doi:10.1016/j.tics.2005.06.009

Doubts about double dissociations

between short- and long-term memory

Charan Ranganath and Robert S. Blumenfeld

Center for Neuroscience and Department of Psychology, University of California at Davis, 1544 Newton Ct, Davis, CA 95616, USA

Historically, psychologists and neuroscientists have

distinguished between processes supporting memory

for events across retention delays of several seconds

(short-term memory, STM), and those supporting

memory for events across longer retention delays of

minutes or more (long-term memory, LTM). Dis-

sociations reported in some neuropsychological studies

have contributed to a popular view that there must be

neurally distinct memory stores that differentially

support STM and LTM. In this article, we review

evidence from recent studies regarding dissociations

between STM and LTM. We suggest that the evidence

reveals problems with claims of selective STM or LTM

impairments, which in turn questions whether theories

of memory need to propose neurally distinct stores for

short- and long-term retention. We consider alternative

ways to explain the neural mechanisms of memory

across different retention intervals.

Introduction

In 1957, Scoville and Milner published a case report on

‘H.M.’, a patient who underwent a bilateral medial

temporal lobe resection for the treatment of intractable

epilepsy [1]. Research on H.M. revolutionized the study of

memory for two reasons: First, following the surgery, H.M.

was virtually unable to form new memories for events,

suggesting that the medial temporal lobes are essential for

new episodic memory formation. Second, clinical testing

indicated that H.M. could exhibit intact performance on

tests of memory with short retention delays of a few

seconds (short-term memory, or STM tasks), despite his

severe impairments on tests of memory with longer

retention delays (long-term memory, or LTM tasks).

Subsequent research identiﬁed patients who seemed to

exhibit impaired immediate memory in the face of normal

memory performance at long retention delays, suggesting

a neuropsychological double dissociation between STM

and LTM [2].

These ﬁndings, along with other neuropsychological

and behavioral results, have been used to advance the idea

that there must be at least two kinds of memory stores:

one that is used to retain information across short delays

(spanning several seconds) and another that is used to

retain information acrosslongerdelays(spanning

minutes, hours, days, etc.). Although the ‘multi-store’

view has been challenged [3–6] (see Box 1), it is widely

cited as fact, based on initial reports of dissociations

between STM and LTM (e.g. [7,8]). However, there has

been an accumulation of new evidence since these initial

Box 1. Psychological models of processes contributing to

STM and LTM

Although psychological models have long distinguished between

processes supporting STM and LTM, they did not necessarily

suggest neurally distinct memory stores. For example, Hebb [65]

proposed a ‘dual-trace mechanism’, suggesting: (1) that STM and

LTM can be supported by the re-activation of distributed networks of

strongly interconnected neurons, which he termed a ‘cell assembly’;

and (2) that STM can additionally be supported by ‘reverberating

activity’ within a cell assembly. In a similar vein, Atkinson and

Shiffrin [66] noted: ‘Our account of short-term and long-term storage

does not require that the two stores necessarily be in different parts

of the brain or involve different physiological structures. One might

consider the short-term store simply as being a temporary activation

of some portion of the long-term store.’

In contrast to this view, neuropsychological results suggested

dissociations between STM and LTM tasks. Related work focused on

functional and neuropsychological dissociations between primacy

and recency effects in verbal list learning, which were assumed to

reﬂect long-term and short-term retention processes, respectively

(this view has been rejected in more recent studies). These

dissociations led many to conclude that there must be a memory

store that speciﬁcally supports temporary retention of information,

independent of the memory store that supports retention across

long delays [2].

The idea of a temporary memory store was further developed by

Baddeley and Hitch [67] in their inﬂuential ‘working memory’ (WM)

model. The Baddeley and Hitch model proposed that short-term

retention (or ‘maintenance’) and manipulation of information across

short delays is mediated by interactions between a ‘central

executive’ and different short-term ‘buffers’ for different types of

information (i.e. visuospatial and phonological). Based largely on

neuropsychological evidence, it was assumed that each of these

short-term buffers is a temporary memory store, separate from the

stores for long-term retention. The Baddeley and Hitch model has

been remarkably successful, to the point that the term ‘working

memory’ is now often used to describe all types of STM tasks.

More recent models have attempted to explain relationships

between STM and LTM without suggesting structurally distinct

stores for temporary and long-term retention. For example, Cowan’s

model [3], like that of Baddeley and Hitch, has a ‘central executive’

that mediates goal-directed control processes. However, in Cowan’s

model active maintenance is accomplished by activating the same

representations that support LTM, rather than by transferring

information to temporary memory buffers. Similar ideas have been

advanced by neuroscientists [4,5], and computational modelers [68].

Although there are substantial differences between these activation-

based and structural accounts, such as the Baddeley and Hitch

model, there have been few experimental efforts to directly contrast

the two types of models.

Corresponding author: Ranganath, C. ([email protected]).

Available online 5 July 2005

Opinion TRENDS in Cognitive Sciences Vol.9 No.8 August 2005

reports, and there have been few efforts to evaluate

previously reported dissociations within the context of

these recent ﬁndings.

The goal of this article is to evaluate current

neuropsychological evidence for dissociations between

STM and LTM. More speciﬁcally, we will evaluate

evidence relevant to the idea that medial temporal lobe

lesions selectively affect LTM, whereas lesions to other

regions, such as the perisylvian cortex or the prefrontal

cortex, selectively affect STM. After reviewing this

evidence, we conclude that claims of dissociations between

STM and LTM are not well-supported. Finally, we

consider theoretical alternatives to the multi-store view

of memory.

Do the medial temporal lobes contribute to STM?

As noted above, patients with medial temporal lobe lesions

can perform normally on many STM measures, despite

their severely impaired LTM performance. Consequently,

it has been proposed that medial temporal cortical areas

(including perirhinal, parahippocampal and entorhinal

cortex), along with the hippocampus, comprise a ‘medial

temporal lobe memory system’ that is speciﬁcally necess-

ary for LTM, but not STM [9,10]. By contrast, other

temporal lobe areas, such as inferior temporal area TE,

have been associated with functions more relevant to

perception and STM.

Although the idea that the medial temporal lobes are

not required for STM is widespread (e.g. [7,8]), there have

been few rigorous attempts to test STM for different types

of information in patients with restricted medial temporal

damage. Such patients can clearly exhibit intact attention

and concentration, and can hold in mind instructions to

perform many complex tasks. Additionally, patients with

medial temporal damage can exhibit intact STM for

simple visual features or shapes ([9] and Prisko, unpub-

lished data), and even intact immediate memory for the

gist of lengthy, complex stories [11]. Such observations

suggest that, despite their LTM impairments, patients

with medial temporal damage can actively retain many

kinds of information across short delays.

However, it is unclear whether all forms of STM are

spared following medial temporal lobe lesions. For

example, most clinical tests of STM involve simple and

overlearned materials (e.g. words, digits, etc.) that are

well-represented in cortical areas outside of the medial

temporal region. Therefore, it might be more appropriate

to investigate STM for materials that are more likely to be

uniquely processed by the medial temporal region, such as

complex, novel objects [12,13]. Interestingly, several lines

of evidence suggest that medial temporal regions –

particularly the perirhinal cortex – play an essential role

in STM for complex, novel visual objects.

For example, neuropsychological studies have exam-

ined retention of novel visual objects in human amnesic

groups that included patients with extensive medial

temporal lobe lesions [14–18]. All but one of these studies

reported memory deﬁcits in medial temporal lobe

amnesics across retention delays as short as 2–10s (see

Figure 1). One could argue that STM deﬁcits in these

patients might have been caused by damage extending

into temporal cortical areas adjacent to the medial

temporal lobes, such as area TE. However, results from

controlled lesion studies in monkeys suggest that medial

temporal lesions that include the perirhinal cortex can

cause severe STM deﬁcits for novel objects, even when

damage to adjacent areas is minimal [19–27]. Interest-

ingly, one of these studies even reported impaired

performance with a zero second retention delay [20].These

data seriously question the assumption that the medial

temporal lobes are not necessary for any form of STM.

Results from single-unit recording studies of monkeys

[28–31] complement the lesion evidence by demonstrating

that perirhinal and entorhinal neurons exhibit persistent,

object-selective activity during retention delays in STM

tasks – a putative neural mechanism for active mainten-

ance (see Box 2). Consistent with these ﬁndings, neuro-

imaging studies of humans have reported perirhinal,

parahippocampal, entorhinal, and hippocampal activity

during the performance of STM tasks with novel visual

objects [32–35], faces [36], or scenes [37,38]. This activity

is enhanced for novel stimuli relative to repeated stimuli

[34,36,38], and enhanced for stimuli that are successfully

remembered after both short and long delays [34,37].

In summary, the available evidence indicates that

medial temporal lesions can impair retention of infor-

mation about complex objects across short delays and that

medial temporal regions exhibit persistent activity during

active maintenance of novel visual objects. These ﬁndings

provide compelling evidence for the idea that, at least

under some circumstances, medial temporal cortical

regions are necessary for, and contribute to normal STM.

Do patients with perisylvian cortex lesions show

selective STM deﬁcits?

Another kind of evidence cited to support the idea of

neurally distinct short- and long-term memory stores

comes from studies of patients with severe impairments in

phonological STM [39]. These patients typically have

damage to left perisylvian cortex and/or underlying white

matter (Ravizza, S. et al., unpublished data). Such

patients can show intact LTM for meaningful words

[2,40], suggesting that STM can be impaired without

affecting LTM.

However, this inference is problematic because the

types of tasks and measures used to assess STM and LTM

differ in several ways that have nothing to do with the

retention interval. For example, in the span tasks used to

assess phonological STM, one must immediately recall a

short sequence of spoken digits in the correct order. In a

typical LTM task, one must learn a long list of meaningful

words (presumably presented at a slow rate), and recall

performance is assessed across multiple learning trials. At

the most basic level, comparisons between such tests are

confounding differences in retention interval and/or list

length with differences in the type of information to be

maintained (e.g. meaningless digits versus meaningful

words). Therefore, to test whether patients with phonolo-

gical STM deﬁcits can exhibit intact phonological LTM, it

is at least necessary to test memory for information that is

difﬁcult to encode semantically or visually. Crucially,

patients with phonological STM deﬁcits exhibit severely

Opinion TRENDS in Cognitive Sciences Vol.9 No.8 August 2005 375

www.sciencedirect.com

impaired LTM for auditorily-presented, meaningless non-

words [41–44].Thus,theoverallpatternofresults

suggests that patients with perisylvian lesions can exhibit

normal LTM for information that can be encoded visually

or semantically, but they clearly have deﬁciencies in

phonological STM and LTM. Accordingly, there is no

reason to believe that these patients exhibit a dissociation

between STM and LTM.

Does the prefrontal cortex disproportionately contribute

to STM?

Another area that is often cited as speciﬁcally supporting

STM is the prefrontal cortex [7]. Indeed, the discovery that

prefrontal neurons exhibit persistent, stimulus-selective

activity during short memory delays [45] has fueled

intensive research on neural STM signals in the pre-

frontal cortex. However, this phenomenon is not speciﬁc to

prefrontal cortex (see Box 2). Recent studies have shown

that several neocortical areas exhibit persistent, stimulus-

speciﬁc activity during short retention delays [46,47].

More crucially, results from neuropsychological studies do

not suggest that the lateral prefrontal cortex is a critical

site for short-term storage per se.Lesionstudiesin

monkeys and humans have failed to ﬁnd consistent effects

of lateral prefrontal damage on the retention of objects or

verbal information across short delays, and the effects that

have been reported were attributed to deﬁciencies in task

learning, attention, or response selection, rather than

mnemonic deﬁcits [15,48–55].

For example, D’Esposito and colleagues [55] recently

examined verbal STM performance in patients with

unilateral prefrontal lesions. Results from this study

showed that digit span (an STM measure used to assess

deﬁcits in the patients with perisylvian lesions described

earlier) was normal in these patients, consistent with

results of a previous meta-analysis of neuropsychological

studies [53]. Patients and controls next performed tasks in

which either one letter or 3–4 letters (the load was

matched to the digit span of each patient) were retained

across a 6.5s delay. On some trials, distracting words were

presented during the retention delay, whereas on other

trials the delay was unﬁlled. As shown in Figure 2,

patients with prefrontal lesions performed normally, even

when retaining multiple letters in the face of distraction.

These ﬁndings are difﬁcult to reconcile with the view that

the prefrontal cortex is necessary for verbal short-term

storage.

Neuropsychological evidence additionally suggests that

prefrontal lesions can affect performance on LTM tasks.

Whereas patients with prefrontal lesions can perform at or

near normal levels on LTM tasks when given structured

encoding tasks and simple test formats, they exhibit

impaired performance when forced to initiate strategies to

actively encode information or when given tests that

require strategic processing during retrieval [56,57].

100

2 5 10 30 60

Delay (s)

% correct

CON (n=12)

AMN (n=5)

100

10 30 90

Delay (s)

% correct

CON (n=7)

MTL (n=3)

Delay (s)

% correct

CON (n=44)

MTL (n=11)

0412

0 – 2 25 – 40

% correct

CON (n=8)

MTL (n=2)

100

025102030

% correct

MTL (n=4)

CON (n=9)

TRENDS in Cognitive Sciences

6 – 10

Delay (s) Delay (s)

Holdstock

et al.

(1995)Aggleton

et al.

(1992) Owen

et al.

(1995)

(a) (b) (c)

Buffalo

et al.

(1998) Holdstock

et al.

(2000)

(d) (e)

Figure 1. Recognition memory for novel visual objects across short retention intervals in patients with medial temporal lesions. (a) Results from an alcoholic control group

(CON) and a group of 3 patients with medial temporal damage following viral encephalitis (MTL) [18]. (b) Results from a healthy control group (CON) and a mixed amnesic

patient group (AMN) consisting of three patients with medial temporal damage following viral encephalitis, one patient with bilateral thalamic damage, and one patient with

a fornix lesion following surgical removal of a colloid cyst [14]. (c) Results from a healthy control group (CON) and a group of 11 patients who underwent unilateral amygdalo-

hippocampectomies (MTL) for treatment of epilepsy [15]. (d) Results from a healthy control group (CON) and from two patients with bilateral temporal lobe damage due to

herpes simplex encephalitis (MTL) [16]. (e) Results from a healthy control (CON) group and from four patients with bilateral temporal lobe lesions due to encephalitis or

meningitis (MTL) [17]. Each graph shows percentage accuracy as a function of the retention interval, plotted from results depicted in the original reports.

Opinion TRENDS in Cognitive Sciences Vol.9 No.8 August 2005376

www.sciencedirect.com

Considered along with the range of nonmnemonic deﬁcits

induced by prefrontal lesions [58,59], these ﬁndings

suggest that the prefrontal cortex is not a short-term

storage buffer, but rather that it is necessary for

implementing control processes that can contribute to

both STM and LTM performance [47,56,60,61].

How can the lesion ﬁndings described above be reconciled

with neuroimaging and single-unit recording studies

showing persistent prefrontal activity during maintenance

of information across short delays? One possibility is that

persistent activity in prefrontal cortex serves to modulate

activity in posterior cortical regions that represent

Box 2. Different types of neural signals for short-term memory

Results from single-unit recording studies of monkeys and neuroima-

ging studies of humans suggest that there are two classes of neural

signals that might disproportionately contribute to STM relative to

LTM (Figure I). One class would be best described as ‘active’ STM

signals, in that they are primarily seen under circumstances when one

is attempting to actively maintain information across short delays.

Most research has focused on persistent, stimulus-speciﬁc activity,

which is thought by many to be a neural correlate of rehearsal or active

maintenance (Figure Ib). The ability of persistent activity to maintain

information across short delays has been described in biologically-

plausible computational models based on known properties of

neocortical neurons [68]. In addition to persistent activity, another

active STM signal has been described as ‘match enhancement’. This

refers to an enhanced response to a stimulus that matches one that

has been maintained across a delay period (Figure Ic). Match

enhancement effects have been reported in lateral prefrontal, inferior

temporal, and medial temporal cortex in both single-unit recording

studies in monkeys [69,70] and neuroimaging studies of humans

[71,72].

A second class of neural signals that might contribute to STM can be

described as ‘passive’, because they are seen in response to a stimulus

even if it is not task-relevant [69]. The most commonly observed

passive STM signal is a reduction in neural responses to recently

encountered stimuli, relative to stimuli that have not been recently

encountered (Figure Id). Available evidence implicates several

mechanisms – including neural adaptation, synaptic plasticity, and

neuromodulatory inﬂuences – that bring about activity reductions to

recently encountered items [73]. It has been proposed that activity

reductions in the perirhinal cortex could signal the occurrence of a

recently encountered item [74]. If so, some mechanisms for neural

activity reductions might selectively support LTM (e.g. synaptic

plasticity), whereas others might selectively support STM (e.

g. neural adaptation) [73]. Interestingly, repetition-related activity

reductions in the perirhinal cortex are lag-sensitive, such that the

effects are most pronounced at short lags between the ﬁrst and second

exposure [74,75]. This characteristic is similar to the behavioral

inﬂuence of familiarity on human recognition decisions, which is

also lag-sensitive and more prominent at short retention intervals [76].

Sample Nonmatch

Repeated

nonmatch

Match

Typical STM task

Persistent activity

Repetition-related activity reduction

Match enhancement

Time from sample onset (ms)

Spikes/sec

Time from sample onset (ms)

Spikes/sec

Time from sample onset (ms)

Spikes/sec

(a)

(b)

(c)

(d)

TRENDS in Cognitive Sciences

Figure I. Examples of neural STM signals recorded from the entorhinal cortex. (a) In a typical STM task used in single-unit recording studies, a sample object must be

retained across a delay and compared with a series of test objects, until a matching object is presented. In some tasks, as in the ﬁgure, nonmatching test objects are

repeated, allowing researchers to differentiate between neural signals sensitive to item repetition and signals more speciﬁcally sensitive to the item that is being actively

maintained. (b) Recordings from a neuron showing persistent activity during the retention of an object. The periods of presentation of the sample (SM), nonmatching test

objects (NM), and the matching test object (M) are denoted by gray bars. Note that persistent activity in this neuron remains robust, even following presentation of

nonmatching test items. (c) Recordings from a neuron showing match enhancement. These effects are not seen for repeated nonmatching items. (d) Recordings from a

neuron showing reduced activity in response to nonmatching items. These neurons also show reduced activity to repeated nonmatching items, suggesting that they are

generally sensitive to stimulus repetition. Adapted from Ref. [28].

Opinion TRENDS in Cognitive Sciences Vol.9 No.8 August 2005 377

www.sciencedirect.com

information that is being maintained. Results from studies

combining lesion and physiology techniques are consistent

with this possibility. Speciﬁcally, these studies have shown

that prefrontal lesions might disrupt STM task performance

through dysregulation of task-relevant activity in posterior

cortical areas. For example, in one study, activity was

recorded from inferior temporal temporal neurons before

and after cooling probes were used to induce reversible

prefrontal lesions. Before cooling, inferior temporal neurons

showed persistent sample-speciﬁc activity during the delay

period of the memory task. However, this effect was

attenuated by prefrontal cooling [62], suggesting that top-

down feedback from prefrontal regions was critical for

supporting robust STM-related activity in inferior temporal

cortex. Another study investigated STM performance and

electrophysiological activity in human patients with pre-

frontal lesions [54]. Behaviorally, the patients were

impaired at retaining information across short delays

when distracting sounds were presented during the memory

delay. Concurrent event-related potential recordings

showed that this effect was accompanied by increased

neural responses to the distracters in posterior cortical

areas. The authors interpreted the results to reﬂect

disinhibition of posterior cortical areas following prefrontal

lesions. These ﬁndings indicate that, rather than being a

short-term storage buffer, the prefrontal cortexprovides top-

down feedback to support maintenance of information that

is represented in posterior cortical areas.

Theoretical alternatives to the multi-store view

As described earlier, most arguments for neurally distinct

short- and long-term memory stores are based on reported

neuropsychological dissociations between STM and LTM.

However, consideration of these reports in the context of

more recent ﬁndings suggests a more complex picture.

Contrary to popular belief, neuropsychological evidence

indicates that medial temporal damage can impair STM

under some circumstances. Medial temporal cortical

regions, like other neocortical regions [47], also exhibit

persistent activity during memory delays, a neural signal

that might speciﬁcally support STM. Neuropsychological

evidence also indicates that individuals with phonological

STM deﬁcits following perisylvian damage also show LTM

deﬁcits when tested appropriately. Finally, neuropsycho-

logical evidence suggests that the prefrontal cortex – an

area popularly identiﬁed with STM – contributes to both

STM and LTM in a circumscribed manner. Put together,

these ﬁndings raise serious doubts about popular claims

regarding double dissociations between STM and LTM. It

follows that there are good reasons to question whether

STM and LTM must be supported by separate neocortical

memory stores or systems. Indeed, there are several

alternative ways to characterize the neural mechanisms

that support STM and LTM.

One possibility is that there are no large-scale cortical

areas that uniquely support STM or LTM, but there might

be neural circuits that differentially support STM and

LTM within the same cortical area. By this view, one

might not expect gross lesion studies to reveal double

dissociations between STM and LTM, because the

underlying memory ‘stores’ might be at a much ﬁner

spatial scale. Such a division of labor might be crucial for

normal episodic memory formation [63]. For example, a

short-term storage buffer might be necessary for encoding

of temporal sequences or spatial maps by relating or

binding aspects of events that unfold over time [63].

Another possibility is that distinctions between neural

mechanisms of STM and LTM can be explained through

differences in the dynamics of activation within the same

neocortical memory circuits [4]. The simplest model would

suggest that: (i) LTM and STM are supported by

reactivation of neocortical memory networks either in

response to an external stimulus or through internal, top-

down signals for memory retrieval; and (ii) STM can

additionally be supported by a family of neural signals,

including persistent stimulus-speciﬁc activity, match

enhancement, and repetition-related activity reductions

(Box 2). According to this view, the crucial differences

among neocortical memory stores or systems is not their

contribution to memory at different retention intervals,

but rather the types of information they process and

represent [64]. Thus, one could easily assume that that

there are multiple memory stores, each representing

different types of information across short and long delays.

Conclusions

In conclusion, we have learned a great deal since the

initial studies of H.M. and other patients with memory

disorders. We argue here that the emerging pattern of

evidence raises doubts about popular claims of dis-

sociations between STM and LTM. This provokes new

questions regarding how best to characterize the neural

mechanisms that support STM and LTM (see also Box 3).

0.0

0.2

0.4

0.6

0.8

1.0

Frontal lesions Controls (55–70 yrs) Controls (71–85 yrs)

Proportion correct

distraction

Distraction

distraction

Distraction

1 letter 3 or 4 letters

TRENDS in Cognitive Sciences

Figure 2. Results showing intact STM in patients with prefrontal lesions [55].Inthis

study, patients with unilateral prefrontal lesions and two control groups completed

STM tasks that required maintenance of letters across a 6.5 s delay. Each subject

was pre-tested to identify STM capacity and then completed the STM task with

either 1 item, or 3 or 4 items corresponding to the subjects’ capacity. On some trials,

subjects maintained the letters across an unﬁlled retention delay, whereas on

others, visual distracters were presented during the delay. Recall performance was

intact in patients with prefrontal lesions, even when they had to retain multiple

items across a distracter-ﬁlled retention delay.

Opinion TRENDS in Cognitive Sciences Vol.9 No.8 August 2005378

www.sciencedirect.com

Clearly, more research will be necessary to determine

whether these questions are best answered by models that

propose neurally distinct short-term and long-term stores

or by models that do not assume distinctions based on

retention intervals.

Acknowledgements

Our research is supported by grants from the Institute for Research on

Pathological Gambling and Related Disorders, NIMH grant R01

MH68721, and NINDS grant PO1 NS40813. We gratefully acknowledge

suggestions and constructive criticism from Craig Brozinsky, Aaron

Heller, Debbie Hannula, Rik Henson, Betsy Murray, Linda Murray, Alex

Martin, Colleen Parks, Susan Ravizza, Carter Wendelken, Andy

Yonelinas, the TICS editor Shbana Rahman, and the anonymous

reviewers. Special thanks to the authors of studies whose data are

plotted in the ﬁgures.

References

1 Scoville, W.B. and Milner, B. (1957) Loss of recent memory after

bilateral hippocampal lesions. J. Neurol. Neurosurg. Psychiatry 20,

11–21

2 Shallice, T. and Warrington, E.K. (1970) Independent functioning of

verbal memory stores: A neuropsychological study. Q.J. Exp. Psychol.

22, 261–273

3 Cowan, N. (1997) Attention and Memory, Oxford University Press

4 Fuster, J.M. (1995) Memory in the Cerebral Cortex, MIT Press

5 Ruchkin, D.S. et al. (2003) Working memory retention systems: a state

of activated long-term memory. Behav. Brain Sci. 26, 709–728

6 Davelaar, E.J. et al. (2005) The demise of short-term memory

revisited: empirical and computational investigations of recency

effects. Psychol. Rev. 112, 3–42

7 Speer, N.K. et al. (2003) Strategy-dependent changes in memory:

effects on behavior and brain activity. Cogn. Affect. Behav. Neurosci. 3,

155–167

8 Ryan, J.D. and Cohen, N.J. (2004) Processing and short-term

retention of relational information in amnesia. Neuropsychologia 42,

497–511

9 Cave, C.B. and Squire, L.R. (1992) Intact verbal and nonverbal short-

term memory following damage to the human hippocampus. Hippo-

campus 2, 151–163

10 Alvarez, P. et al. (1994) The animal model of human amnesia: long-

term memory impaired and short-term memory intact. Proc. Natl.

Acad. Sci. U. S. A. 91, 5637–5641

11 Baddeley, A. and Wilson, B.A. (2002) Prose recall and amnesia:

implications for the structure of working memory. Neuropsychologia

40, 1737–1743

12 Lee, A.C. et al. (2005) Perceptual deﬁcits in amnesia: challenging the

medial temporal lobe ‘mnemonic’ view. Neuropsychologia 43, 1–11

13 Murray, E.A. and Bussey, T.J. (1999) Perceptual-mnemonic functions

of the perirhinal cortex. Trends Cogn. Sci. 3, 142–151

14 Holdstock, J.S. et al. (1995) The performance of amnesic subjects on

tests of delayed matching-to-sample and delayed matching-to-

position. Neuropsychologia 33, 1583–1596

15 Owen, A.M. et al. (1995) Visuo-spatial short-term recognition memory

and learning after temporal lobe excisions, frontal lobe excisions or

amygdalo-hippocampectomy in man. Neuropsychologia 33, 1–24

16 Buffalo, E.A. et al. (1998) The human perirhinal cortex and

recognition memory. Hippocampus 8, 330–339

17 Holdstock, J.S. et al. (2000) Perceptual and mnemonic matching-to-

sample in humans: contributions of the hippocampus, perirhinal and

other medial temporal lobe cortices. Cortex 36, 301–322

18 Aggleton, J.P. et al. (1992) The performance of postencephalitic

amnesic subjects on two behavioural tests of memory: concurrent

discrimination learning and delayed matching-to-sample. Cortex 28,

359–372

19 Zola-Morgan, S. et al. (1989) Lesions of perirhinal and parahippo-

campal cortex that spare the amygdala and the hippocampal

formation produce severe memory impairment. J. Neurosci. 9,

4355–4370

20 Eacott, M.J. et al. (1994) Preserved recognition memory for small sets,

and impaired stimulus identiﬁcation for large sets following rhinal

cortex ablations in monkeys. Eur. J. Neurosci. 6, 1466–1478

21 Mahut, H. et al. (1982) Hippocampal resections impair associative

learning and recognition memory in the monkey. J. Neurosci. 2,

1214–1229

22 Meunier, M. et al. (1993) Effects on visual recognition of combined and

separate ablations of the entorhinal and perirhinal cortex in rhesus

monkeys. J. Neurosci. 13, 5418–5432

23 Murray, E.A. and Mishkin, M. (1986) Visual recognition in monkeys

following rhinal cortical ablations combined with either amygdalect-

omy or hippocampectomy. J. Neurosci. 6, 1991–2003

24 Murray, E.A. and Mishkin, M. (1984) Severe tactual as well as visual

memory deﬁcits follow combined removal of the amygdala and

hippocampus in monkeys. J. Neurosci. 4, 2565–2580

25 Zola-Morgan, S. and Squire, L. (1985) Medial temporal lesions on

monkeys impair memory in a variety of tasks sensitive to human

amnesia. Behav. Neurosci. 99, 22–34

26 Zola-Morgan, S. and Squire, L.R. (1986) Memory impairment in

monkeys following lesions limited to the hippocampus. Behav.

Neurosci. 100, 155–160

27 Zola-Morgan, S. et al. (1993) Damage to the perirhinal cortex

exacerbates memory impairment following lesions to the hippocampal

formation. J. Neurosci. 13, 251–265

28 Suzuki, W.A. et al. (1997) Object and place memory in the macaque

entorhinal cortex. J. Neurophysiol. 78, 1062–1081

29 Miyashita, Y. and Chang, H.S. (1988) Neuronal correlate of pictorial

short-term memory in the primate temporal cortex. Nature 331, 68–70

30 Nakamura, K. and Kubota, K. (1995) Mnemonic ﬁring of neurons in

the monkey temporal pole during a visual recognition memory task.

J. Neurophysiol. 74, 162–178

31 Miller, E.K. et al. (1993) Activity of neurons in anterior inferior

temporal cortex during a short-term memory task. J. Neurosci. 13,

1460–1478

32 Davachi, L. and Goldman-Rakic, P.S. (2001) Primate rhinal cortex

participates in both visual recognition and working memory tasks:

functional mapping with 2-DG. J. Neurophysiol. 85, 2590–2601

33 Sybirska, E. et al. (2000) Prominence of direct entorhinal-CA1

pathway activation in sensorimotor and cognitive tasks revealed by

2-DG functional mapping in nonhuman primate. J. Neurosci. 20,

5827–5834

34 Ranganath, C. et al. (2005) Working memory maintenance contributes

to long-term memory formation: Neural and behavioral evidence.

J. Cogn. Neurosci. 17, 994–1010

35 Elliott, R. and Dolan, R. (1999) Differential neural responses during

performance of matching and nonmatching to sample tasks at two

delay intervals. J. Neurosci. 19, 5066–5073

Box 3. Questions for future research

† Are there any neocortical areas or neural circuits that selectively

support perception versus active maintenance versus long-duration

memory storage [64]?

† Are there fundamental differences in the types of encoding or

retrieval processes that are typically engaged during STM and LTM

tasks [77]?

† Can persistent activity represent signals that are useful for STM in

the absence of conscious awareness?

† How do network-level interactions between different neocortical

areas support active maintenance of information across short delays

[4]?

† Does persistent activity during short-term maintenance contribute

to successful LTM for sequences or spatial maps [63]?

† What is the relationship between synchronized neural oscillations

and neural STM signals?

† What kinds of memories, if any, are represented in prefrontal

cortical networks [4]?

† Is the hippocampus or entorhinal cortex necessary for represent-

ing or retaining information about spatial and/or arbitrary non-

spatial relations across short delays [78]?

† Subcortical areas, such as the mediodorsal thalamic nucleus, have

been implicated in both STM and LTM. What are the roles of these

areas, and how are their contributions to STM and LTM different

from neocortical areas?

Opinion TRENDS in Cognitive Sciences Vol.9 No.8 August 2005 379

www.sciencedirect.com

36 Ranganath, C. and D’Esposito, M. (2001) Medial temporal lobe

activity associated with active maintenance of novel information.

Neuron 31, 865–873

37 Schon, K. et al. (2004) Persistence of parahippocampal representation

in the absence of stimulus input enhances long-term encoding: a

functional magnetic resonance imaging study of subsequent memory

after a delayed match-to-sample task. J. Neurosci. 24, 11088–11097

38 Stern, C.E. et al. (2001) Medial temporal and prefrontal contributions

to working memory tasks with novel and familiar stimuli. Hippo-

campus 11, 337–346

39 Warrington, E.K. and Shallice, T. (1969) The selective impairment of

auditory verbal short term memory. Brain 92, 885–896

40 Warrington, E.K. et al. (1971) The anatomical localisation of selective

impairment of auditory-verbal short-term memory. Neuropsychologia

9, 377–387

41 Belleville, S. et al. (1997) Neuropsychological argument for the

activation approach to memory: A case of phonological memory

deﬁcit. Brain Cogn. 35, 382–385

42 Belleville, S. et al. (2003) A neuropsychological argument for a

processing view of memory. J. Mem. Lang. 48, 686–703

43 Baddeley, A. et al. (1988) When long-term learning depends on short-

term storage. J. Mem. Lang. 27, 586–595

44 Martin, R.C. (1993) Short-term memory and sentence processing:

evidence from neuropsychology. Mem. Cogn. 21, 176–183

45 Fuster, J.M. and Alexander, G.E. (1971) Neuron activity related to

short-term memory. Science 173, 652–654

46 Pasternak, T. and Greenlee, M.W. (2005) Working memory in primate

sensory systems. Nat. Rev. Neurosci. 6, 97–107

47 Ranganath, C. and D’Esposito, M. (2005) Directing the mind’s eye:

prefrontal, inferior and medial temporal mechanisms for visual

working memory. Curr. Opin. Neurobiol. 15, 175–182

48 Rushworth, M.F. et al. (1997) Ventral prefrontal cortex is not essential

for working memory. J. Neurosci. 17, 4829–4838

49 Bachevalier, J. and Mishkin, M. (1986) Visual recognition impairment

follows ventromedial but not dorsolateral prefrontal lesions in

monkeys. Behav. Brain Res. 20, 249–261

50 Kowalska, D.M. et al. (1991) The role of the inferior prefrontal

convexity in performance of delayed nonmatching-to-sample. Neu-

ropsychologia 29, 583–600

51 Meunier, M. et al. (1997) Effects of orbital frontal and anterior

cingulate lesions on object and spatial memory in rhesus monkeys.

Neuropsychologia 35, 999–1015

52 Mishkin, M. and Manning, F.J. (1978) Non-spatial memory after

selective prefrontal lesions in monkeys. Brain Res. 143, 313–323

53 D’Esposito, M. and Postle, B.R. (1999) The dependence of span and

delayed-response performance on prefrontal cortex. Neuropsychologia

37, 1303–1315

54 Chao, L.L. and Knight, R.T. (1998) Contribution of human prefrontal

cortex to delay performance. J. Cogn. Neurosci. 10, 167–177

55 D’Esposito, M. et al. The role of prefrontal cortex on component

processes of working memory: evidence from lesion and fMRI data.

J. Int. Neuropsychol. Soc. (in press)

56 Ranganath, C. and Knight, R.T. (2003) Prefrontal cortex and episodic

memory: Integrating ﬁndings from neuropsychology and event-

related functional neuroimaging. In The Cognitive Neuroscience of

Memory Encoding and Retrieval (Parker, A. et al., eds), pp. 83–99,

Psychology Press

57 Shimamura, A.P. (1995) Memory and frontal lobe function. In The

Cognitive Neurosciences (Gazzaniga, M.S., ed.), pp. 803–813, MIT

Press

58 Stuss, D.T. and Benson, D.F. (1986) The Frontal Lobes, Raven Press

59 Shimamura, A.P. (2000) The role of the prefrontal cortex in dynamic

ﬁltering. Psychobiology 28, 207–218

60 Fletcher, P.C. and Henson, R.N. (2001) Frontal lobes and human

memory: Insights from functional neuroimaging. Brain 124, 849–881

61 Ranganath, C. et al. (2003) Prefrontal activity associated with

working memory and episodic long-term memory. Neuropsychologia

41, 378–389

62 Fuster, J.M. et al. (1985) Functional interactions between inferotem-

poral and prefrontal cortex in a cognitive task. Brain Res. 330,

299–307

63 Howard, M.W. et al. (2005) The temporal context model in spatial

navigation and relational learning: toward a common explanation of

medial temporal lobe function across domains. Psychol. Rev. 112,

75–116

64 Gaffan, D. (2002) Against memory systems. Philos. Trans. R. Soc.

Lond. B Biol. Sci. 357, 1111–1121

65 Hebb, D.O. (1949) Organization of Behavior: A Neuropsychological

Theory, Wiley

66 Atkinson, R.C. and Shiffrin, R.M. (1971) The control of short-term

memory. Sci. Am. 225, 82–90

67 Baddeley, A. and Hitch, G.J. (1974) Working memory. In Recent

Advances in Learning and Motivation (Vol. VIII) (Bower, G., ed.),

pp. 47–90, Academic Press

68 Amit, D.J. and Mongillo, G. (2003) Selective delay activity in the

cortex: phenomena and interpretation. Cereb. Cortex 13, 1139–1150

69 Miller, E.K. and Desimone, R. (1994) Parallel neuronal mechanisms

for short-term memory. Science 263, 520–522

70 Holscher, C. and Rolls, E.T. (2002) Perirhinal cortex neuronal activity

is actively related to working memory in the macaque. Neural Plast. 9,

41–51

71 Druzgal, T.J. and D’Esposito, M. (2001) A neural network reﬂecting

decisions about human faces. Neuron 32, 947–955

72 Jiang, Y. et al. (2000) Complementary neural mechanisms for tracking

items in human working memory. Science 287, 643–646

73 Ranganath, C. and Rainer, G. (2003) Neural mechanisms for detecting

and remembering novel events. Nat. Rev. Neurosci. 4, 193–202

74 Xiang, J.Z. and Brown, M.W. (1998) Differential neuronal encoding of

novelty, familiarity and recency in regions of the anterior temporal

lobe. Neuropharmacology 37, 657–676

75 Brozinsky, C.J. et al. (2005) Lag-sensitive repetition suppression

effects in the anterior parahippocampal gyrus. Hippocampus. doi: 10.

1002/hipo.20087

76 Yonelinas, A.P. and Levy, B.J. (2002) Dissociating familiarity from

recollection in human recognition memory: different rates of

forgetting over short retention intervals. Psychon. Bull. Rev. 9,

575–582

77 Cabeza, R. et al. (2002) Similarities and differences in the neural

correlates of episodic memory retrieval and working memory. Neuro-

image 16, 317–330

78 Buckmaster, C.A. et al. (2004) Entorhinal cortex lesions disrupt the

relational organization of memory in monkeys. J. Neurosci. 24,

9811–9825

Free journals for developing countries

The WHO and six medical journal publishers have launched the Access to Research Initiative, which enables nearly 70 of the world’s

poorest countries to gain free access to biomedical literature through the Internet.

Gro Harlem Brundtland, director-general for the WHO, said that this initiative was ‘perhaps the biggest step ever taken towards reducing

the health information gap between rich and poor countries’.

See http://www.healthinternetwork.net for more information.

Opinion TRENDS in Cognitive Sciences Vol.9 No.8 August 2005380