concurrent task, images for the classify condition were drawn from the large set of 1400 images, and non-classifiable images for the image condition consisted of a set of 400 relatively-novel images. 2.1.3 ProcedureMonkeys completed four 300-trial sessions, two with the small set of familiar images and two with the large set of unfamiliar images, alternating and counterbalanced for testing order across monkeys. Half the trials in each session contained no secondary task; the other half were equally divided among motor, image, and classification tasks. The four levels of cognitive demand, the four categories, and the four possible response locations were intermixed pseudorandomly within each session.
To-be-classified images were never from the same category as the sample and were not presented as distractors for that trial. Trials proceeded as in, separated by a 10-sec ITI. Because matching-to-sample accuracy in monkeys is typically higher with large sets of unfamiliar images than with small sets of familiar images (; ), we matched baseline performance by testing the large set of unfamiliar images at a 30-sec delay and the small set of familiar images at a 1-sec delay (values determined during pre-testing). At test, selection of the sample produced a positive audio stimulus and a 75% chance of food, whereas selection of a distractor produced a negative audio stimulus and a 2-sec timeout. To ensure that monkeys were attending to, and processing, the concurrent task, incorrect responses in the concurrent tasks aborted the trial.
Proportions were arcsine transformed prior to statistical analysis. 2.2 Results and DiscussionThe distraction tasks affected memory performance for the two image sets differently (; two-factor repeated measures ANOVA; interaction: F (3,15) = 57.83, p. Memory performance for familiar but not unfamiliar images is impaired by concurrent cognitive demand in a demand-dependent manner in monkeysProportion correct (±SEM) on the final recognition test is graphed for both the familiar small image set (red dashed line) and the unfamiliar large image set (solid blue line) as a function of the four levels of concurrent cognitive demand imposed during the memory interval.
The gray horizontal dashed line represents the proportion correct expected by chance. Experiments 2a–2c: Alternative explanationsThe results of Experiment 1 suggest that the concurrent tasks impaired performance because holding familiar, but not unfamiliar, images in memory required limited cognitive resources, and the concurrent tasks competed for those resources. Prior to accepting this interpretation, we investigated four alternative explanations. In Experiment 2a, we evaluated whether the decrement occurred because completing the concurrent task lengthened the memory interval. In Experiment 2b, we evaluated whether the decrement was due to the concurrent task occurring immediately after study in the familiar image condition, rather than after a relatively long interval in the unfamiliar image condition. In Experiment 2c, we evaluated whether the decrement occurred only when to-be-remembered samples were classifiable, and also whether the selective decrement was due to the two image sets being tested at different memory delays.
3.1.2 Experiment 2b: Timing of concurrent taskWe ran two 100-trial sessions using the unfamiliar image set, one with the concurrent task at the end of the 30-second delay and one with the concurrent task at the start of the 30-second delay. If the decrement in accuracy with the small set of familiar items was due to the secondary task following quickly after the sample, then moving the secondary task to the beginning of the delay with the large set of unfamiliar items should produce a similar decrement to that found with the small image set. 3.1.3 Experiment 2c: Image content and constant memory intervalWe ran two 100-trial sessions at a consistent memory interval of 4 seconds. The to-be- remembered images for the two sessions were the relatively unfamiliar set of 400 non- classifiable images used in the image condition of Experiment 1 and the set of 4 highly-familiar non-classifiable images, respectively. If the classification task produced a large decrement because the samples were classifiable, then using samples that the monkeys were unable to classify should eliminate the effect. Additionally, if the difference between the two sets was due to them being tested at different memory delays, then testing them at the same memory delay should eliminate the effect. 3.2.1 Experiment 2a: Lengthened memory intervalThe performance impairment seen with the familiar images in Experiment 1 was not due to elongation of the memory interval by the addition of time spent completing the concurrent tasks.
On average, the concurrent tasks increased the memory interval of the familiar images from 1s to 2.5s in Experiment 1; however, memory performance following the unfilled 4s delay in Experiment 2a was significantly higher than the filled 2.5s delay from the Experiment 1 (mean proportion correct at 4s =.79, t 5 = 8.47, p. 3.2.2 Experiment 2b: Timing of concurrent taskThe lack of performance impairment with the unfamiliar images in Experiment 1 was not due to the concurrent task following more quickly after the study phase with the familiar images than with the unfamiliar images in Experiment 1. In Experiment 2b, memory performance with the unfamiliar images was equivalent when the classification task occurred at the end and at the beginning of the 30s delay (mean proportion correct: end of delay = 0.65, beginning of delay = 0.64; t 5 = 0.39, p =.709). 3.2.3 Experiment 2c: Image content and constant memory intervalThe selective performance impairment seen in Experiment 1 was not due to the to-be-remembered images being classifiable by the monkeys. With non-classifiable images, we again observed selective impairment for the familiar stimuli but not the unfamiliar stimuli (non-classifiable familiar images mean proportion correct: none =.86, concurrent classification task =.51, t 5 = 13.08, p.
General DiscussionConcurrent cognitive demands during the memory delay impaired performance for familiar, but not unfamiliar, images in a demand-dependent manner. This indicates that remembering familiar information is cognitively effortful for monkeys. This establishes a strong parallel with human working memory. It also raises the intriguing possibility that monkeys hold familiar images in working memory via an effortful maintenance process akin to human rehearsal (but see ).
Primacy, or superior memory for items appearing early in a list, is often due to rehearsal in humans. We recently found that memory performance for lists of familiar, but not unfamiliar, images showed a primacy effect in monkeys , again suggesting a rehearsal-like process for familiar information (but see ). This difference between processing of familiar and unfamiliar memoranda parallels fMRI results from humans showing that the prefrontal cortex is more active when remembering familiar images.The discrepancy of these results, which provide evidence of active maintenance of monkey memory, with previous results, which found no evidence of active maintenance (; ), may be due to the relatively high familiarity of the images being remembered. In these previous studies, samples were either drawn from a medium-sized set of 32 photographs , or an unbounded set of algorithmically-generated grid patterns. Because target stimuli did not repeat every trial, it is possible that they could be discriminated from distractors at test on the basis of familiarity, and thus monkeys could perform accurately without needing to maintain them in working memory. Although set size appears to be a likely factor considering the current findings, there are too many differences between the current study and the previous ones to draw a firm conclusion without additional experiments.It is a challenge to select appropriate language to accurately describe cognitive processes in nonhumans.
Passive familiarity describes well the immunity to interference we saw in recognition of targets from the large set of unfamiliar images. With the large set, the target had been seen much more recently than the distractors and thus was presumably more familiar, memory for the target was unaffected by concurrent cognitive demands and thus primarily passive, and studies in humans have shown that familiarity judgments are primarily passive (; ). However, one could also describe this as a recency judgment or as a novelty judgment. We cannot distinguish between these descriptions in the current study, and it is not immediately clear whether these are different ways of describing the same type of judgment or different types of judgments.
Future studies may help illuminate these distinctions.Our results indicate that future studies of working memory in nonhumans should contrast performance with familiar and unfamiliar images. Because of the relative ease with which monkeys learn memory tasks with large sets of unfamiliar images , large sets have become the standard in primate memory research; however, the present results show that large and small sets are remembered differently. Failure to recognize this difference may have created the perception of inconsistencies between the nonhuman and human literatures, in which humans are often tested with familiar items and monkeys are often tested with relatively unfamiliar items. Secondary tasks that manipulate concurrent cognitive load can be used to identify instances of active working memory and may help resolve these apparent inconsistencies.
Neurophysiological studies of working memory that contrast performance with large and small stimulus sets or that use other methods to contrast passive familiarity and active maintenance will prove especially informative.Humans often maintain information in working memory through verbal rehearsal, but our results with monkeys indicate that active memory maintenance does not require language. There is evidence that humans engage in nonverbal memory maintenance , but it is difficult to block the human tendency to name visual stimuli, and recoding unfamiliar visual stimuli into familiar words does facilitate memory.
Based on these findings and ours, one intriguing possibility is that the capacity for active control of memory may have more to do with familiarity than with other properties of linguistic material. The ease with which humans recode unfamiliar memoranda into familiar words, an option unavailable to monkeys, may be one of the reasons that cognitive control over memory is more robust in humans than it is in monkeys.
The authors declare no competing financial interests.Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Aron A, Aron E. Statistics for psychology. Upper Saddle River, NJ: Prentice Hall; 1999.
Baddeley A. Working memory: Looking back and looking forward. Nature Reviews Neuroscience. 2003; 4(10):829–839. Basile BM, Hampton RR. Rhesus monkeys ( Macaca mulatta) show robust primacy and recency in memory for lists from small, but not large, image sets.
Behavioural Processes. 2010; 83(2):183–190.
Basile BM, Hampton RR. Monkeys show recognition without priming in a classification task. Behavioural Processes (in press). Bisby JA, Leitz JR, Morgan CJA, Curran HV. Decreases in recollective experience following acute alcohol: a dose-response study.
2010; 208(1):67–74. Constantinidis C, Franowicz MN, Goldman-Rakic PS. The sensory nature of mnemonic representation in the primate prefrontal cortex. Nature Neuroscience. 2001; 4(3):311–316. Cook RG, Wright AA, Sands SF. Interstimulus-Interval and Viewing Time Effects in Monkey List Memory.
Animal Learning & Behavior. 1991; 19(2):153–163. What are the differences between long-term, short-term, and working memory? In: Sossin WS, Lacaille JC, Castellucci VF, Belleville S, editors. Essence of Memory. Amsterdam: Elsevier Science Bv; 2008.
D’esposito M, Postle BR, Ballard D, Lease J. Maintenance versus manipulation of information held in working memory: An event-related fMRI study. Brain and Cognition. 1999; 41(1):66–86. Eacott MJ, Gaffan D, Murray EA.
Preserved Recognition Memory for Small Sets, and Impaired Stimulus Identification for Large Sets, Following Rhinal Cortex Ablations in Monkeys. European Journal of Neuroscience. 1994; 6(9):1466–1478.
Elmore LC, Ma WJ, Magnotti JF, Leising KJ, Passaro AD, Katz JS, Wright AA. Visual Short-Term Memory Compared in Rhesus Monkeys and Humans. Current Biology. 2011; 21(11):975–979. Fuster JM, Alexander GE. Neuron Activity Related to Short-term Memory.
1971; 173(3997):652. Hannula DE, Tranel D, Cohen NJ. The long and the short of it: Relational memory impairments in amnesia, even at short lags.
Journal of Neuroscience. 2006; 26(32):8352–8359. Heuer E, Bachevalier J. Neonatal Hippocampal Lesions in Rhesus Macaques Alter the Monitoring, but Not Maintenance, of Information in Working Memory. Behavioral Neuroscience. 2011; 125(6):859–870.
Heyselaar E, Johnston K, Pare M. A change detection approach to study visual working memory of the macaque monkey.
Journal of Vision. 2011; 11(3). Hourihan KL, Ozubko JD, Macleod CM. Directed forgetting of visual symbols: Evidence for nonverbal selective rehearsal. Memory & Cognition.
2009; 37(8):1059–1068. Jacoby LL.
A Process Dissociation Framework - Separating Automatic from Intentional Uses of Memory. Journal of Memory and Language. 1991; 30(5):513–541. Jeneson A, Mauldin KN, Hopkins RO, Squire LR. The role of the hippocampus in retaining relational information across short delays: The importance of memory load. Learning & Memory.
2011; 18(5):301–305. Jeneson A, Squire LR.
Working memory, long-term memory, and medial temporal lobe function. Learning & Memory. 2012; 19(1):15–25. Logie RH.
Visuospatial Processing in Working Memory. Quarterly Journal of Experimental Psychology Section a-Human Experimental Psychology. 1986; 38(2):229–247. Marshall PH, Werder PR. Effects of Elimination of Rehearsal on Primacy and Recency.
Journal of Verbal Learning and Verbal Behavior. 1972; 11(5):649. Meyer T, Qi XL, Constantinidis C.
Persistent discharges in the prefrontal cortex of monkeys naive to working memory tasks. Cerebral Cortex. 2007; 17:I70–I76. Miller EK, Desimone R. Scopolamine Affects Short-Term-Memory but not Inferior Temporal Neurons.
1993; 4(1):81–84. Miller EK, Erickson CA, Desimone R. Neural mechanisms of visual working memory in prefrontal cortex of the macaque.
Journal of Neuroscience. 1996; 16(16):5154–5167.
Milner B. Memory and the medial temporal regions of the brain. In: Pribram KH, Broadbent DE, editors. Biology of Memory. New York: Academic Press; 1970. Mishkin M, Delacour J.
Analysis of Short-term Visual Memory in Monkeys. Journal of Experimental Psychology-Animal Behavior Processes. 1975; 1(4):326–334. Phillips WA, Christie DFM. Interference with Visualization. Quarterly Journal of Experimental Psychology.
1977; 29(V):637–650. Prendergast MA, Jackson WJ, Terry AV, Kille NJ, Arneric SP, Decker MW, Buccafusco JJ. Age-related differences in distractibility and response to methylphenidate in monkeys. Cerebral Cortex. 1998; 8(2):164–172. Shettleworth SJ.
Cognition, Evolution, and Behavior. New York: Oxford University Press; 1998. Stern CE, Sherman SJ, Kirchhoff BA, Hasselmo ME. Medial temporal and prefrontal contributions to working memory tasks with novel and familiar stimuli. 2001; 11(4):337–346. Unsworth N, Engle RW.
The nature of individual differences in working memory capacity: Active maintenance in primary memory and controlled search from secondary memory. Psychological Review. 2007; 114(1):104–132. Uttal WR. The new phrenology: the limits of localizing cognitive processes in the brain.
Cambridge, Mass: MIT Press; 2001. Washburn DA, Astur RS.
Nonverbal working memory of humans and monkeys: Rehearsal in the sketchpad? Memory & Cognition. 1998; 26(2):277–286. Wright AA, Cook RG, Rivera JJ, Shyan MR, Neiworth JJ, Jitsumori M. Naming, Rehearsal, and Interstimulus-Interval Effects in Memory Processing.
Journal of Experimental Psychology-Learning Memory and Cognition. 1990; 16(6):1043–1059. Wynn T, Coolidge FL. The expert neandertal mind.
Journal of Human Evolution. 2004; 46(4):467–487. Yonelinas AP. The nature of recollection and familiarity: A review of 30 years of research. Journal of Memory and Language. 2002; 46(3):441–517.
Yonelinas AP, Jacoby LL. Dissociations of processes in recognition memory - Effects of interference and of response speed. Canadian Journal of Experimental Psychology-Revue Canadienne De Psychologie Experimentale.
1994; 48(4):516–535.
. Working MemoryWorking MemoryBy, updated 2012Atkinson’s and Shiffrin’s (1968) was extremely successful in terms of the amount of research it generated. However, as a result of this research, it became apparent that there were a number of problems with their ideas concerning the characteristics of short-term memory.Baddeley and Hitch (1974) argue that the picture of short-term memory (STM) provided by the Multi-Store Model is far too simple.According to the, STM holds limited amounts of information for short periods of time with relatively little processing. It is a unitary system. This means it is a single system (or store) without any subsystems. Working Memory is not a unitary store.Fig 1. The Working Memory Model (Baddeley and Hitch, 1974)Working memory is.
However, instead of all information going into one single store, there are different systems for different types of information. Part of working memory that deals with spoken and written material.
It can be used to remember a phone number. It consists of two parts. Phonological Store (inner ear) – Linked to speech perception.
Holds information in a speech-based form (i.e., spoken words) for 1-2 seconds. Articulatory control process (inner voice) – Linked to speech production. Used to rehearse and store verbal information from the phonological store.Fig 2.
The Working Memory Model Components (Baddeley and Hitch, 1974)The labels given to the components (see fig 2) of the working memory reflect their function and the type of information they process and manipulate. The phonological loop is assumed to be responsible for the manipulation of speech based information, whereas the visuospatial sketchpad is assumed to be responsible for manipulating visual images. The central executive decides which information is attended to and which parts of the working memory to send that information to be dealt with.
For example, two activities sometimes come into conflict, such as driving a car and talking. Rather than hitting a cyclist who is wobbling all over the road, it is preferable to stop talking and concentrate on driving.
The central executive directs attention and gives priority to particular activities.The central executive is the most versatile and important component of the working memory system. However, despite its importance in the working-memory model, we know considerably less about this component than the two subsystems it controls.Baddeley suggests that the central executive acts more like a system which controls attentional processes rather than as a memory store. This is unlike the phonological loop and the visuospatial sketchpad, which are specialized storage systems. The central executive enables the working memory system to selectively attend to some stimuli and ignore others.Baddeley (1986) uses the metaphor of a company boss to describe the way in which the central executive operates. The company boss makes decisions about which issues deserve attention and which should be ignored. They also select strategies for dealing with problems, but like any person in the company, the boss can only do a limited number of things at the same time. The boss of a company will collect information from a number of different sources.If we continue applying this metaphor, then we can see the central executive in working memory integrating (i.e., combining) information from two assistants (the phonological loop and the visuospatial sketchpad) and also drawing on information held in a large database (long-term memory).The Phonological Loop.
The phonological loopThe articulatory control process (linked to speech production) acts like an inner voice rehearsing information from the phonological store. It circulates information round and round like a tape loop. This is how we remember a telephone number we have just heard. As long as we keep repeating it, we can retain the information in working memory.The articulatory control process also converts written material into an articulatory code and transfers it to the phonological store.The Visuospatial Sketchpadthe visuospatial sketchpad ( inner eye) deals with visual and spatial information. Visual information refers to what things look like. It is likely that the visuospatial sketchpad plays an important role in helping us keep track of where we are in relation to other objects as we move through our environment (Baddeley, 1997).As we move around, our position in relation to objects is constantly changing and it is important that we can update this information. For example, being aware of where we are in relation to desks, chairs and tables when we are walking around a classroom means that we don't bump into things too often!The sketchpad also displays and manipulates visual and spatial information held in long-term memory.
For example, the spatial layout of your house is held in LTM. Try answering this question: How many windows are there in the front of your house? You probably find yourself picturing the front of your house and counting the windows.
An image has been retrieved from LTM and pictured on the sketchpad.Evidence suggests that working memory uses two different systems for dealing with visual and verbal information. A visual processing task and a verbal processing task can be performed at the same time. It is more difficult to perform two visual tasks at the same time because they interfere with each other and performance is reduced. The same applies to performing two verbal tasks at the same time.
This supports the view that the phonological loop and the sketchpad are separate systems within working memory.Empirical Evidence for WMWhat evidence is there that working memory exists, that it is made up of a number of parts, that it performs a number of different tasks?The working memory model makes the following two predictions:1. If two tasks make use of the same component (of working memory), they cannot be performed successfully together.2. Critical EvaluationStrengthsResearchers today generally agree that short-term memory is made up of a number of components or subsystems. The working memory model has replaced the idea of a unitary (one part) STM as suggested by the multistore model.The working memory model explains a lot more than the multistore model.
It makes sense of a range of tasks - verbal reasoning, comprehension, reading, problem-solving and visual and spatial processing. And the model is supported by considerable experimental evidence.The working memory applies to real-life tasks:- reading (phonological loop)- problem solving (central executive)- navigation (visual and spatial processing)The KF Case Study supports the Working Memory Model. KF suffered brain damage from a motorcycle accident that damaged his short-term memory. KF's impairment was mainly for verbal information - his memory for visual information was largely unaffected. This shows that there are separate STM components for visual information (VSS) and verbal information (phonological loop).Working memory is supported by dual-task studies (Baddeley and Hitch, 1976).The working memory model does not over emphasize the importance of rehearsal for STM retention, in contrast to the multi-store model.
WeaknessesLieberman (1980) criticizes the working memory model as the visuospatial sketchpad (VSS) implies that all spatial information was first visual (they are linked).However, Lieberman points out that blind people have excellent spatial awareness, although they have never had any visual information. Lieberman argues that the VSS should be separated into two different components: one for visual information and one for spatial.There is little direct evidence for how the central executive works and what it does. The capacity of the central executive has never been measured.Working memory only involves STM, so it is not a comprehensive model of memory (as it does not include SM or LTM).The working memory model does not explain changes in processing ability that occur as the result of practice or time.ReferencesAtkinson, R.
C., & Shiffrin, R. Chapter: Human memory: A proposed system and its control processes. In Spence, K. W., & Spence, J. The psychology of learning and motivation (Volume 2).
New York: Academic Press. 89–195.Baddeley, A. Working memory. Oxford: Oxford University Press.Baddeley, A. The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4, (11): 417-423.Baddeley, A.
D., & Hitch, G. Working memory. Bower (Ed.), The psychology of learning and motivation: Advances in research and theory (Vol. New York: Academic Press.Baddeley, A. D., & Lieberman, K.
Spatial working memory. Attention and Performance, VIII.
Hillsdale, N): Erlbaum.How to reference this article:McLeod, S. Working memory. Retrieved from https://www.simplypsychology.org/working%20memory.html.