the effects of prefrontal lesions on working memory ... · test the effects of prefrontal lesions....

Copyright 2004 Psychonomic Society Inc 528

Cognitive Affective amp Behavioral Neuroscience2004 4 (4) 528-539

The effects of experimental lesions of the monkey pre-frontal cortex (PFC) have played a predominant role incurrent conceptualizations of prefrontal function andmost notably of working memory Working memory al-lows organisms to use information that is not currentlypresent in the environment but is vital to adaptive be-havior Current conceptualizations include both a main-tenance component akin to short-term memory and anldquoexecutiverdquo component that includes various control pro-cesses that in some way transform maintained represen-tations into a more usable form The loss or preservationof certain working memory abilities has been shown tobe attributable to differences in the specific requirementsof behavioral testing (eg spatial vs nonspatial delayvs no delay) along with differences in the specific loca-tions of applied ablations (eg dorsal vs ventral PFC)Such findings which have accumulated now for over acentury have led to widespread acceptance that the dor-solateral (BA 46 946 and 9) and ventrolateral (BA 45

and 4712) aspects of the PFC (see Figure 1) may per-form different specialized roles in higher order cogni-tion1 Nonetheless it remains unclear and controversialhow the lateral PFC is functionally organized

Two main views propose different types of functionalspecialization for the dorsal and ventral PFC The firstthe material-specific model contends that the lateralPFC is segregated according to the processing of spatialand nonspatial domains of information (Levy amp Goldman-Rakic 2000) The dorsolateral PFC is engaged in ldquoon-linerdquo maintenance of spatial memoranda while the ven-trolateral PFC supports nonspatial (eg face objects)memoranda In essence the segregation of the dorsalldquowhererdquo and ventral ldquowhatrdquo visual pathways in posteriorextrastriate parietal and temporal cortices are proposedto be conserved in the lateral PFC This appealing hy-pothesis is founded on the relatively greater proportionof projections from the dorsal stream to the dorsolateralPFC and similarly the greater proportion of projectionsfrom the ventral stream to the ventrolateral PFC (see Fig-ure 1) The second process-specific model contends thatdomain specialization is not the key to the organizationof the PFC This model does not deny that material or do-main specificity exists in the PFC it simply attempts todescribe another functional topography To this end thedorsal and ventral prefrontal cortices are proposed to per-

This work was supported by grants from the National Institutes ofHealth and the James S McDonnell Foundation to CEC Specialthanks to Gregory M Weiner for illustrating the monkeys in Figure 2Correspondence concerning this article should be addressed to C E Cur-tis Department of Psychology Center for Neural Science New YorkUniversity 6 Washington Place Room 859 New York NY 10003 (e-mail claytoncurtisnyuedu)

The effects of prefrontal lesions on workingmemory performance and theory

CLAYTON E CURTISNew York University New York New York

and

MARK DrsquoESPOSITO University of California Berkeley California

The effects of experimental lesions of the monkey prefrontal cortex have played a predominant rolein current conceptualizations of the functional organization of the lateral prefrontal cortex especiallywith regard to working memory The loss or sparing of certain performance abilities has been shownto be attributable to differences in the specific requirements of behavioral testing (eg spatial vs non-spatial memoranda) along with differences in the specific locations of applied ablations (eg dorsalvs ventral prefrontal cortex) Such findings which have accumulated now for over a century have ledto widespread acceptance that the dorsolateral and ventrolateral aspects of the prefrontal cortex mayperform different specialized roles in higher order cognition Nonetheless it remains unclear and con-troversial how the lateral prefrontal cortex is functionally organized Two main views propose differ-ent types of functional specialization of the dorsal and ventral prefrontal cortex The first contends thatthe lateral prefrontal cortex is segregated according to the processing of spatial and nonspatial do-mains of information The second contends that domain specialization is not the key to the organiza-tion of the prefrontal cortex but that instead the dorsal and ventral prefrontal cortices perform qual-itatively different operations This report critically reviews all relevant monkey lesion studies that haveserved as the foundation for current theories regarding the functional organization of the prefrontalcortex Our goals are to evaluate how well the existing lesion data support each theory and to enu-merate caveats that must be considered when interpreting the relevant literature

PREFRONTAL LESIONS AND WORKING MEMORY 529

form qualitatively different operations (Petrides 2000b)The process-specific model proposes a hierarchy in whichthe ventrolateral PFC supports processes such as the ac-tive encoding and retrieval of information and the mid-dorsolateral PFC supports higher order ldquoexecutiverdquo con-trol functions like the monitoring and manipulation ofstored information This report critically reviews themonkey lesion studies that have provided the foundationfor current theories regarding the functional organizationof the PFC

History

Hitzig (1874) and Ferrier (1886) described over 100years ago the effects of experimental ablations of the

PFC Early naturalistic observations made by these ex-perimental psychologists led them to conclude that thefrontal cortex is particularly concerned with intellectualand attentional rather than sensory and motor functionsLater more extensive comparative studies and structuredbehavioral methods were used to investigate the effectsof frontal lesions The Italian physiologist Bianchi (1922)concluded that removal of the frontal lobes destroyed theability of the animal to synthesize various incoming per-cepts and integrate these with outgoing motor commandsIn his view the frontal cortex was the top of the sensoryand motor hierarchies and was the final stage of integra-tion The American psychologist Franz (1907) unlikehis predecessors used structured and objective behav-ioral testing methods (eg Thorndike boxes and basicvisual discrimination problems) to study the effects offrontal lobe removal Unfortunately though the behav-ioral tests he chose proved inadequate for capturing thetypes of mental processes that are impaired after frontallesions Up to this point Hitzig Ferrier Bianchi andFranz had produced intriguing but still somewhat im-pressionistic work that failed to establish what role thePFC might play in high-level cognition What was clearlylacking from their studies was a means for reliably andflexibly assessing the behavioral effects of prefrontal le-sions the pioneering researchers had noted these effectsbut were as yet unable to capture them In the hands ofJacobsen (1936) Hunterrsquos (1913) delayed-response task(Figure 2) proved to be a selective index of primate pre-frontal deficit Jacobsenrsquos highly influential monographdetailed the delayed-response and visual discriminationcapabilities of monkeys that had been given prefrontalpremotor or temporal lobe lesions He demonstrated thatonly prefrontal monkeys showed a selective and delay-dependent deficit on delayed-response tasks Delayed-response tasks depend upon the animalrsquos ability to main-tain an internal representation over the course of a delayAt the time of choice this internal representation mustguide behavior because no features present in the envi-ronment signal the correct response With these ideas inmind Jacobsen concluded that the deficit caused by pre-frontal cortical damage was mnemonic in nature specif-ically the monkeyrsquos ability to use ldquoimmediate memoryrdquoto guide behavior was impaired by the prefrontal lesion

The Effects of Prefrontal Lesions on thePerformance of Delayed-Response Tasks

Following Jacobsenrsquos (1936) pivotal demonstration ofimpaired delayed-response task performance after pre-frontal ablations several decades of investigations haveattempted to identify exactly which area(s) of the PFCare necessary for delayed-response performance (ie lo-calization) and which mental process or processes nec-essary for delay-response performance are impaired bydamage to the PFC (ie functional specification)

Figure 1 Subdivisions of the macaque prefrontal cortex as de-fined by Petrides and Pandya (1994) on the basis of connectivityand cytoarchitecture The mid-dorsolateral PFC is composed ofareas 46 and 946d and the mid-ventrolateral PFC of areas 471245A and 946v Note that many of the lesion studies reviewed hereused Walkerrsquos earlier cytoarchetectonic map (Walker 1940) inwhich Petrides and Pandyarsquos (1994) subdivisions 46 946d and946v are combined into a simpler area 46 Also depicted are thedorsal and ventral streams that respectively process spatial andnonspatial visual information (Ungerleider amp Mishkin 1982)The dorsal stream has relatively greater dorsolateral PFC pro-jections and the ventral stream has relatively greater ventrolat-eral PFC projections (Petrides amp Pandya 2002 UngerleiderCourtney amp Haxby 1998) Projections from the posterior pari-etal cortex (PG) primarily target 8A and to a lesser degree 946das depicted by the thickness of the line Projections from the in-ferior temporal cortex (IT) primarily target 4712 and to a lesserextent 45 The dashed line represents the principal sulcus whichdelineates the dorsal and ventral aspects of the prefrontal cor-texrsquos lateral convexity

8Ad946d44

8Av946v45A

8B4645B

9104712

530 CURTIS AND DrsquoESPOSITO

Localization I The Importance ofthe Principal Sulcus

Earlier studies illustrating that an intact PFC was nec-essary for delayed-response task performance typicallymade large lesions of the entire lateral prefrontal cortexoften including the medial and orbital surfaces as wellAs first demonstrated by Blum (1952) an equally dis-abling deficit on delayed-response tasks can be producedby more focal lesions of just the ldquomidlateralrdquo portion ofthe PFC In fact small circumscribed lesions of just theprincipal sulcus (ie areas 46 and 946d) can produce

deficits as great as larger total prefrontal lesions (But-ters amp Pandya 1969 Butters Pandya Stein amp Rosen1972 Goldman amp Rosvold 1970 Goldman RosvoldVest amp Galkin 1971 Gross amp Weiskrantz 1962 Mishkin1957) Lesions of the periacuate (ie areas 8Ad and 8Av)superior frontal (ie areas 8B and 9) orbital (ie areas 11and 12) premotor (ie areas 6 8 and 44) inferior tem-poral and parietal cortex do not typically cause signifi-cant or long-lasting impaired delayed-response perfor-mance (Butters amp Pandya 1969 Goldman amp Rosvold1970 Goldman et al 1971 Jacobsen 1936 Rosenkilde

Food

Reward

Correct

Choice

Cue

Delay

Response

Correct

Choice

Cue

Delay

Response

Cue

Delay

Response

Delay

Response

Delayed

Response

Delayed

Matching-to-Sample

Self

Ordered

Figure 2 Schematic depictions of working memory tasks used totest the effects of prefrontal lesions The spatial delayed-responsetask (left) is a test of spatial working memory and requires the mon-key to remember the baited location of a food well over a delay Ona variant of this task (not shown) called delayed spatial alternationthe monkey selects the location not chosen on the immediate pasttrial The delayed matching-to-sample task (middle) is a nonspatialworking memory task and requires the monkey to remember anobject over a delay After the delay only the well beneath the sam-ple object is baited A common variant (not shown) is the delayednonmatching-to-sample task in which the monkey remembers thesample object but after the delay must select an object that does notmatch the sample objectrsquos form Another nonspatial working mem-ory variant called delayed object alternation (not shown) is similarto spatial alternation but requires the monkey to alternatively se-lect the object not selected on the past trial On a self-ordered task(right) a set of objects is shown and on each trial the monkey mustselect one object until they have all been selected


Rosvold amp Mishkin 1981) Importantly lesions of theinferior convexity or ventrolateral PFC (ie areas 471245A and 45B) which are discussed below do not pro-duce substantial or long-lasting deficits in delayed-response performance (Passingham 1975) Thus by theearly 1970s a fair amount of evidence from lesion analy-ses implicated the principal sulcus as the most necessarystructure required for normal performance of delayed-response and delayed-alternation tasks (see Figure 2)The nature of the deficit caused by lesions of the princi-pal sulcus however was far from being understood

Functional SpecificationSimultaneous with probes into which part of the PFC

supported delayed-response task performance investi-gators began asking exactly what cognitive processeswere being measured Were the deficits domain-specific(eg spatial vs nonspatial memoranda) or modality-specific (eg visual vs auditory cues) Delayed-responsetasks are complex and failure can result from a multi-tude of impaired processes Were the deficits most con-sistent with impaired stimulus discrimination Did theyseem attributable to alterations in motor control Or asJacobsen (1936) speculated were they attributable toimpairments in the ability to maintain an internal repre-sentation over a period of time To seek answers varia-tions on the classic delayed-response task were conceived

The importance of space and time In the classicalversions of the delayed-response and delayed-alternationproblems an animal must use spatial factors to solvethem correctly For the delayed-response task the spatialposition of the baited well must be remembered over adelay Similarly for the delayed-alternation task the spa-tial position of the well chosen on the last trial must beremembered Besides their spatial position the two wellsare otherwise indistinguishable (eg same color shapeetc) We have already seen that prefrontal lesions im-pair performance on such spatial delayed-response anddelayed-alternation tasks To address whether the samepattern would emerge if the animal must rely on nonspa-tial information Mishkin and Pribram conducted a seriesof studies (Mishkin 1954 Mishkin amp Pribram 1954)They trained monkeys to perform delayed-response anddelayed-alternation tasks that were modified so that mon-keys were required to use nonspatial information to per-form correctly Remarkably the monkeys with prefrontallesions were able to perform these versions normally butstill performed poorly on tasks for which spatial infor-mation was necessary Therefore the deficit followingprefrontal damage appeared limited to the spatial do-main However Mishkin and Pribram then discoveredthat monkeys with prefrontal lesions were just as im-paired on a delayed object alternation task as on the spa-tial version (Mishkin amp Pribram 1955 1956 Pribram ampMishkin 1956) calling into question the domain speci-ficity of the lesion For the object alternation task ani-mals alternately chose between two easily distinguish-

able objects on each trial the positions of the objectswere random and did not guide correct performance Ina followup study with prefrontal lesions limited to thedorsal and sparing the ventral lateral prefrontal surfacethere was a marked deficit on spatial but not object delayed-alternation tasks (Mishkin Vest Waxler amp Rosvold 1969)this effect was later replicated by Petrides (1995) Thusdamage to the dorsolateral PFC caused alternation defi-cits when spatial but not object information must be usedto guide behavior

In a pair of influential studies Goldman-Rakic and col-leagues (Goldman amp Rosvold 1970 Goldman et al 1971)presented evidence that spatial factors are necessary butnot sufficient to cause deficits in monkeys with principalsulcus lesions the imposition of a delay between cue andresponse is also necessary Monkeys with lesions of theprincipal sulcus could not perform spatial delayed-response(Goldman et al 1971) or spatial delayed-alternationtasks (Goldman amp Rosvold 1970) However these mon-keys performed normally on a conditional position re-sponse task that did not impose an intratrial delay Themonkeys were rewarded for choosing a left or right boxdepending on a sound that originated from above orbelow the chamber (Goldman amp Rosvold 1970) Thistask contained a spatial component but no delay Addi-tionally in a study that replicated the findings of Mishkinand Pribram (1954) lesions of the principal sulcus didnot cause impairments in a delayed gono-go alternationtask which contained a delay but not a spatial component(Goldman et al 1971) The researchers argued that inorder for a behavioral task to be sensitive to lesions of theprincipal sulcus it must necessarily contain both a spatialand a delay factor rather than either of these factors alone

Other compelling ways have also been found to dem-onstrate that the presence of a delay is crucial For in-stance instead of comparing performance on differenttasks that have or do not have delays the same task canbe used and performance can be assessed for varying de-lays M H Miller and Orbach (1972) demonstrated thatmonkeys with dorsolateral PFC lesions show a delay-dependent spatial alternation deficit The monkeys canperform spatial alternation tasks when there is no delayshow mild deficits at a minimal 1-sec delay and show pro-nounced deficits when the delay is lengthened to 5-secSimilar delay-dependent effects on delayed-responsetasks have been observed after cooling of the PFC (Baueramp Fuster 1976 Fuster amp Alexander 1970)

Most notably using an eight-position oculomotordelayed-response task with restrained monkeys Funa-hashi Bruce and Goldman-Rakic (1993) were able todemonstrate convincingly that lesions of the principalsulcus interfere in a delay-dependent manner with theaccuracy of saccades made to remembered locations ofbriefly presented cues in the contralesional hemifieldThe deficits did not appear to be strictly sensory or motorin nature since the monkeys had no difficulties detectingvisual cues and making accurate saccades to them as


long as they were still visible In fact significant deficitswere rarely noted at a brief delay of 15 sec but becameprogressively worse as the delay was lengthened show-ing a delay dependency Funahashi and colleagues hadpreviously shown that directionally selective delay-periodneuronal activity as measured by unit recordings of cel-lular activity in the principal sulcus in awake-behavingmonkeys performing oculomotor delayed-response tasksmay represent maintenance of the spatial cue over thedelay (Funahashi Bruce amp Goldman-Rakic 1989 19901991) Also notably Funahashi et al (1993) reported forthe first time impairments when only one PFC hemi-sphere was lesioned Past studies done in the WisconsinGeneral Testing Apparatus a testing chamber in whichthe monkey was free to move its head during the testinghad not found significant impairments unless both PFChemispheres were damaged However with the monkeyrsquoshead restrained and eye position controlled Funahashiet al (1993) were able to demonstrate that unilateral le-sions were sufficient to cause contralesional impairedspatial working memory performance (see Figure 3) To-gether these data strongly link the principal sulcus to themaintenance of contralateral spatial information

Non-spatial working memory performance afterdorsolateral prefrontal lesions In the 1970s Fusterand his colleagues began experimenting with localizedcryogenic depressing or ldquocoolingrdquo of cortical tissue as ameans of inducing a reversible lesion These investiga-tors demonstrated that cooling of the principal sulcusimpaired performance of not only spatial delayed-response tasks (Bauer amp Fuster 1976 Fuster amp Alexan-der 1970 Fuster amp Bauer 1974) but also nonspatial de-layed matching-to-sample tasks (Bauer amp Fuster 1976

Quintana amp Fuster 1993) These data argue against thehypothesis that the principal sulcus supports short-termmemory for spatial information only and suggests that ithas a more general role in working memory To bolsterthis claim Fuster and others have cited unit recordingstudies in which delay period activity of neurons in theprincipal sulcus is just as great during spatial as duringnonspatial working memory tasks (Fuster Bauer amp Jer-vey 1982 Quintana amp Fuster 1992 Quintana Yajeyaamp Fuster 1988 Rao Rainer amp E K Miller 1997)

Other investigations around the same time showed thatlesions limited to the principal sulcus do not cause non-spatial delayed alternation or delayed matching-to-sampledeficits (Mishkin amp Manning 1978 Mishkin et al 1969Passingham 1975 Stamm 1973 Stamm amp Weber-Levine 1971) but lesions ventral to the principal sulcusdo cause such impairments severely (Mishkin amp Manning1978 Passingham 1975) Moreover further studies haveconfirmed that lesions of the dorsolateral PFC do not causeimpaired nonspatial working memory (Bachevalier ampMishkin 1986 Levy amp Goldman-Rakic 1999 Petrides1991 1995 2000a) How do we reconcile these discrep-ancies Some investigators attributed the nonspatial work-ing memory deficits noted after cooling of the principalsulcus to a spreading of the cooling to adjacent prefrontalcortical tissue Although cryogenic depression does resultin the dampening of synaptic activity the effects are notvery localizable and there is even a halo zone around thecooled center that can exhibit hyperactive synaptic ac-tivity (Volgushev Vidyasagar Chistiakova amp Eysel2000) This spatial uncertainty combined with recentstudies that show that dorsolateral PFC lesions do notcause impaired maintenance of object stimuli (Levy amp

90o

270o

0o

180o

315o

225o

135o

45o

50

25

0

15-sec delay

30-sec delay

60-sec delay

A Oculomotor Delayed-Response Task

90o

270o

0o

180o

315o

225o

135o

45o

100

50

0

B Oculomotor Control Task

Figure 3 Unilateral lesion to the principal sulcus causes impaired memory for spatial locations in thecontralesional hemifield Example data are shown from an oculomotor version of a spatial delayed-response task in which monkeys made saccades to the remembered location of cues (A) Accuracy ofmemory-guided saccade is impaired for locations in the upper right quadrant after left principal sulcuslesion Note that the lengthening of the delay exacerbates the deficit Accuracy is depicted as a percentageof decline in accuracy after the lesion (B) Accuracy of visually guided saccades is not impaired suggest-ing intact visual and motor function From ldquoDorsolateral prefrontal lesions and oculomotor delayed-response performance Evidence for mnemonic lsquoscotomasrsquordquo by S Funahashi C J Bruce and P SGoldman-Rakic 1993 Journal of Neuroscience 13 p 1482 Copyright 1993 by the Society for Neuro-science Adapted with permission


Goldman-Rakic 1999 Petrides 2000a) cause us to con-clude that dorsolateral PFC lesions cause specific im-pairments in the maintenance of spatial memoranda

Sensory modality of memoranda is not limited tovision Almost all of the studies we have discussed up tothis point have used visually presented stimuli whichleaves open the possibility that the deficits they havenoted following prefrontal damage are modality spe-cific However this does not appear to be the case Pre-frontal lesions also cause delayed-response impairmentswhen spatial memoranda cues are auditory (Blum 1952Oscar-Berman 1975) Cooling of the PFC also causesimpairments when memoranda are perceived haptically(Shindy Posley amp Fuster 1994) Therefore it appearsthat lesions to the PFC cause impaired delayed-responseperformance regardless of the input sensory modality

Monitoring and self-ordered task performance afterdorsolateral prefrontal lesions Michael Petrides themajor proponent of the process-specific theory of mid-lateral PFC organization developed a variation on thedelayed-response task called the self-ordered task (see Fig-ure 2 Petrides amp Milner 1982) Self-ordered tasks notonly require that the animal remember some stimulusfeature over a delay but also critically tax the animalrsquosability to monitor the contents of working memoryMonitoring according to Petrides (Petrides 2000b) isan executive control process that is used to track updateand order the number of items stored in working mem-ory According to the process-specific model the mid-dorsolateral PFC (ie area 946d and 46) supports thisexecutive process irrespective of the nature of the mate-rial (ie spatial vs nonspatial) Recall that monkeyswith dorsolateral PFC lesions are not impaired on testsof nonspatial working memory that rely on recognitionmemory (eg delayed matching or nonmatching-to-sample or object alternation tasks) However they do failon nonspatial self-ordered tasks that in addition to main-tenance require the executive process of monitoring(Petrides 1991 1995 2000a) These results are the cor-nerstone of the process-specific modelrsquos position that themid-dorsolateral PFC supports working memory pro-cesses both spatial and nonspatial by tracking or mon-itoring this information

Levy and Goldman-Rakic (1999) directly challengedthese conclusions with data from monkeys with dorso-lateral PFC lesions that performed spatial and nonspa-tial versions of self-ordered and delayed-response ormatching tasks Lesions to the dorsolateral PFC causedimpairments on spatial delayed-response tasks but noton object delayed nonmatching-to-sample tasks repli-cating the effects that had been shown many times be-fore (see above) Lesions to the dorsolateral PFC causedlasting impairments on the spatial but not the object self-ordered task Animals had not been tested on a spatialversion of the self-ordered task previous to this studybut impairments on such a task are consistent with boththe material- and process-specific models However the

intact performance of monkeys with lesions to the dor-solateral PFC on the object version of the self-orderedtask a version very similar to the one that yielded defi-cits in Petridesrsquos studies is particularly problematic (Levyamp Goldman-Rakic 2000)

These data are strong evidence in favor of the material-specific model at least with regard to the specific role ofthe dorsolateral PFC in spatial working memory Thecomparability of the exact location of the dorsolateralPFC lesions in Petrides (2000b) and Levy and Goldman-Rakic (1999) must be carefully considered howeverBoth sets of investigators acknowledged that the lesionedloci were not the same and that this fact could be relatedto the discrepancy in their results (see Figure 4) Levyand Goldman-Rakic (1999) lesioned the tissue in thedepths of the sulcus in areas 46 946d and the anteriorportion of sulcal 8Ad The Petrides (2000b) lesions weremostly restricted to the dorsal lip of the principal sulcusextending up around the midlateral gyrus (areas 46 946dand inferior-lateral portions of area 9) but importantlyspared most of the tissue in the depths of the sulcus it-self Therefore it may be that intact performance on theobject self-ordered task can be seen after damage to themid-dorsolateral PFC as long as the lesion spares the gyralconvexity

Dorsolateral prefrontal cortex lesion summaryOverall there is compelling evidence much of it fromthe work of Goldman-Rakic that damage to the dorso-lateral PFC especially in the depths of the principal sulcusin areas 46 946 and 8Ad causes profound impairmentsin spatial working memory Performance is typically intacton tests of nonspatial working memory unless executivecontrol processes are required as has been demonstratedby the work of Petrides Several outstanding issues haveyet to be resolved however First the differential contri-butions of the sulcal and gyral portions of the dorsolat-eral PFC to working memory performance need to beaddressed For instance it is unknown whether or not le-sions like those in Petrides (2000b) but extending to theconvexity of the mid-dorsolateral PFC in areas 46 and946 would result in impaired performance on a spatialdelayed-response task which does not depend criticallyon monitoring Second how much does impaired spatialworking memory depend on damage to area 8Ad whichis known to contain the frontal eye fields and was includedin the lesions in Levy and Goldman-Rakic (1999) Thesecond question has been partially addressed by Funa-hashi et al (1993) in which two of the monkeysrsquo lesionsdid not extend into caudal 8Ad these monkeys did in-deed show impaired oculomotor delayed-response taskperformance

Localization II Ventrolateral Prefrontal Cortex(or Inferior Convexity)

As discussed lesions of the principal sulcus cause im-paired spatial but unimpaired nonspatial working mem-ory deficits a pattern of findings consistent with the


material-specific model In addition that model pro-poses that the tissue below the principal sulcus (areas46v 946v 4712) is specialized for the maintenance ofnonspatial memoranda like objects and faces Indeedneurons in area 45 and part of area 12 often show selec-tive responses to visually presented faces and objects(Oacute Scalaidhe Wilson amp Goldman-Rakic 1999 WilsonOacute Scalaidhe amp Goldman-Rakic 1993) In contrast withthe dorsolateral PFC literature relatively little data hasaccumulated on the effects of lesions to the ventrolateralPFC but nonetheless we review these findings below

Spatial working memory performance after ven-trolateral prefrontal lesions Contrary to the commonclaim that lesions to the ventrolateral PFC do not nor-mally impair spatial working memory (eg Levy ampGoldman-Rakic 1999 2000) substantial impairments

on tests of spatial working memory have often been ob-served following such lesions (Iversen amp Mishkin 1970Mishkin et al 1969 Passingham 1975 Rosenkildeet al 1981)

Nonspatial working memory performance afterventrolateral prefrontal lesions Lesions to the ventro-lateral PFC have been shown to produce impaired per-formance on tests of nonspatial working memory likedelayed object alternation tasks (Mishkin amp Manning1978) and tests of recognition memory for objects or col-ors (Iversen amp Mishkin 1970 Mishkin amp Manning1978 Passingham 1975) However these impairmentsare often reported to be relatively mild and transient(Kowalska Bachevalier amp Mishkin 1991 Passingham1975) Also unlike the results following dorsolateralPFC lesions where performance deficits necessarily de-

Figure 4 Lesions from (A) Petrides (2000a) and (B) Levy and Goldman-Rakic (1999) Levyand Goldman-Rakic (1999) lesioned the tissue in the depths of the sulcus in areas 46 946dportions of 946v and the anterior portion of sulcal 8Ad The Petrides (2000a) lesions weremostly restricted to the dorsal lip of the principal sulcus extending up around the midlateralgyrus (areas 46 946d and inferior-lateral portions of area 9) but spared most of the tissuein the depths of the sulcus itself (A) From ldquoDissociable roles of mid-dorsolateral prefrontaland anterior inferotemporal cortex in visual working memoryrdquo by M Petrides 2000 Jour-nal of Neuroscience 20 p 7497 Copyright 2000 by the Society for Neuroscience Reprintedwith permission (B) From ldquoAssociation of storage and processing functions in the dorsolat-eral prefrontal cortex of the nonhuman primaterdquo by R Levy and P S Goldman-Rakic1999 Journal of Neuroscience 19 p 5152 Copyright 1999 by the Society for NeuroscienceReprinted with permission

A

B


pend on the imposition of a delay ventrolateral PFC le-sions often cause impairments even when no delay ispresent (Iversen amp Mishkin 1970 Passingham 1975Rushworth Nixon Eacott amp Passingham 1997) In-deed no delay-dependent working memory impairmenthas been reported following ventrolateral PFC lesions

The most compelling data that questions the role ofthe ventrolateral PFC in working memory generally andnot just in nonspatial working memory is that of Rush-worth et al (1997) who reported that removal of theventrolateral PFC (area 4712) resulted in no effect onpattern or color working memory Lesions were then ex-tended to include area 45A which resulted in impair-ments in a color matching-to-sample task with no delay(ie the sample and response choices were presented si-multaneously) but the imposition of increasing delaysdid not exacerbate the effect of the lesion (see Figure 5)These data led the authors to suggest that the ventrolat-eral PFC is not essential for working memory

Associative learning after ventrolateral prefrontallesions Given that the effects of lesions to the ventro-lateral PFC may not be best characterized as impairedworking memory what role may this portion of the cor-tex have in cognition There have been several reportsof impaired performance on what are called conditionalvisuomotor learning or arbitrary visuomotor mappingtasks after lesions to the ventrolateral PFC which thus

deserve comment These tasks require that the animallearn arbitrary associations between visual cues andmotor responses The word ldquoarbitraryrdquo is important herebecause it means that the stimulusndashresponse mapping isnot spatially guided For instance the position of a redlight at an intersection does not in itself provide anyguide as to where one should drive but it does indicatethat one should stop there is an arbitrary nonspatialmapping between a red signal light and the motor be-havior associated with the light

Using a crossed disconnection procedure in which theventrolateral PFC is damaged in one hemisphere and theinferior temporal (IT) cortex is damaged in the otherhemisphere followed by a transection of the corpus cal-losum Gaffan and colleagues have shown that interac-tions between the ventrolateral PFC and IT are necessaryfor normal learning of arbitrary visuomotor associations(Gaffan amp Harrison 1988 Parker amp Gaffan 1998) Tran-section of the major fiber bundle connecting IT and ven-trolateral PFC the uncinate fasciculus also impairslearning of arbitrary visuomotor associations (Eacott ampGaffan 1992) Moreover monkeys with bilateral ven-trolateral PFC lesions have terrible difficulties learningarbitrary visuomotor mappings (Bussey Wise amp Mur-ray 2001 see Figure 6) These deficits do not appear tobe caused by sensory or motor failures per se as themonkeys can achieve substantial levels of performance

Figure 5 Results of lesions to the ventrolateral PFC (areas 4712 and 45A) on a color matching-to-sample task (A) No-delay matching Number of errors made during task learning before andafter lesion The sample cue and the two color choices were presented simultaneously with no delay(B) Delayed matching Number of errors made at various delays before and after lesion After re-learning the task with extended training the monkeys were retested with a delayed matching-to-sample task Note that the introduction of a delay did not interact with the lesion From ldquoVentralprefrontal cortex is not essential for working memoryrdquo by M F Rushworth P D Nixon M J Ea-cott and R E Passingham 1997 Journal of Neuroscience 17 pp 4835ndash4836 Copyright 1997 bythe Society for Neuroscience Adapted with permission

0

300

100

500

200Err

ors

to

Crite

rio

n

Operation

Pre Post

400

600

A B

0 2 4 6 80

50

100

150

Err

ors

Delay

Preoperation (2 testings)

Postoperation

200

250Preoperation

Postoperation


on a concurrent visual discrimination task and analysisof the errors indicates that they are not consistent withperseveration or systematic response biases (Busseyet al 2001) Even small reversible lesions applied withthe local infusion of bicuculline into a restricted area ofthe ventrolateral PFC cause impaired arbitrary visuomo-tor learning while infusions into the dorsolateral PFC donot (Wang Zhang amp Li 2000) The pattern of perfor-mance deficits following ventrolateral PFC damage maybe due to a failure to form andor use arbitrary visuo-motor associations These stimulusndashresponse mappingsare essentially rules (Murray Bussey amp Wise 2000 Wiseamp Murray 2000) that could be low-level (for instancestop at a red signal light) or high-level (for instance sig-nal lights guide traffic flow) in nature Alternatively theventrolateral PFC lesion could impair performance byimpacting the animalrsquos strategic approach to the prob-lem Error analyses indicate that the monkeys fail toapply win-stay lose-shift and change-shift strategiesduring the learning sessions (Bussey et al 2001 Wanget al 2000)

Ventrolateral prefrontal lesion summary Some ev-idence suggests that lesions to the ventrolateral PFCcause nonspatial working memory impairments How-ever there are several problems with this interpretationFirst no study has demonstrated a selective deficit in thespatial as compared with the nonspatial domain When

deficits were found in the spatial domain they werenoted in the nonspatial domain as well Second there hasbeen no demonstration of a delay dependency Impair-ments have been observed with tasks that do not requiremaintenance and increasing the delay does not exacer-bate the deficit Third damage to the ventrolateral PFCcauses impairments with tasks that make no workingmemory demands For instance performance on condi-tional visuomotor association tasks is impaired afterventrolateral PFC damage (Bussey et al 2001 Gaffanamp Harrison 1988 Parker amp Gaffan 1998 Wang et al2000) Therefore impairments following ventrolateralPFC lesions may be best conceptualized as an inabilityto form andor use associations between cues and motorresponses or may be more generally associated with in-ability to use learning strategies (Murray et al 2000)

Evaluating the Theories of the FunctionalSegregation of the Mid-Lateral PFC

Material-Specific ModelThis model proposed by Goldman-Rakic predicts that

lesions to the dorsolateral PFC cause specific impair-ments in spatial but not in nonspatial working memoryperformance Indeed the specificity of dorsolateral PFClesions to spatial working memory is the strongest sup-port for this model Lesions to the dorsolateral PFC reli-

Figure 6 Effects of ventrolateral PFC lesions on conditional visuomotor learning Monkeysattempted to learn an arbitrary joystick movement (leftrightpull) associated with three vi-sual cues On repeat trials the visual cue was the same as on the immediately preceding trial(filled markers) but on change trials the visual cue was different (open markers) Prior tothe lesion monkeys learned the visuomotor associations as can be seen by the increasingslope (Pre preoperative) Since performance is already above chance very early in the ses-sion before the associations have been learned the monkeys must have appreciated thestrategies of repeat-stay and change-shift After the lesion the monkeys showed no evidenceof learning the visuomotor associations performance was at chance levels (Post postop-erative) Additionally they failed to use adaptive strategies On repeat trials they failed toemit the same response after correctly responding (repeat-stay) and on change trials theyfailed to emit a different response after correctly responding (change-shift) From ldquoThe roleof ventral and orbital prefrontal cortex in conditional visuomotor learning and strategy usein rhesus monkeys (Macaca mulatta) by T J Bussey S P Wise and E A Murray 2001 Be-havioral Neuroscience 115 p 975 Public domain

70

60

50

40

30

20

10

0

ndash104 8 12 16 20 48444036322824

Pre repeat

Pre change

Post repeat

Post change

Trial

Pe

rce

nta

ge

Im

pro

ve

me

nt

Fro

m C

ha

nce


ably cause significant impairments in spatial workingmemory performance and the effect is exacerbated withincreasing delays The model also predicts that ventro-lateral PFC lesions cause specific impairments in non-spatial working memory performance So far very littlesupport exists that indicates that the ventrolateral PFC isnecessary for nonspatial working memory Clearly theoverall pattern of results does not suggest a double dis-sociation between dorsolateral and ventrolateral PFC le-sions and spatial and nonspatial working memory

Process-specific modelThis model proposed by Petrides predicts that lesions

to the mid-dorsolateral PFC cause an impaired ability tomaintain and monitor or track multiple items in workingmemory The strongest support for this prediction is thefinding that when monitoring demands are increasedeven when the memoranda are objects the performanceof monkeys with mid-dorsolateral PFC lesions is im-paired However these same monkeys can still maintainobjects over varying retention intervals without problemIf monitoring is indeed the critical process that dependson mid-dorsolateral PFC integrity then it remains un-clear why such lesions cause impairments on simplertests of spatial working memory dorsolateral PFC le-sions impair performance on spatial delayed responseand oculomotor delayed response tasks that do not de-pend on monitoring The model also predicts that ven-trolateral PFC lesions will cause impairments in activeor strategic encoding and retrieval of information How-ever these aspects of the model have yet to be tested andtheir relation to learning arbitrary visuomotor associa-tions has yet to be formalized

Conclusions

The material- and process-specific theories have beencritical catalysts in our attempts to understand the func-tional specializations of PFC subregions by motivatingresearch and providing a heuristic framework withinwhich hypotheses can be tested It cannot be overstatedhow important the findings from monkey lesion studieshave been in the development of these two theories of thefunctional organization of the midlateral PFC The useof lesions to map structurendashfunction relationships ex-isted before the development of new imaging techniqueswhen the roots of the current PFC theories were also de-veloping Recently the lessons from these studies haveno longer been given primary importance despite thefact that lesions can be used to test the necessity of abrain region for cognition powerfully Theory testing iscurrently being done with the use of correlative techniquessuch as electrophysiology (Funahashi 2001 Fuster 2001E K Miller 2000 Wilson et al 1993) and neuroimaging(Curtis amp DrsquoEsposito 2003 Curtis Zald Lee amp Pardo2000 DrsquoEsposito Postle amp Rypma 2000 Funahashi2001 Fuster 2001 E K Miller 2000 Owen 2000 Owen

et al 1998 Passingham Toni amp Rushworth 2000 Unger-leider Courtney amp Haxby 1998 Wilson et al 1993) Thegoal of this review was to revive awareness of the im-portant contributions that lesion studies have made Tothis end we summarized the effects of PFC lesions withspecial emphasis on how the findings fit with predic-tions made by dominant PFC theories We conclude thatthe most consistent results from numerous lesions to thePFC suggest that the mid-dorsolateral PFC is necessaryfor some aspect of spatial working memory task perfor-mance and that the ventrolateral PFC is necessary forsome aspect of learning arbitrary stimulusndashresponse as-sociations These conclusions only partially overlap withthe predictions from the material- and process-specificmodels indicating that in their current forms neither ofthese models can adequately account for the functionaltopography of the midlateral PFC

REFERENCES

Bachevalier J amp Mishkin M (1986) Visual recognition impair-ment follows ventromedial but not dorsolateral prefrontal lesions inmonkeys Behavioral amp Brain Research 20 249-261

Bauer R H amp Fuster J M (1976) Delayed-matching and delayed-response deficit from cooling dorsolateral prefrontal cortex in monkeysJournal of Comparative amp Physiological Psychology 90 293-302

Bianchi L (1922) The mechanism of the brain and the function of thefrontal lobes Edinburgh E amp S Livingstone

Blum R A (1952) Effects of subtotal lesions of frontal granular cor-tex on delay reaction in monkeys Archives of Neurology amp Psychia-try 67 375-386

Bussey T J Wise S P amp Murray E A (2001) The role of ventraland orbital prefrontal cortex in conditional visuomotor learning andstrategy use in rhesus monkeys (Macaca mulatta) Behavioral Neu-roscience 115 971-982

Butters N amp Pandya D (1969 September 19) Retention of delayed-alternation Effect of selective lesions of sulcus principalis Science165 1271-1273

Butters N Pandya D Stein D amp Rosen J (1972) A search forthe spatial engram within the frontal lobes of monkeys Acta Neuro-biologiae Experimentalis 32 305-329

Curtis C E amp DrsquoEsposito M (2003) Persistent activity in the pre-frontal cortex during working memory Trends in Cognitive Sciences7 415-423

Curtis C E Zald D H Lee J T amp Pardo J V (2000) Object andspatial alternation tasks with minimal delays activate the right ante-rior hippocampus proper in humans NeuroReport 11 2203-2207

DrsquoEsposito M Postle B R amp Rypma B (2000) Prefrontal corti-cal contributions to working memory Evidence from event-relatedfMRI studies Experimental Brain Research 133 3-11

Eacott M J amp Gaffan D (1992) Inferotemporal-frontal discon-nection The uncinate fascicle and visual associative learning in mon-keys European Journal of Neuroscience 4 1320-1332

Ferrier D (1886) The functions of the brain (2nd ed) LondonSmith Elder amp Co

Franz S I (1907) On the functions of the cerebrum The frontal lobesNew York Science Press

Funahashi S (2001) Neuronal mechanisms of executive control bythe prefrontal cortex Neuroscience Research 39 147-165

Funahashi S Bruce C J amp Goldman-Rakic P S (1989) Mne-monic coding of visual space in the monkeyrsquos dorsolateral prefrontalcortex Journal of Neurophysiology 61 331-349

Funahashi S Bruce C J amp Goldman-Rakic P S (1990) Visuo-spatial coding in primate prefrontal neurons revealed by oculomotorparadigms Journal of Neurophysiology 63 814-831

Funahashi S Bruce C J amp Goldman-Rakic P S (1991) Neu-


ronal activity related to saccadic eye movements in the monkeyrsquos dor-solateral prefrontal cortex Journal of Neurophysiology 65 1464-1483

Funahashi S Bruce C J amp Goldman-Rakic P S (1993) Dorso-lateral prefrontal lesions and oculomotor delayed-response perfor-mance Evidence for mnemonic ldquoscotomasrdquo Journal of Neuroscience13 1479-1497

Fuster J M (2001) The prefrontal cortexmdashan update Time is of theessence Neuron 30 319-333

Fuster J M amp Alexander G E (1970) Delayed response deficitby cryogenic depression of frontal cortex Brain Research 20 85-90

Fuster J M amp Bauer R H (1974) Visual short-term memory def-icit from hypothermia of frontal cortex Brain Research 81 393-400

Fuster J M Bauer R H amp Jervey J P (1982) Cellular dischargein the dorsolateral prefrontal cortex of the monkey in cognitive tasksExperimental Neurology 77 679-694

Gaffan D amp Harrison S (1988) Inferotemporal-frontal discon-nection and fornix transection in visuomotor conditional learning bymonkeys Behavioral amp Brain Research 31 149-163

Goldman P S amp Rosvold H E (1970) Localization of functionwithin the dorsolateral prefrontal cortex of the rhesus monkey Ex-perimental Neurology 27 291-304

Goldman P S Rosvold H E Vest B amp Galkin T W (1971)Analysis of the delayed-alternation deficit produced by dorsolateralprefrontal lesions in the rhesus monkey Journal of Comparative ampPhysiological Psychology 77 212-220

Gross C G amp Weiskrantz L (1962) Evidence for dissociation of im-pairment on auditory discrimination and delayed response following lat-eral frontal lesions in monkeys Experimental Neurology 5 453-476

Hitzig E (1874) Untersuchungen uber das Gehirn [Investigations ofthe brain] Berlin A Hirschwald

Hunter W S (1913) The delayed reaction in animals BehavioralMonographs 2 21-30

Iversen S D amp Mishkin M (1970) Perseverative interference inmonkeys following selective lesions of the inferior prefrontal con-vexity Experimental Brain Research 11 376-386

Jacobsen C F (1936) Studies of cerebral function in primates I Thefunctions of the frontal association areas in monkeys ComparativePsychology Monographs 13 1-60

Kowalska D M Bachevalier J amp Mishkin M (1991) The roleof the inferior prefrontal convexity in performance of delayed non-matching-to-sample Neuropsychologia 29 583-600

Levy R amp Goldman-Rakic P S (1999) Association of storage andprocessing functions in the dorsolateral prefrontal cortex of the non-human primate Journal of Neuroscience 19 5149-5158

Levy R amp Goldman-Rakic P S (2000) Segregation of workingmemory functions within the dorsolateral prefrontal cortex Experi-mental Brain Research 133 23-32

Miller E K (2000) The prefrontal cortex and cognitive control Na-ture Reviews Neuroscience 1 59-65

Miller M H amp Orbach J (1972) Retention of spatial alternationfollowing frontal lobe resections in stump-tailed macaques Neuro-psychologia 10 291-298

Mishkin M (1954) Visual discrimination performance following partialablations of the temporal lobe II Ventral surface vs hippocampusJournal of Comparative amp Physiological Psychology 47 187-193

Mishkin M (1957) Effects of small frontal lesions on delayed alter-nation in monkeys Journal of Neurophysiology 20 615-622

Mishkin M amp Manning F J (1978) Non-spatial memory after se-lective prefrontal lesions in monkeys Brain Research 143 313-323

Mishkin M amp Pribram K H (1954) Visual discrimination perfor-mance following partial ablations of the temporal lobe I Ventral vslateral Journal of Comparative amp Physiological Psychology 47 14-20

Mishkin M amp Pribram K H (1955) Analysis of the effects offrontal lesions in monkey I Variations of delayed alternation Jour-nal of Comparative amp Physiological Psychology 48 492-495

Mishkin M amp Pribram K H (1956) Analysis of the effects offrontal lesions in monkey II Variations of delayed response Journalof Comparative amp Physiological Psychology 49 36-40

Mishkin M Vest B Waxler M amp Rosvold H E (1969) A re-examination of the effects of frontal lesions on object alternationNeuropsychologia 7 357-364

Murray E A Bussey T J amp Wise S P (2000) Role of prefrontalcortex in a network for arbitrary visuomotor mapping ExperimentalBrain Research 133 114-129

Oacute Scalaidhe S P Wilson F A amp Goldman-Rakic P S (1999)Face-selective neurons during passive viewing and working memoryperformance of rhesus monkeys Evidence for intrinsic specializa-tion of neuronal coding Cerebral Cortex 9 459-475

Oscar-Berman M (1975) The effects of dorsolateral-frontal and ventrolateral-orbitofrontal lesions on spatial discrimination learningand delayed response in two modalities Neuropsychologia 13 237-246

Owen A M (2000) The role of the lateral frontal cortex in mnemonicprocessing The contribution of functional neuroimaging Experi-mental Brain Research 133 33-43

Owen A M Stern C E Look R B Tracey I Rosen B R ampPetrides M (1998) Functional organization of spatial and nonspatialworking memory processing within the human lateral frontal cortexProceedings of the National Academy of Sciences 95 7721-7726

Parker A amp Gaffan D (1998) Memory after frontaltemporal dis-connection in monkeys Conditional and nonconditional tasks uni-lateral and bilateral frontal lesions Neuropsychologia 36 259-271

Passingham R [E] (1975) Delayed matching after selective pre-frontal lesions in monkeys (Macaca mulatta) Brain Research 9289-102

Passingham R E Toni I amp Rushworth M F (2000) Specializa-tion within the prefrontal cortex The ventral prefrontal cortex and as-sociative learning Experimental Brain Research 133 103-113

Petrides M (1991) Monitoring of selections of visual stimuli and theprimate frontal cortex Proceedings of the Royal Society of LondonSeries B 246 293-298

Petrides M (1995) Impairments on nonspatial self-ordered and exter-nally ordered working memory tasks after lesions of the mid-dorsalpart of the lateral frontal cortex in the monkey Journal of Neuro-science 15 359-375

Petrides M (2000a) Dissociable roles of mid-dorsolateral prefrontaland anterior inferotemporal cortex in visual working memory Jour-nal of Neuroscience 20 7496-7503

Petrides M (2000b) The role of the mid-dorsolateral prefrontal cor-tex in working memory Experimental Brain Research 133 44-54

Petrides M amp Milner B (1982) Deficits on subject-ordered tasksafter frontal- and temporal-lobe lesions in man Neuropsychologia20 249-262

Petrides M amp Pandya D N (1994) Comparative architectonicanalysis of the human and macaque frontal cortex In F Boller ampJ Grafman (Eds) Handbook of neuropsychology (Vol 9 pp 17-58)Amsterdam Elsevier

Petrides M amp Pandya D N (2002) Comparative cytoarchitectonicanalysis of the human and the macaque ventrolateral prefrontal cor-tex and corticocortical connection patterns in the monkey EuropeanJournal of Neuroscience 16 291-310

Pribram K H amp Mishkin M (1956) Analysis of the effects offrontal lesions in monkey III Object alternation Journal of Com-parative amp Physiological Psychology 49 41-45

Quintana J amp Fuster J M (1992) Mnemonic and predictive func-tions of cortical neurons in a memory task NeuroReport 3 721-724

Quintana J amp Fuster J M (1993) Spatial and temporal factors inthe role of prefrontal and parietal cortex in visuomotor integrationCerebral Cortex 3 122-132

Quintana J Yajeya J amp Fuster J M (1988) Prefrontal repre-sentation of stimulus attributes during delay tasks I Unit activity incross-temporal integration of sensory and sensory-motor informa-tion Brain Research 474 211-221

Rao S C Rainer G amp Miller E K (1997 May 2) Integration ofwhat and where in the primate prefrontal cortex Science 276 821-824

Rosenkilde C E Rosvold H E amp Mishkin M (1981) Time dis-crimination with positional responses after selective prefrontal le-sions in monkeys Brain Research 210 129-144


Rushworth M F Nixon P D Eacott M J amp Passingham R E(1997) Ventral prefrontal cortex is not essential for working memoryJournal of Neuroscience 17 4829-4838

Shindy W W Posley K A amp Fuster J M (1994) Reversible def-icit in haptic delay tasks from cooling prefrontal cortex CerebralCortex 4 443-450

Stamm J S (1973) Functional dissociation between the inferior andarcuate segments of dorsolateral prefrontal cortex in the monkeyNeuropsychologia 11 181-190

Stamm J S amp Weber-Levine M L (1971) Delayed alternation im-pairments following selective prefrontal cortical ablations in mon-keys Experimental Neurology 33 263-278

Ungerleider L G Courtney S M amp Haxby J V (1998) Aneural system for human visual working memory Proceedings of theNational Academy of Sciences 95 883-890

Ungerleider L G amp Mishkin M (1982) Two cortical visual sys-tems In R J W Mansfield (Ed) Analysis of visual behavior (pp 549-586) Cambridge MA MIT Press

Volgushev M Vidyasagar T R Chistiakova M amp Eysel U T(2000) Synaptic transmission in the neocortex during reversiblecooling Neuroscience 98 9-22

Walker A E (1940) A cytoarchitectural study of the prefrontal area

of the macaque monkey Journal of Comparative Neurology 73 59-86

Wang M Zhang H amp Li B M (2000) Deficit in conditional vi-suomotor learning by local infusion of bicuculline into the ventralprefrontal cortex in monkeys European Journal of Neuroscience 123787-3796

Wilson F A Oacute Scalaidhe S P amp Goldman-Rakic P S (1993June 25) Dissociation of object and spatial processing domains inprimate prefrontal cortex Science 260 1955-1958

Wise S P amp Murray E A (2000) Arbitrary associations betweenantecedents and actions Trends in Neurosciences 23 271-276

NOTE

1 Other areas of the PFC are not addressed in this review for severalreasons Although in some of the earlier less precise PFC lesions area 10was involved this region has not been systematically studied The termdorsolateral PRC lesions do not refer to damage to the dorsomedial por-tion of area 9 and area 44 which is a premotor area in the frontal cortex

(Manuscript received August 3 2004revision accepted for publication November 11 2004)


form qualitatively different operations (Petrides 2000b)The process-specific model proposes a hierarchy in whichthe ventrolateral PFC supports processes such as the ac-tive encoding and retrieval of information and the mid-dorsolateral PFC supports higher order ldquoexecutiverdquo con-trol functions like the monitoring and manipulation ofstored information This report critically reviews themonkey lesion studies that have provided the foundationfor current theories regarding the functional organizationof the PFC

History

Hitzig (1874) and Ferrier (1886) described over 100years ago the effects of experimental ablations of the

PFC Early naturalistic observations made by these ex-perimental psychologists led them to conclude that thefrontal cortex is particularly concerned with intellectualand attentional rather than sensory and motor functionsLater more extensive comparative studies and structuredbehavioral methods were used to investigate the effectsof frontal lesions The Italian physiologist Bianchi (1922)concluded that removal of the frontal lobes destroyed theability of the animal to synthesize various incoming per-cepts and integrate these with outgoing motor commandsIn his view the frontal cortex was the top of the sensoryand motor hierarchies and was the final stage of integra-tion The American psychologist Franz (1907) unlikehis predecessors used structured and objective behav-ioral testing methods (eg Thorndike boxes and basicvisual discrimination problems) to study the effects offrontal lobe removal Unfortunately though the behav-ioral tests he chose proved inadequate for capturing thetypes of mental processes that are impaired after frontallesions Up to this point Hitzig Ferrier Bianchi andFranz had produced intriguing but still somewhat im-pressionistic work that failed to establish what role thePFC might play in high-level cognition What was clearlylacking from their studies was a means for reliably andflexibly assessing the behavioral effects of prefrontal le-sions the pioneering researchers had noted these effectsbut were as yet unable to capture them In the hands ofJacobsen (1936) Hunterrsquos (1913) delayed-response task(Figure 2) proved to be a selective index of primate pre-frontal deficit Jacobsenrsquos highly influential monographdetailed the delayed-response and visual discriminationcapabilities of monkeys that had been given prefrontalpremotor or temporal lobe lesions He demonstrated thatonly prefrontal monkeys showed a selective and delay-dependent deficit on delayed-response tasks Delayed-response tasks depend upon the animalrsquos ability to main-tain an internal representation over the course of a delayAt the time of choice this internal representation mustguide behavior because no features present in the envi-ronment signal the correct response With these ideas inmind Jacobsen concluded that the deficit caused by pre-frontal cortical damage was mnemonic in nature specif-ically the monkeyrsquos ability to use ldquoimmediate memoryrdquoto guide behavior was impaired by the prefrontal lesion

The Effects of Prefrontal Lesions on thePerformance of Delayed-Response Tasks

Following Jacobsenrsquos (1936) pivotal demonstration ofimpaired delayed-response task performance after pre-frontal ablations several decades of investigations haveattempted to identify exactly which area(s) of the PFCare necessary for delayed-response performance (ie lo-calization) and which mental process or processes nec-essary for delay-response performance are impaired bydamage to the PFC (ie functional specification)

Figure 1 Subdivisions of the macaque prefrontal cortex as de-fined by Petrides and Pandya (1994) on the basis of connectivityand cytoarchitecture The mid-dorsolateral PFC is composed ofareas 46 and 946d and the mid-ventrolateral PFC of areas 471245A and 946v Note that many of the lesion studies reviewed hereused Walkerrsquos earlier cytoarchetectonic map (Walker 1940) inwhich Petrides and Pandyarsquos (1994) subdivisions 46 946d and946v are combined into a simpler area 46 Also depicted are thedorsal and ventral streams that respectively process spatial andnonspatial visual information (Ungerleider amp Mishkin 1982)The dorsal stream has relatively greater dorsolateral PFC pro-jections and the ventral stream has relatively greater ventrolat-eral PFC projections (Petrides amp Pandya 2002 UngerleiderCourtney amp Haxby 1998) Projections from the posterior pari-etal cortex (PG) primarily target 8A and to a lesser degree 946das depicted by the thickness of the line Projections from the in-ferior temporal cortex (IT) primarily target 4712 and to a lesserextent 45 The dashed line represents the principal sulcus whichdelineates the dorsal and ventral aspects of the prefrontal cor-texrsquos lateral convexity

8Ad946d44

8Av946v45A

8B4645B

9104712





Food

Reward

Correct

Choice

Cue

Delay

Response

Correct

Choice

Cue

Delay

Response

Cue

Delay

Response

Delay

Response

Delayed

Response

Delayed

Matching-to-Sample

Self

Ordered
















90o

270o

0o

180o

315o

225o

135o

45o

50

25

0

15-sec delay

30-sec delay

60-sec delay


90o

270o

0o

180o

315o

225o

135o

45o

100

50

0



















A

B








0

300

100

500

200Err

ors

to

Crite

rio

n

Operation

Pre Post

400

600

A B

0 2 4 6 80

50

100

150

Err

ors

Delay


Postoperation

200

250Preoperation

Postoperation









70

60

50

40

30

20

10

0

ndash104 8 12 16 20 48444036322824

Pre repeat

Pre change

Post repeat

Post change

Trial

Pe

rce

nta

ge

Im

pro

ve

me

nt

Fro

m C

ha

nce





Conclusions



REFERENCES















































































NOTE







Food

Reward

Correct

Choice

Cue

Delay

Response

Correct

Choice

Cue

Delay

Response

Cue

Delay

Response

Delay

Response

Delayed

Response

Delayed

Matching-to-Sample

Self

Ordered
















90o

270o

0o

180o

315o

225o

135o

45o

50

25

0

15-sec delay

30-sec delay

60-sec delay


90o

270o

0o

180o

315o

225o

135o

45o

100

50

0



















A

B








0

300

100

500

200Err

ors

to

Crite

rio

n

Operation

Pre Post

400

600

A B

0 2 4 6 80

50

100

150

Err

ors

Delay


Postoperation

200

250Preoperation

Postoperation









70

60

50

40

30

20

10

0

ndash104 8 12 16 20 48444036322824

Pre repeat

Pre change

Post repeat

Post change

Trial

Pe

rce

nta

ge

Im

pro

ve

me

nt

Fro

m C

ha

nce





Conclusions



REFERENCES















































































NOTE

















90o

270o

0o

180o

315o

225o

135o

45o

50

25

0

15-sec delay

30-sec delay

60-sec delay


90o

270o

0o

180o

315o

225o

135o

45o

100

50

0



















A

B








0

300

100

500

200Err

ors

to

Crite

rio

n

Operation

Pre Post

400

600

A B

0 2 4 6 80

50

100

150

Err

ors

Delay


Postoperation

200

250Preoperation

Postoperation









70

60

50

40

30

20

10

0

ndash104 8 12 16 20 48444036322824

Pre repeat

Pre change

Post repeat

Post change

Trial

Pe

rce

nta

ge

Im

pro

ve

me

nt

Fro

m C

ha

nce





Conclusions



REFERENCES















































































NOTE



















A

B








0

300

100

500

200Err

ors

to

Crite

rio

n

Operation

Pre Post

400

600

A B

0 2 4 6 80

50

100

150

Err

ors

Delay


Postoperation

200

250Preoperation

Postoperation









70

60

50

40

30

20

10

0

ndash104 8 12 16 20 48444036322824

Pre repeat

Pre change

Post repeat

Post change

Trial

Pe

rce

nta

ge

Im

pro

ve

me

nt

Fro

m C

ha

nce





Conclusions



REFERENCES















































































NOTE










0

300

100

500

200Err

ors

to

Crite

rio

n

Operation

Pre Post

400

600

A B

0 2 4 6 80

50

100

150

Err

ors

Delay


Postoperation

200

250Preoperation

Postoperation









70

60

50

40

30

20

10

0

ndash104 8 12 16 20 48444036322824

Pre repeat

Pre change

Post repeat

Post change

Trial

Pe

rce

nta

ge

Im

pro

ve

me

nt

Fro

m C

ha

nce





Conclusions



REFERENCES















































































NOTE











70

60

50

40

30

20

10

0

ndash104 8 12 16 20 48444036322824

Pre repeat

Pre change

Post repeat

Post change

Trial

Pe

rce

nta

ge

Im

pro

ve

me

nt

Fro

m C

ha

nce





Conclusions



REFERENCES















































































NOTE







Conclusions



REFERENCES















































































NOTE
































































NOTE



the effects of prefrontal lesions on working memory ... · test the effects of prefrontal lesions....

Documents