strategic control of memory - kuhllab.com

This article was originally published in the Encyclopedia of Neuroscience published by Elsevier, and the attached copy is provided by Elsevier for the

author's benefit and for the benefit of the author's institution, for non-commercial research and educational use including without limitation use in

instruction at your institution, sending it to specific colleagues who you know, and providing a copy to your institution’s administrator.

All other uses, reproduction and distribution, including without limitation commercial reprints, selling or licensing copies or access, or posting on open

internet sites, your personal or institution’s website or repository, are prohibited. For exceptions, permission may be sought for such use through

Elsevier's permissions site at:

http://www.elsevier.com/locate/permissionusematerial

Kuhl B A and Wagner A D (2009) Strategic Control of Memory. In: Squire LR (ed.) Encyclopedia of Neuroscience, volume 9, pp. 437-444.

Oxford: Academic Press.

Strategic Control of MemoryB A Kuhl and A D Wagner, Stanford University,Stanford, CA, USA

ã 2009 Elsevier Ltd. All rights reserved.

Introduction

Cognitive control mechanisms permit memory to beaccessed strategically and so aid in bringing knowl-edge to mind that is relevant to current decisions andactions. A fundamental component of the strategiccontrol of memory is the resolution of interferencefrom competing, irrelevant representations. This arti-cle considers how the ventrolateral prefrontal cortex(VLPFC) regulates mnemonic competition in multiplememory systems. We initially discuss how damage tolateral prefrontal cortex impacts mnemonic functionand then consider recent neuroimaging and focal lesionfindings that highlight the distinct roles that subregionsof the VLPFC play in the control of memory.

Lateral Prefrontal Cortex

The lateral prefrontal cortex (PFC) consists of ven-tral, dorsal, and frontopolar subregions. In this arti-cle, we primarily focus on the function of VLPFC. Inthe human (Figure 1(a)), VLPFC corresponds to theinferior frontal gyrus, which includes (moving cau-dally to rostrally) the inferior frontal pars opercularis(Brodmann area (BA) 44), inferior frontal pars trian-gularis (BA 45), and inferior frontal pars orbitalis (anarea that Petrides and Pandya term area 47/12, whichcorresponds to the lateral portion of BA 47). WhereasPetrides and Pandya refer to area 47/12 and BA 45collectively as the mid-VLPFC, distinguishing theseregions from the caudally situated BA 44, in thisarticle we functionally distinguish area 47/12 fromBA 45. Thus, we use anterior VLPFC to refer tothe inferior frontal pars orbitalis (area 47/12), mid-VLPFC to refer to the pars triangularis (BA 45), andposterior VLPFC to refer to the pars opercularis(BA 44). We note that the caudal portion of area47/12 actually lies ventral (rather than rostral) toBA 45 (Figure 1(a)). The inferior frontal sulcus inhumans (and the principal sulcus in monkeys) marksthe approximate boundary between the VLPFC andthe dorsolateral PFC (DLPFC). In terms of anatomi-cal connectivity, the VLPFC is strongly connectedwith cortical areas in the lateral temporal lobe andmedial temporal lobe (MTL), including (but not lim-ited to) the inferotemporal cortex, superior temporalcortex, and, more medially, perirhinal and parahip-pocampal cortices.

Mnemonic Deficits Following LateralPrefrontal Damage

Neuropsychological studies in humans and lesionstudies in animals indicate that insult to the lateralPFC can produce memory impairments that arequalitatively distinct from those observed followingMTL damage. Whereas MTL damage results in anamnesic condition that reflects the inability toencode and retrieve new declarative memories –long-term memories for events (episodic memory)and facts (semantic memory) – lateral PFC damageimpairs the strategic regulation of multiple forms ofmemory, including declarative memory and workingmemory. Significantly, impairments in episodic mem-ory, semantic memory, and working memory follow-ing lateral PFC damage are often most apparentwhen performance requires the resolution of interfer-ence, which led Moscovitch, Shimamura, and othersto propose that lateral PFC subserves cognitive con-trol mechanisms that regulate how we work withor dynamically filter memory.

Episodic Memory Deficits

Patients with lateral PFC damage show modestimpairments in their ability to encode and retrieveepisodic memories, with these deficits being particu-larly apparent when memory for target informationis required in the face of distraction. Here we illus-trate these deficits by highlighting a number of well-documented paradigms in which PFC patients areimpaired.

First, PFC patients show disproportionate deficitson tests of free recall compared to item recognition.Theorists have argued that the unconstrained natureof free recall increases the likelihood of mnemoniccompetition or interference. Second, even withinrecognition tests, PFC patients show impairmentswhen distracter items at test (foils) are similar toitems from the study list. In other words, PFC patientssuffer interference from items that are similar to thosethey studied, as evidenced by false claims of havingstudied these items. Third, in tests of source memory,in which memories for individual items must beattributed to a particular learning context – a discrim-ination that presents high interference because irrele-vant contexts are often highly salient – PFC patientsexhibit disproportionate deficits relative to tests ofitem recognition. Fourth, the use of AB–AC learningparadigms reveals that learning an AC associa-tion is disproportionately impaired in PFC patients(i.e., patients show a heightened sensitivity to pro-active interference). In such paradigms, the patient’s

Strategic Control of Memory 437

Encyclopedia of Neuroscience (2009), vol. 9, pp. 437-444

Author's personal copy

memory for AC pairs is largely compromised byincreased intrusions of the earlier learned AB asso-ciations. Moreover, even when sequentially studyingtwo unrelated lists of items (i.e., the lists do notshare common associates), PFC patients show dispro-portionate list 1 interference when trying to learnlist 2. Finally, when PFC patients are provided witha subset of items from a previously studied list asretrieval cues (i.e., part-list cueing), the high salienceof the provided items results in increased impair-ment in the patients’ ability to recall the remaining(unpresented) items from the list. Collectively, theseand other observations indicate that the ability toresolve interference in episodic memory is compro-mised following lateral PFC damage.

Semantic Memory Deficits

Just as PFC patients show impairments in uncon-strained episodic memory tests, they also show defi-cits in unconstrained tests of semantic retrieval. Forexample, patients show reduced total output on ver-bal fluency tasks, which involve freely generatingas many unique exemplars as possible in responseto a semantic or orthographic cue. This impairmentis thought to partially reflect patients’ heightenedsensitivity to output interference stemming fromthe initially retrieved exemplars. Significantly, thisinterference-dependent deficit during semantic or

lexical retrieval stands in contrast to the typicallyunimpaired ability of PFC patients to recognizesemantic structure or evaluate semantic relationships.Beyond verbal fluency tasks, patients with damageto the left lateral PFC fail to show normal semanticpriming when the meaning of a semantic cue (theprime) is a context-dependent homograph, comparedto words with less ambiguous meanings. This impair-ment may be due to PFC patients failing to retrievethe appropriate stimulus meaning when the primeis ambiguous, thereby preventing priming fromoccurring. Both these deficits are consistent with theperspective that the lateral PFC is recruited duringcontrolled semantic retrieval, as well as when inter-ference between competing semantic or lexical repre-sentations must be resolved.

Working Memory Deficits

In tests of working memory, PFC patients are oftenunimpaired, as measured by simple tests of verbalspan. However, in tests in which maintenance iscomplicated by interference from irrelevant infor-mation, lateral PFC damage tends to result in impair-ment. For example, in delayed recognition tests, PFCpatients are particularly sensitive to the presence ofdistracters during the delay. When compared topatients with damage to the MTL, PFC patientsshow impairments at all delay intervals as long as

8B

8Ad

8/w

44

9/46v

9/46d

96

4

6

10

47/12

45A 45B

Anterior VLPFC(pars orbitalis)

Mid-VLPFC(pars triangularis)

Posterior VLPFC(pars opercularis)

a b

(1)

y= 16

y= 30

(1)

(3)

(4)

Mid-VLPFC

Mid-VLPFC

Anterior VLPFC

(2)

46

Figure 1 Anatomical divisions of the ventrolateral prefrontal cortex (VLPFC) in the human: (a) representation of the cytoarchitectonicsubdivisions of the lateral prefrontal cortex (PFC); (b) coronal slices through the PFC depicting the anatomical boundaries that define themid-VLPFC and anterior VLPFC. The subregions of the VLPFC include the pars orbitalis, pars triangularis, and pars opercularis,corresponding to the anterior VLPFC (area 47/12), mid-VLPFC (area 45), and posterior VLPFC (area 44), respectively. (Note that weuse the term mid-VLPFC to refer to area 45 only, adopting anterior VLPFC to refer to area 47/12.) As shown in (b), both the anterior andmid-VLPFC lie ventral to the inferior frontal sulcus (1). In caudal slices, the mid-VLPFC is bounded ventrally by the insular sulcus (2), andin rostral slices it is bounded by the horizontal ramus of the lateral fissure (3). The anterior VLPFC is bounded ventrally and medially by theorbital gyrus (4). (a) Adapted from Petrides M and Pandya DN (2002) Comparative cytoarchitectonic analysis of the human and themacaque ventrolateral prefrontal cortex and corticocortical connection patterns in the monkey. European Journal of Neuoscience 16:291–310. (b) From Badre D and Wagner AD (2007) Left ventrolateral prefrontal cortex and the cognitive control of memory. Neuropsy-chologia 45(13): 2883–2901.

438 Strategic Control of Memory



distracters are present, whereas patients with MTLdamage show impairments principally at long delays.A similar observation comes from studies of PFC-lesioned monkeys. Whereas it was initially believedthat PFC-lesioned monkeys were unable to rememberthe location of a food item that was hidden for only ashort delay, it was later shown that this apparentmemory deficit could be eliminated if the delay periodwas held in total darkness. In other words, the mon-key’s memory deficit was a function of the interferingeffects of visual input during the delay period. Strik-ingly, not only do PFC patients show impaired work-ing memory performance due to distracters, but PFCpatients also show exaggerated electrophysiologicalresponses in cortical areas that process the modalityof the distracting information, suggesting that whendistraction is present, the PFC is responsible for gat-ing activity in regions that would otherwise processthe distracting information. Further evidence that lat-eral PFC damage results in an inability to filter or gateirrelevant distracting information comes from studiesof negative priming. In such studies, healthy controlstypically showed increased reaction times when pro-cessing a stimulus that was previously a distracter(i.e., these previous distracters showed negativepriming). By contrast, frontal patients showed thereverse effect – facilitated processing of stimuli thatpreviously served as distracters – which suggests thatthe patients had difficulty preventing the processingof the distracters.

Functional Specificity within LateralPrefrontal Cortex

Although data from frontal patients provide compel-ling evidence that one contribution of the lateralPFC to the strategic control of memory is to regulateinterference, the lack of anatomical specificity thattypically accompanies naturally occurring PFC lesionsin humans often precludes the determination of thespecific functional contributions of particular PFC sub-regions. In an attempt to specify the mechanistic con-tributions of the subregions of theVLPFC to the controlof memory, we consider in the next section evidencefrom functional neuroimaging studies and from recentreports of patients with lesions focused on subregionsof VLPFC. As we discuss, extant data implicate themid-VLPFC as a key component of the neural circuitrythat resolvesmnemonic interference, be itwithinwork-ing memory, semantic memory, or episodic memory.

Memory and the VLPFC

Focal lesion studies with monkeys have shown thatdamage to the inferior convexity (the monkey homo-log of the human inferior frontal gyrus), but not to the

DLPFC, causes perseverative tendencies in contextsin which the prior response learning must be reversed.Similar perseverative errors are a classic hallmarkof human frontal lobe damage, as is often revealedthrough errors on the Wisconsin Card SortingTask, when subjects must override a previously estab-lished, but now irrelevant, response set. As alreadynoted, the strategic control of working memory,semantic memory, and episodic memory oftenrequires conceptually similar mechanisms that permitthe overriding of interference from irrelevant repre-sentations. Functional neuroimaging data in humans,complemented by focal lesion evidence, now indicatethat the left mid-VLPFC regulates mnemonic interfer-ence by enabling the selection of relevant representa-tions in the face of mnemonic conflict, whereas theleft anterior VLPFC controls access to, and guidesthe retrieval of, long-term semantic knowledge.Response override or inhibition, by contrast, appearsto differentially depend on regions in the human rightVLPFC.

Left Mid-VLPFC and Interference in WorkingMemory

Extensive evidence for the role of the left mid-VLPFC(BA 45) in resolving mnemonic interference comesfrom a variant of the Sternberg working memoryparadigm (Figure 2(a)). In this task, participantsencode a target set of stimuli (e.g., four letters),which they then attempt to maintain in workingmemory across a brief delay. Following the delay, aprobe (e.g., a letter) is presented and participantsmake a yes/no decision as to whether the probe isin the currently maintained memory set. Critically,some negative probes (i.e., probes to which the sub-ject should respond no) come from the immediatelypreceding memory set (negative recent probes),whereas other negative probes have not appearedin either the present or the preceding memory sets(negative nonrecent probes) (Figure 2(a)). Given thisstructure, negative recent probes are associated withinterference because subjects must attribute the famil-iarity of the probe to its having been in the precedingmemory set to correctly reject the probe as not beinga member of the current memory set. Thus, compar-isons of negative recent to negative nonrecent trialsprovide leverage on the neural mechanisms thatare engaged in the face of this mnemonic interference.

Jonides and colleagues were the first to demon-strate that functional activation in the left mid-VLPFC is greater during negative recent relative tonegative nonrecent trials, a pattern that has beenreplicated and extended by others (Figure 2(b)).Subsequent data indicated that (1) the response inthe left mid-VLPFC during negative recent trials is




restricted to the probe period of the trial, whichis consistent with the fact that interference is a func-tion of the probe’s familiarity, and (2) when directlyinterrogating the DLPFC, a similar difference betweennegative recent and nonrecent trials is not typicallyobserved, suggesting a specific role of the leftmid-VLPFC in resolving interference in this task.Moreover, in a compelling study motivated by theaforementioned neuroimaging data, Thompson-Schilland colleagues showed that PFC patients with damagethat spared the left VLPFC demonstrated interferenceeffects that were comparable in magnitude to those inhealthy controls, whereas one patient, R.C., with afocal lesion that damaged almost the entire extent ofthe left BA 45, exhibited an exacerbated interferenceeffect despite relatively normal working memory per-formance when interference was minimal. Comple-menting these data, Postle and colleagues observedthat when repetitive transcranial magnetic stimulation(TMS) – which transiently disrupts cortical function –was applied to the left VLPFC during the probe periodin healthy humans there was a similar increase in sus-ceptibility to interference to that seen in patient R.C.

By contrast, TMS disruption of the primary motor,primary somatosensory, or supplemental motor areadid not affect the ability to resolve interference in thistask. Collectively, these data demonstrate that the leftmid-VLPFCmakes necessary contributions to resolvinginterference from currently irrelevant information inworking memory. A leading hypothesis is that thisregion enables the selection of relevant active informa-tion over irrelevant active representations.

Left VLPFC and Declarative Memory Retrieval

Complementing the working memory literature im-plicating the left mid-VLPFC in resolving interferencefrom recently activated but now irrelevant mnemonicrepresentations are patient and imaging data thatsuggest that the left mid-VLPFC also supports selec-tion during retrieval from declarative memory (i.e.,semantic and episodic memory). For example, asalready discussed, demands on interference resolu-tion during semantic retrieval may be greater whenavailable retrieval cues are unconstrained, and it isunder these conditions that left PFC patients showimpairments in generating semantic knowledge.

Negative nonrecent

+ g+a

a

b

de

f

+ c+d ba

c

+ b+a

Trial N

Trial N−1

de

f

Negative recent

L mid-VLPFC

Sig

nal c

hang

e (%

) −4

−3

−2

−1

0

Nonrecent

Recent

Figure 2 Interference resolution in the Sternberg working memory task: (a) a representative version of the task; (b) representative fMRIdata. As shown in (a), during the task subjects maintain a set of four letters in working memory until a probe letter appears, at which pointsubjects indicate whether the probe is a member of the currently maintained set (positive probe) or is not a member of the current set(negative probe). Interference occurs when a probe is not a member of the current set, but is a member of the immediately precedingset (negative recent probe), relative to situations in which the probe is a member neither of the current set nor of the immediatelypreceding set (negative nonrecent probe). In (b), the fMRI data reveal greater activation during negative recent vs. negative nonrecenttrials in the left mid-VLPFC (circled), reflecting that the presence of interference in workingmemory is accompanied by greater recruitmentof the left mid-VLPFC. fMRI, functional magnetic resonance imaging; L, left; VLPFC, ventrolateral prefrontal cortex. Adapted fromBadre D and Wagner AD (2005) Frontal lobe mechanisms that resolve proactive interference. Cerebral Cortex 15: 2003–2012.




Such impairments may reflect an inability to selectrelevant representations from among competingalternatives.Functional neuroimaging and focal lesion data

converge on the left mid-VLPFC as being centralfor mediating the selection of relevant semanticrepresentations. For example, in an influential func-tional magnetic resonance imaging (fMRI) study,Thompson-Schill and colleagues compared semanticprocessing under conditions that varied the demandsplaced on selection. Across several tasks, low-selectionconditions were constructed such that they involvedretrieving dominant, or prepotent, semantic infor-mation, whereas high-selection conditions involvedretrieving nondominant semantic information fromamong competing representations. Significantly, theimaging data revealed that the left VLPFC (BAs 44and 45) was consistently more active during thehigh-selection (vs. low-selection) conditions, demon-strating that this is a replicable characteristic of leftmid-and/or posterior VLPFC function. Subsequentneuropsychological work by Thompson-Schill andcolleagues revealed that it is the proportion of dam-age to this left VLPFC region, but not the grosslesion size, that strongly predicts the magnitude ofbehavioral impairment on high-selection semanticretrieval tasks.Related data have increased the precision of our

understanding of left mid-VLPFC function and itsrelation to the functions of the surrounding anteriorand posterior VLPFC subregions. In particular, Badreand colleagues demonstrated that selection demandsduring semantic processing tasks are specificallyassociated with the same left mid-VLPFC subregion(BA 45; pars triangularis) that resolves interferencein working memory tasks (Figure 3(a)). Moreover,these researchers observed a functional dissocia-tion within the left VLPFC, indicating that the leftmid-VLPFC supports selection from among activerepresentations, whereas the more rostrally and ven-trally situated left anterior VLPFC subregion (area47/12; pars orbitalis) controls the retrieval of seman-tic representations stored in the lateral temporalcortical areas. This functional distinction betweenthe left mid-VLPFC and anterior VLPFC appearsreplicable and generalizeable; Gold and colleaguesreported a similar functional pattern using a lexicaldecision task – namely that the left anterior VLPFCwas associated with controlled semantic retrieval andthe left mid-VLPFC was associated with resolvinginterference from irrelevant active representations.Evidence for functional segregation within the

left VLPFC has also come from studies comparingsemantic versus phonological control. For example,neuroimaging studies of phonological rehearsal and

phonological analysis of stimuli have implicated theleft posterior VLPFC (BA 44; pars opercularis), sug-gesting that this VLPFC subregion is critical forrepresenting and maintaining phonological codes.On the other hand, neuroimaging studies of semanticretrieval and analysis of stimuli have implicated theleft anterior VLPFC. Significantly, these dissociationsbetween the left anterior and posterior VLPFC havebeen observed in studies that directly contrastedtasks that differentially depend on semantic andphonological control. Moreover, Devlin and collea-gues have shown that TMS to the left anterior VLPFCdifferentially disrupts concurrent semantic proces-sing, whereas TMS to the left posterior VLPFC differ-entially disrupts concurrent phonological processing,demonstrating that these subregions are necessary fordistinct forms of control.

Functional distinctions between the left VLPFCsubregions are also apparent during tasks that probeepisodic memory. For example, Dobbins and collea-gues observed that the left mid-VLPFC is associatedwith selecting between episodic details in the courseof making a source judgment (Figure 3(b)), whereasthe left anterior VLPFC is associated with semanticallyelaborating on the cues used to probe episodicmemory.Further, during episodic encoding, Dolan, Fletcher,and colleagues demonstrated that left mid-VLPFCactivity increases when prior learning interferes withto-be-encoded information (i.e., when the resolution ofproactive interference is required). Collectively, thesedata suggest that the left mid-VLPFC contributes to theresolution of interference during all forms of declara-tive memory, whereas the left anterior VLPFC controlsthe retrieval of semantic knowledge that is not retrievedin an automatic (bottom-up) manner. Moreover,when considered along with the evidence that theleft mid-VLPFC resolves interference within workingmemory, extant data favor the hypothesis that the leftmid-VLPFC serves a domain-general role of resolvingmnemonic interference.

Right VLPFC and the Control of Memory

We have thus far focused on how the left VLPFC con-tributes to interference resolution and other forms ofcognitive control; however, it should be emphasizedthat the right VLPFC is sometimes coactive withthe left VLPFC during tasks that require overcomingmnemonic conflict. Nevertheless, although the rightVLPFC may support control functions that are con-ceptually analogous to those supported by the leftVLPFC, it appears that (in the human) the rightVLPFC at least partially differs from the left VLPFCin the domain of knowledge on which is operates.For example, Aron and colleagues, among others,




Left mid-VLPFC

Sem

antic

inte

rfer

ence

0.0Conceptual

detailPerceptual

detailItem

recognition

0.1−2

−0.2

0

R= 0.76

0.2

0.4

−1 0

0.400.300.200.100.00

−0.10−0.20

0.300.250.20

0.100.15

0.050.00

−0.05−0.10

0.100.150.20

0.050.00

−0.05

0.10

0.15

0.05

0.00

−0.05

−0.10

Selection

L mid-VLPFC

1 2

0.2

Sig

nal c

hang

e (%

) 0.3

0.4

a b

1. L anterior VLPFC 2. Middle temporal cortex

Conceptual detail

Perceptual detail

Item recognition

3. R posterior VLPFC 4. Inferior temporal cortex

c

Conceptual recollection Perceptual recollection

Figure 3 PFC contributions to the control of declarative memory: (a) a behavioral selection measure derived from performance during various semantic retrieval tasks; (b) the left mid-VLPFC during episodic retrieval; (c) the left anterior and right posterior VLPFC during episodic retrieval. In (a), the selection measure was exclusively associated with the magnitude ofactivation in the left mid-VLPFC (left). For example, this selection measure was tightly correlated with the magnitude of interference-related activation in the left mid-VLPFC during semanticretrieval (right). As shown in (b), during episodic retrieval, the left mid-VLPFCwas differentially engaged when source details (conceptual or perceptual) were selectively retrieved, compared tosimple item recognition. As shown in (c), during episodic retrieval, the left anterior VLPFC (1) was selectively engaged during recollection of conceptual source details, functionally couplingwith the middle temporal cortex (2). By contrast, the right posterior VLPFC (3) was engaged during perceptual recollection and during recognition of familiar vs. novel objects, functionallycoupling with the inferior temporal cortex (4). L, left; PFC, prefrontal cortex; VLPFC, ventrolateral prefrontal cortex. (a) Adapted from Badre D andWagner AD (2005) Frontal lobe mechanismsthat resolve proactive interference. Cerebral Cortex 15: 2003–2012. (b, c) Adapted from Dobbins IG and Wagner AD (2005) Domain-general and domain-sensitive prefrontal mechanisms forrecollecting events and detecting novelty. Cerebral Cortex 15: 1768–1778.

442Strate

gic

Contro

lofMemory



reported that functional activation in the posterioraspect of the right VLPFC is consistently implicatedin situations in which prepotent motor responses mustbe inhibited or when well-learned stimulus–responsecontingencies must be reconfigured. Moreover, theseresearchers demonstrated that when the right VLPFCis damaged, subjects display increased susceptibilityto interference at the response level, highlightingthe necessary contribution of this VLPFC subregion toresponse inhibition.By contrast, other data indicate that the right

VLPFCmay contribute to the orienting of visual atten-tion toward task-relevant object representations (andaway from distracters). Significantly, with respect tothe control of memory, this latter role of the rightVLPFC in orienting to visuo-object representationshas been observed in a number of mnemonic contexts,including discriminating novel fromencountered visualobjects, encoding faces and complex visual scenes intoepisodic memory, recollecting object–object associa-tions, and recollecting perceptual details about pre-viously encountered objects (Figure 3(c)). Thus, theextent to which the right VLPFC (as opposed to theleft VLPFC) is engaged at least partially dependson the modality of the mnemonic representationsattended,with the right VLPFC preferentially engagedduring visuospatial processing and the left VLPFCengaged during phonological and semantic processing(Figure 3(c)). Significantly, although the right VLPFChas been implicated in visuo-object attention and inresolving competition at the response level, functionalunderstanding of the subregions within the rightVLPFC is not as well advanced as that of the homolo-gous structures in the left VLPFC.

Concluding Comments

Interference resolution is a fundamental aspect ofefficient mnemonic processing and may represent aprincipal way in which cognitive control interactswith memory. This interaction is critically enabledby control mechanisms that depend on the lateralPFC for their operation. This article focuses on thedelineation of the mechanisms supported by theVLPFC, but it should be noted that the VLPFC mustdynamically interact with the posterior neocorticalassociation areas and with the MTL to implementthe control of memory. Further, although VLPFCfunctions are clearly important for the resolution ofmnemonic interference, the VLPFC also interactswith the DLPFC and frontopolar cortical areas thatsubserve other forms of cognitive control. Next webriefly illustrate each of these points.To regulate memory-relevant processes, PFC control

mechanismsmust interact with systems that support or

store long-termmnemonic representations. One line ofevidence for such frontal-posterior interactions comesin the form of coactivations within functional neuro-imaging studies. For example, fMRI studies of episodicremembering indicate that the left anterior VLPFCcoactivates with the left middle temporal corticalareas that represent conceptual knowledge during con-trolled attempts to recollect such details about the past(Figure 3(c)). Analogously, the right VLPFC coactiveswith bilateral occipito-temporal areas that representvisuo-object form during attempts to recollect suchdetails (Figure 3(c)). Moreover, as previously alludedto, lesions to the lateral PFC result in the failure toregulate processing in such posterior neocorticalareas, demonstrating that the lateral PFC is necessaryfor gating the processing of distracting or interferingstimuli. Collectively, these data illustrate the top-downrole of PFC control mechanisms in regulating percep-tion, memory, and action.

With respect to cross-regional interactions withinthe PFC, it is important to emphasize that other struc-tures in the DLPFC and frontopolar cortex are alsooften engaged in situations in which control must beimplemented to accomplish a mnemonic goal. Forexample, DLPFC activation is frequently observedduring complex working memory tasks, duringattempts to remember past events, and during manytasks that require response selection or the resolutionof response conflict. Although the specific nature ofinteractions between the DLPFC and VLPFC have yetto be well characterized, it is hypothesized that theDLPFCmay operate at a higher stage in the processinghierarchy relative to the VLPFC, serving to performoperations on the products of VLPFC processing.

Thus, it is clear that the strategic control of mem-ory is multifaceted. Future research promises tofurther illuminate how cognitive control emergesthrough lateral PFC function, allowing us to strategi-cally wrest control of our memories, bringing themin line with current goals.

See also: Cognition: An Overview of NeuroimagingTechniques; Event-Related Potentials (ERPs); InhibitoryControl over Action and Memory; Neuroimaging;Prefrontal Cortex: Structure and Anatomy; PrefrontalCortex; Short Term and Working Memory; SpatialCognition and Executive Function; Working Memory:Capacity Limitations.

Further Reading

Aron AR, Robbins TW, and Poldrack RA (2004) Inhibition and theright inferior frontal cortex. Trends in Cognitive Sciences 8:170–177.

Badre D, Poldrack RA, Pare-Blagoev EJ, Insler RZ, andWagner AD(2005) Dissociable controlled retrieval and generalized selection




mechanisms in ventrolateral prefrontal cortex. Neuron 47:907–918.

Badre D and Wagner AD (2005) Frontal lobe mechanisms thatresolve proactive interference. Cerebral Cortex 15: 2003–2012.

Badre D and Wagner AD (2007) Left ventrolateral prefrontal cor-tex and the cognitive control of memory. Neuropsychologia45(13): 2883–2901.

Dobbins IG and Wagner AD (2005) Domain-general and domain-sensitive prefrontal mechanisms for recollecting events anddetecting novelty. Cerebral Cortex 15: 1768–1778.

Dolan RJ and Fletcher PC (1997) Dissociating prefrontal and hip-pocampal function in episodic memory encoding. Nature 388:582–585.

Feredoes E, Tononi G, and Postle BR (2006) Direct evidence for aprefrontal contribution to the control of proactive interferencein verbal working memory. Proceedings of the NationalAcademy of Sciences of the United States of America 103:19530–19534.

Gold BT, Balota DA, Jones SJ, Powell DK, Smith CD, andAnderson AH (2006) Dissociation of automatic and strategiclexical-semantics: Functional magnetic resonance imaging evi-dence for differing roles of multiple frontotemporal regions.Journal of Neuroscience 26: 6523–6532.

Gough PM, Nobre AC, and Devlin JT (2005) Dissociating linguis-tic processes in the left inferior frontal cortex with transcranialmagnetic stimulation. Journal of Neuroscience 25: 8010–8016.

Henson RN, Shallice T, Josephs O, and Dolan RJ (2002) Functionalmagnetic resonance imaging of proactive interference duringspoken cued recall. Neuroimage 17: 543–558.

Iversen SD and Mishkin M (1970) Perseverative interference inmonkeys following selective lesions of the inferior prefrontalconvexity. Experimental Brain Research 11: 376–386.

Jonides J, Smith EE, Marshuetz C, Koeppe RA, and Reuter-LorenzPA (1998) Inhibition in verbal working memory revealed by

brain activation. Proceedings of the National Academy ofSciences of the United States of America 95: 8410–8413.

Martin RC and Cheng Y (2006) Selection demands versus associa-tion strength in the verb generation task. Psychonomic Bulletin& Review 13: 396–401.

Moscovitch M and Melo B (1997) Strategic retrieval and the fron-tal lobes: Evidence from confabulation and amnesia. Neurop-sychologia 35: 1017–1034.

Petrides M and Pandya DN (2002) Comparative cytoarchitectonicanalysis of the human and the macaque ventrolateral prefrontalcortex and corticocortical connection patterns in the monkey.European Journal of Neuoscience 16: 291–310.

Shimamura AP (2000) The role of the prefrontal cortex in dynamicfiltering. Psychobiology 28: 207–218.

Shimamura AP, Jurica PJ, Mangels JA, Gershberg FB, and KnightRT (1995) Susceptibility to memory interference effectsfollowing frontal lobe damage: findings from tests of paired-associate learning. Journal of Cognitive Neuroscience 7:144–152.

Thompson-Schill SL, D’Esposito M, Aguirre GK, and Farah MJ(1997) Role of left inferior prefrontal cortex in retrieval ofsemantic knowledge: A reevaluation. Proceedings of theNational Academy of Sciences of the United States of America94: 14792–14797.

Thompson-Schill SL, Jonides J, Marshuetz C, et al. (2002) Effectsof frontal lobe damage on interference effects in workingmemory. Cognitive, Affective, & Behavioral Neuroscience 2:109–120.

Thompson-Schill SL, Swick D, Farah MJ, D’Esposito M, Kan IP,and Knight RT (1998) Verb generation in patients with focalfrontal lesions: A neuropsychological test of neuroimaging find-ings. Proceedings of the National Academy of Sciences of theUnited States of America 95: 15855–15860.




strategic control of memory - kuhllab.com

Documents