Context-dependent, neural system-specific neurophysiological concomitants of ageing: mapping PET...

Brain (1999),122,963–979

Context-dependent, neural system-specificneurophysiological concomitants of ageing:mapping PET correlates during cognitive activationGiuseppe Esposito, Brenda S. Kirkby, John D. Van Horn, Timothy M. Ellmore andKaren Faith Berman

Unit on Integrative Neuroimaging, Clinical Brain Correspondence to: Karen Faith Berman, MD, NationalDisorders Branch, Intramural Research Program, National Institutes of Health, Building 10, Room 4C101,Insitute of Mental Health, Bethesda, MD, USA 9000 Rockville Pike, Bethesda, MD 20892-1365 USA

E-mail: [email protected]

SummaryWe used PET to explore the neurophysiological changesthat accompany cognitive disability in ageing, with afocus on the frontal lobe. Absolute regional cerebralblood flow (rCBF) was measured in 41 healthyvolunteers, evenly distributed across an age range of18–80 years, during two task paradigms: (i) theWisconsin Card Sorting Test (WCST), which dependsheavily on working memory and is particularly sensitiveto dysfunction of the dorsolateral prefrontal cortex(DLPFC); and (ii) Raven’s Progressive Matrices (RPM),which may also have a working memory component,but depends more on visuo-spatial processing and ismost sensitive to dysfunction of postrolandic regions.We used voxel-wise correlational mapping to determineage-related changes in WCST and RPM activation anddeveloped a method to quantitate and localize statisticaldifferences between the correlation maps for thetwo task paradigms. Because both WCST and RPMperformance declined with age, as expected, correlationalanalyses were performed with and without partialling outthe effect of task performance. Task-specific reductions ofrCBF activation with age were found in the DLPFCduring the WCST and in portions of the inferolateraltemporal cortex involved in visuo-spatial processingduring the RPM. We also found reduced ability tosuppress rCBF in the right hippocampal region during

Keywords: ageing; brain activation; cognition; frontal lobe; PET

Abbreviations: BA 5 Brodmann area; DLPFC5 dorsolateral prefrontal cortex; rCBF5 regional cerebral blood flow;RPM 5 Raven’s Progressive Matrices; RPMc5 control task for RPM; WCST5 Wisconsin Card Sorting Test; WCSTc5control task for WCST

© Oxford University Press 1999

the WCST and in mesial and polar portions of theprefrontal cortex during both task conditions. Task-dependent alterations with age in the relationshipbetween the DLPFC and the hippocampus were alsodocumented; because the collective pattern of changesin the hippocampal–DLPFC relationship with ageingwas opposite to that seen in a previous study usingdextroamphetamine, we postulated a dopaminergicmechanism. These results indicate that, despite somecognitive overlap between the two tasks and the age-related cognitive decline in both, many of the changesin rCBF activation with age were task-specific, reflectingfunctional alteration of the different neural circuitsnormally engaged by young subjects during the WCSTand RPM. Reduced activation of areas critical for taskperformance (i.e. the DLPFC during the WCST andposterior visual association areas of the inferolateraltemporal cortex during the RPM), in conjunction withthe inability to suppress areas normally not involvedin task performance (i.e. the left hippocampal regionduring the WCST and mesial polar prefrontal cortexduring both the WCST and RPM), suggest that, overall,reduced ability to focus neural activity may be impairedin older subjects. The context dependency of the age-related changes is most consistent with systems failureand disordered connectivity.

by guest on September 8, 2014

http://brain.oxfordjournals.org/D

ownloaded from

http://brain.oxfordjournals.org/

964 G. Espositoet al.

IntroductionIt is generally accepted that normal ageing is associated withdecline of certain higher cognitive functions. Notwithstandingwide individual differences, elderly individuals show acognitive pattern characterized by performance decrementson memory tasks and on tests involving abstract reasoning,problem solving, visuospatial skills and selective attention(for reviews, see Botwinick, 1984; Craik and Jennings,1992). Most investigators also find that cognitive behavioursspecifically linked to the prefrontal cortex, including workingmemory, mental flexibility and response to external feedback,are particularly affected by the ageing process (Bellevilleet al., 1996, for review, see Hochanadel and Kaplan, 1994)and that signs of prefrontal impairment, such as perseveration,may appear (Offenbach, 1974; Morriset al., 1990; Daigneaultand Braun, 1993).

Consistent with the notion of particular prefrontalinvolvement in ageing is the fact that a number of functionalbrain imaging studies have shown reduced blood flow and/orglucose metabolism in the prefrontal cortex (Kuhlet al.,1984; Pantanoet al., 1984; Devouset al., 1986; Martinet al.,1991; Waldemaret al., 1991). However, neither the functionalneuroimaging nor the neuropsychological literature is withoutcontroversy. Both Duaraet al. (1983) using PET and Yoshiiet al.(1988) with SPECT failed to confirm reduced prefrontalcortical activity in older subjects. Neuropsychologically, age-related cognitive decline has received a variety ofexplanations, some focusing on prefrontally related constructsand others not. These have included primary reduction ofworking memory capacity (Light, 1982), general cognitiveslowing (Salthouse, 1990) or decreased attentional abilitiesprobably due to failure of inhibitory processes that controlaccess to relevant information and removal of irrelevantinformation (Hasher and Zacks, 1988; Richardson, 1996).Thus, the exact nature of the cognitive decline thataccompanies ageing, and its neural substrate, haveremained unclear.

Recent PET studies of brain activation during cognitionhave demonstrated the utility of integrating neuro-psychological and neurophysiological investigations of thecognitive sequelae of the ageing process (Gradyet al., 1994,1995; Cabezaet al., 1997a). These studies have reported thatage-related cognitive changes are accompanied by alteredcerebral activation in temporo-occipital and parietalexstrastriate regions during tasks of visuospatial processing(Grady et al., 1994), and in medial temporal/hippocampalareas during tasks of memory encoding and retrieval (Gradyet al., 1995; Cabezaet al., 1997b), suggesting that importantpathophysiological underpinnings of age-related cognitivechanges may lie in shifts in thepatternof neural activationrather than, or perhaps in addition to, altered absolute globalor regional activity. In addition, the alterations in prefrontalactivation reported in older subjects collectively suggest thatthe qualitative nature of the changes is fluid and dependsupon the behaviour during the scan: increased prefrontal

activation was found both during face and location processing(Gradyet al., 1994) and during memory recall (Cabezaet al.,1997a), while reduced prefrontal activation was reportedduring memory encoding (Gradyet al., 1995; Cabezaet al., 1997b).

In the present study, to explore age-related changes inprefrontal function, we measured cognitive activation withPET regional cerebral blood flow (rCBF) in 41 carefullyscreened healthy subjects evenly distributed over an agerange of 18–80 years during performance of a task, theWisconsin Card Sorting Task (WCST), which appears todepend heavily on working memory (Goldman-Rakic, 1987;Berman et al., 1995) and to be particularly sensitive todysfunction of the dorsolateral prefrontal cortex (DLPFC)(Milner, 1963; Luria, 1973). We mapped anatomical areas ofsignificant correlation between age and neurophysiologicalactivation related to performing the WCST and statisticallycompared these correlative results for the WCST with thoseduring another abstract reasoning and problem-solving task,Raven’s Progressive Matrices (RPM). Like the WCST, RPMmay have a working memory component (Carpenteret al.,1990; Prabhakaranet al., 1997), but it appears to dependmore upon visuospatial processing and computationalproblem solving and to be most sensitive to lesions anddysfunction of postrolandic regions rather than of theprefrontal cortex (Bassoet al., 1973). We hypothesized,therefore, that participation of the prefrontal cortex and otherrelated regions in the two tasks could have different cognitiveand neurobiological import and that the neurophysiologicalmanifestations of the ageing process would, thus, be differentduring the tasks.

Material and methodsSubjectsForty-one healthy volunteers (20 men and 21 women) signedinformed consent in accordance with the National Institutesof Health Institutional Review Board and Radiation SafetyCommittee guidelines. Subjects were evenly distributedamong the age range of 18–80 years: fiveø 20 years (threefemale, two male) and six (three female, three male) for eachdecade from 21 to 80 years. Mean age was 45.56 19.7years (45.96 19.8 years for men and 45.26 20.1 years forwomen). Subjects received on average 16 years of education.There was no correlation between subjects’ age and years ofeducation. Six subjects were left handed. Becausecerebrovascular risk factors have been postulated to affectfunctional brain imaging results in ageing studies (Naritomiet al., 1979; for reviews, see Dasturet al., 1963; Meyeret al., 1994), we carefully screened this cohort for history ofpast or present medical condition or pharmacologicaltreatment that could have relevance for rCBF or metabolism,and for psychiatric or neurological disorders, including headtrauma or substance abuse. MRI scans were also obtained



ownloaded from


Context-dependent rCBF changes with ageing 965

(double-echo sequence; repetition time, TR5 2000; echotime, TE 5 20/80) and were reviewed for structuralabnormalities by a neuroradiologist. Screening of subjectsù 40 years of age also included physical examination andlaboratory tests to rule out hypertension, as well as pulmonary,cardiovascular, hepatic, haematological, renal and thyroiddisorders. Subjects over 40 years (with two exceptions, aged42 and 54 years, whose test data were not available) werealso evaluated with the Mini-Mental State exam (Folsteinet al., 1975), and only those who scoredù 28/30 underwenta PET study. Subjects were required to abstain from caffeineand nicotine for 4 h preceding the PET study. Womenreceiving oestrogen replacement therapy were also excluded.

PET scansIndividually fitted thermoplastic masks were used to minimizehead movements. Data were acquired with a ScanditronixPC2048-15B PET scanner (15 contiguous slices per scanwith spatial resolution 6–6.5 mm both in-plane and axially).Transmission scans were obtained with a rotating pin sourceof 68Ga/68Ge and were used to correct for radiation attenuationthrough skull and brain tissue. Each patient underwent fourPET scans, each following an i.v. bolus of ~42 mCi of15Owater, in a single scanning session. Emission data wereacquired in 16 frames over 4 min (12 frames of 10 s eachand four of 30 s each). Scan data were corrected forscatter, random coincidences and deadtime. Arterial inputfunctions were measured with automated arterial bloodsampling (Daube-Witherspoonet al., 1992), and absoluterCBF (ml/min/100 g) was calculated with a rapid leastsquares method (Koeppeet al., 1985) as describedelsewhere (Espositoet al., 1996).

Cognitive conditionsSubjects were administered four different cognitive tasks, oneper scan: the WCST, RPM and their respective sensorimotorcontrol tasks (WCSTc and RPMc) (see Bermanet al., 1988,1995). The sensorimotor control tasks were designed tomatch the WCST and the RPM for visual stimulation andresponse modality but did not involve higher order cognitiveprocessing and abstract reasoning. All tasks were begun 1min before the injection of tracer and continued throughoutthe 4 min of the scan. Subjects responded using a button-pressing device held in the right hand during the WCST andthe WCSTc. During RPM and RPMc, they verbally reportedthe number of the chosen response. Tasks were presented incounterbalanced order on a computer screen. The tasks wereself-paced in order to engage the neurophysiological ‘dutycycle’ fully by avoiding non-cognitively occupied time duringthe measurement period. Subjects were told to concentrateon the task and to try their best. For both tasks, the percentageof correct to total responses was calculated as the cognitiveoutcome measure.

The Wisconsin Card Sorting Task (WCST)The WCST is a problem-solving, abstract reasoning taskinvolving working memory (Goldman-Rakic, 1987; Bermanet al., 1995). It has been shown to be particularly sensitiveto dysfunction of the prefrontal cortex (Milner, 1963; Luria,1973), although its specificity has been questioned (Mountainand Snow, 1992), and it is known to activate prefrontal cortexphysiologically as well as an associated network of regionsincluding the inferior parietal lobule and the inferior posteriortemporal lobe (Bermanet al., 1995). During our computerizedWCST, subjects were required to match a central targetstimulus to four surrounding reference stimuli on the basisof three possible abstract categories: colour, shape and number(for sample stimuli, see Bermanet al., 1995). Subjects mustnot only discover the three correct matching principles, theymust also determine which is correct by using feedbackdisplayed on the screen that indicates whether each responseis right or wrong, maintain the rule through 10 trials, andthen use feedback to shift to a new category after it ischanged by the experimenter without warning to the subject.The ability to create an internal representation of the resultsof previous trials and then to use the internal representationas feedback to formulate a strategy for present and futureresponses, as well as the ability to shift abstract categorizationwhen needed, are critical for good performance and have allbeen linked to the prefrontal cortex (Goldman-Rakic, 1987;Berman et al., 1995). Previous studies have shown thatageing is associated with reduced performance on the WCST(Whehilan and Lesher, 1985; Daigneault and Braun, 1993;Nagahamaet al., 1997). The sensorimotor control task forthe WCST was a no-delay, matching-to-sample task whichwas of a visual complexity similar to the WCST. Subjectsmatched each central target stimulus to one of four unchangingsurrounding reference stimuli (for sample stimuli, see Bermanet al., 1995).

RPMRPM is an abstract reasoning task first developed in 1938(Raven, 1938). It is widely regarded as a good indicator ofgeneral intelligence (Jensen, 1982) and has found particularutility in patient populations in whom non-verbal tests aredesirable. As with the WCST, there is evidence thatperformance on the RPM declines in older subjects (forreviews, see Burke, 1985; Salthouse, 1992). During RPMperformance, subjects are shown patterns in a 33 3 cellmatrix with a missing piece [see, for example, Raven (1938);Carpenteret al. (1990) for sample stimuli]. They are requiredto determine which of six to eight possible alternatives bestcompletes the matrix so that the interrelational rules amongthe elements (e.g. along the rows and the columns) aresatisfied. Good RPM performance requires that subjectsperceive the relationships between cells in the matrix,determine the relationships between the columns and rowsof the matrix and then integrate this information. A crucial



ownloaded from





ownloaded from



difference between the RPM and the WCST is that while theWCST depends heavily upon internal representations ofprevious trials and the current conceptual set, all theinformation necessary to solve each RPM trial is availableexternally to the subject throughout that trial. The correctanswer does not depend on the results of previous trials asdoes the WCST. While it has been argued alternatively thatthe necessity of keeping several conceptual formulations inmind during RPM is itself a working memory function(Carpenteret al., 1990) involving the prefrontal cortex(Prabhakaranet al, 1997), studies of patients with lesionssuggest that postrolandic structures may be more critical forthis task (Bassoet al., 1973). The control task was ano-delay, matching-to-sample task. Subjects matched thereference stimulus displayed at the top of the screen to oneof eight possible answer stimuli displayed on the bottom.

Image processingFor each subject, the four PET scans were realigned withthe program of Woodset al. (1992). The realigned PETimages were roll–yaw corrected and then interpolated to 43slices, spatially normalized into the stereotaxic space of theTalairach and Tournoux atlas (1988), and smoothed with a20 3 20 3 12 mm filter in the x, y and z dimensions,respectively, with the SPM95 package (Fristonet al., 1995).For the voxel-level analyses, we chose to use rationormalization for global flow. This method assumes theexistence of a proportional relationship between regional andglobal CBF values (Shimosegawaet al., 1995), whereasANCOVA normalization assumes an additive relationship(Fristonet al., 1990). Normalized images from the respectivecontrol tasks were subtracted from the WCST and the RPMto isolate rCBF activation linked to the higher cognitiveprocessing during these two tasks. For voxel-by-voxel andcorrelational analyses, local maxima of the statistical resultswere found by searching within a volume of20 3 20 3 20 mm (103 10 3 5 voxels inx, y andz). Onlythose maxima representing volumes of.10 contiguousvoxels were considered further. Activation and correlationalmaps were thresholded atP , 0.005 for purposes ofillustration, but only those maxima significant atP , 0.001were tabulated and explored further withpost hocanalyses.

Fig. 1 (A) rCBF activation in the younger group during the WCST (top row) and the RPM (middle row). Red/yellow indicates greaterrCBF during the task (WCST or RPM) compared with the control (WCSTc or RPMc); blue indicates greater rCBF during the control(WCSTc or RPMc) compared with the task (WCST or RPM). The bottom row shows the comparison of rCBF activation between theWCST and the RPM paradigms. In red/yellow are shown areas in which the differences between task and control were greater during theWCST paradigm than during the RPM paradigm. In blue are shown areas in which the differences between task and control were greaterduring the RPM paradigm than during the WCST paradigm. (B) Voxel-by-voxel correlation maps between age and rCBF activationduring the WCST (top row) and the RPM (middle row). Only those voxels in which the correlation exceeded a threshold ofP , 0.005are highlighted. Red/yellow indicate areas where activation increased with age (positive correlations,P , 0.005); green indicatesnegative correlations (P , 0.005). The bottom row shows a voxel-by-voxel comparison of the age–rCBF correlations between the WCSTand RPM paradigms conducted with Williams’ test (P , 0.005). Here, red/yellow indicate areas in which the correlation coefficient wasgreater during the WCST; in green are areas in which the correlation coefficient was greater during RPM. WCST5 Wisconsin CardSorting Test; RPM5 Raven’s Progressive Matrices; rCBF5 regional cerebral blood flow.

Specific notations are made in the text wherepost hocresultsthat clarified the primary findings are reported at lessersignificance.

Statistical analysisEffect of ageing on global CBF and cognitiveperformanceAge-related changes in task performance were evaluated withPearson’s correlation analysis. Pearson’s correlation analysiswas also used to explore the relationship between normalageing and global CBF (determined as the average of allintracerebral voxels) for each of the four tasks. This was ofinterest given the fact that the literature is controversial withregard to the possibility of reduced global CBF (Dasturet al.,1963; Shawet al., 1984; for a review, see Meyeret al.,1994). Evaluation of age-related changes in global CBF isalso important methodologically because most strategies forregional activation analysis depend on some method ofnormalization for intertask and intersubject differences inglobal flow, a procedure that potentially can produceartefactual findings if between-task or between-group globaldifferences do exist.

rCBF activation in younger subjectsPrior to evaluating ageing effects on rCBF, we examined thedata for the 20 younger subjects of our population (using amedian split that yielded an age range of 18–42 years forthis subgroup) to ensure that their patterns of regionalactivation were consistent with previous reports and tocompare WCST and RPM activation directly. Meanperformance scores for this young subcohort fell, for boththe WCST and the RPM, within 1 SD of previous resultsobtained from larger populations of young control subjects(Ostrem et al., 1993; Bermanet al., 1995). For the twocognitive paradigms, task and control were compared withinstatistical parametric maps (SPM) using weighted linearcontrasts. The activation maps for the two paradigms werealso statistically compared directly. Age-related changes inrCBF activation were interpreted in light of the pattern ofregional activation in this younger group.



ownloaded from



Voxel-based correlations between age and rCBFactivationWe took advantage of the even age distribution of our cohortto search for linear relationships between rCBF activationand age. For the WCST and RPM paradigms, voxel-by-voxel Pearson’s Product Moment correlation analyses wereperformed between age and normalized rCBF activation (taskminus control) using software developed by our group at theNIMH (T.M.E., J.D.V.H. and G.E.). Local maxima of regionswith significant correlations were investigated further withinthe context of the standard task-related rCBF changes in theyounger group by tabulating for each task (i) the direction,correlation coefficient and significance level of the age–rCBFcorrelation and (ii) the direction and significance of the rCBFactivation (task minus control) at that locale in the youngerand older subgroups. Because we anticipated a declinein performance with age which could, itself, affect rCBFactivation, and because our primary goal was to test theimpact of age on rCBF activation, we also examined areaswith significant age–activation correlations after partiallingout the effect of performance (% correct responses). Wefurther explored with multiple regression analysis whetherthe variance in rCBF activation in these regions was accountedfor better by age, by performance or by age and performancelevel together.

Finally, the crucial step of this experiment was to determinewhether significantly different neurophysiological correlatesof ageing were demonstrated during the two tasks. To testthis hypothesis, we statistically compared, on a voxel-by-voxel basis, the correlation coefficients between age andWCST activation with those between age and RPM activation.For this purpose, we used a voxel-wise version (T.M.E.,J.D.V.H. and G.E.) of Williams’ test (Williams, 1959).Williams’ test is a well-validated statistical approach forassessing the equality of two correlation coefficients that aredependent (as they are in this experiment, since the WCSTand RPM rCBF data were collected from the same sample)(Neill and Dunn, 1975; Steiger, 1980). Williams’ statistic hasthe shape of at-distribution with d.f. 5 n–3, and itsmathematical representation is given by:

T2 5 (rjk–rjh){(n–1)(1 1 rkh)/[2(n–1/n–3)|R|1 r2(1–rkh)3]} 1/2;

where, in our case,r jk represents the matrix for the correlationbetween age (j) and WCST activation (k),r jh represents thematrix for the correlation between age (j) and RPM activation(h) andrkh represents the matrix for the correlation betweenWCST (k) and RPM activation (h). The latter takes intoaccount any correlation between WCST and RPM rCBFactivation data.

|R|5 (1–rjk2–rjh2–rkh

2) 1 (2rjkr jhrkh)

is the determinant of the 33 3 correlation matrix contain-ing the coefficients being tested; andr2 is expressed as1/2(rjk 1 r jh).

Voxels with significant Williams’ test were mapped onto

a representative MRI also spatially normalized to Talairachspace. The direction, magnitude and significance level ofeach local maximum in this correlational difference mapwere examined.

ResultsEffect of ageing on global blood flow andcognitive performanceThere was a significant negative correlation betweenage and percentage correct for both the WCST (r5 –0.53,P , 0.001) and RPM (r5 0.58, P , 0.001), but notfor the control tasks. Mean % correct scores6SD forthe young and the old group were, respectively: 76.16 12.5and 56.46 19.0 during the WCST, and 85.26 13.1 and61.16 22.4 during the RPM. No significant effect of ageingon global CBF was found for any of the four tasks (WCST:r 5 –0.24,P 5 0.14; WCSc:r 5 –0.22,P 5 0.16; RPM:r 5 –0.14,P 5 0.38; RPMc:r 5 –0.03,P 5 0.83).

CBF activation in younger subjectsIn our younger subjects, the patterns of regional activationfor both the WCST and the RPM (Fig. 1A, top two rows)were consistent with earlier studies (Ostremet al., 1993;Berman et al., 1995; Mattayet al., 1996). There were anumber of similarities between the two tasks. Both activatedthe dorsolateral [Brodmann area (BA) 9 and 46] prefrontalcortex, inferior parietal lobule (BA 39/40), anterior cingulate,and inferolateral temporal (BA 21 and 37) and occipital (BA18/19) cortex bilaterally. In both tasks, relative deactivationswere found in mesial polar portions of the prefrontal cortex(BA 10), perisylvian areas of the superior temporal andinferior temporal lobe (BA 38 and 22) and the posteriorcingulate (BA 23/31).

Important differences between the two tasks were alsoobserved (Fig. 1A, bottom row). The WCST activated aventral and more anterior area of the prefrontal cortex (BA10 and 47) that was not activated by RPM (between-taskcomparison,P , 0.0001), while RPM activated visuallyassociated areas including BA 18, 19 and 37 more than theWCST (P , 0.001). Portions of the inferior parietal lobule(BA 39) and the inferior temporal lobe (BA 20/21) wereactivated more by the WCST (between-task comparisonP , 0.0001). The right hippocampal/parahippocampal regionwas activated by RPM but not by the WCST (between-taskcomparison,P , 0.005).

Correlations between age and rCBF activationThe Wisconsin Card Sorting Task (WCST)paradigmDuring the WCST paradigm (Fig. 1B, top row, andTable 1), there were a number of areas in which therewas reduced activation as a function of increasing age



ownloaded from



Table 1 Wisconsin Card Sorting Test

BA Coordinate r P Activation in young Activation in old

x y z Z P Z P

Negative correlations between rCBF change and ageAnterior cingulate R 32 10 8 44 –0.5 ,0.001 1.79 ,0.05 –2.71 ,0.005Frontal–dorsolateral L 9 –32 4 36 –0.59,0.0001 3.30 ,0.001 –1.06 NSParietal–inferior lobule L 39/40 –40 –52 32 –0.55,0.001 4.17 ,0.0001 0.66 NSCerebellum–midline R 4 –88 –28 –0.58,0.0001 3.32 ,0.001 –1.51 NSCerebellum–posterior L –18 –96 –24 –0.6,0.0001 2.45 ,0.01 –2.14 ,0.05

Positive correlations between rCBF change and ageFrontal–mesial polar L 9 –8 56 24 0.6 ,0.0001 –4.16 ,0.0001 2.27 ,0.05Frontal–lateral polar R 9 20 42 28 0.51,0.001 –1.46 NS 2.49 ,0.01Occipital–cuneus L 18 –4 –96 16 0.5 ,0.001 –0.99 NS 2.82 ,0.005Occipital–cuneus R 17 8 –78 12 0.5 ,0.001 –2.46 ,0.01 1.38 NSParahippocampal R 30/19 26 –50 –4 0.55,0.001 –1.83 ,0.05 2.95 ,0.005

Performance expressed in % correct responses; L5 left; R 5 right; BA 5 Brodmann area;x, y, z 5 stereotaxic coordinates;x 5 medial–lateral distance from the middle (positive5 right); y 5 anterior–posterior distance relative to the anterior commissure(positive5 anterior);z 5 superior–inferior distance from the intercommissural line (positive5 superior); NS5 not significant.

(i.e. negative correlations between age and rCBF activation;in all cases P ø 0.001, r ø –0.5). These occurredexclusively in regions that were activated by the youngersubjects (Fig. 1A, top row, and Table 1): left DLPFC,left inferior parietal lobule, right anterior cingulate andcerebellum. These correlations mainly reflected an actualreversal of the quantitative relationship between the WCSTand its sensorimotor control task: areas activated by theWCST in younger subjects showed relatively less activityduring the WCST relative to the control task in oldersubjects (Table 1).

In contrast, significant increases in activation with age(P ø 0.001; r ù 0.5) occurred in regions where therewas no activation or even apparent deactivation (i.e.control . task) in the younger group (Fig. 1B, top row,and Table 1). These included left mesial and right lateralportions of the polar prefrontal cortex, the cuneus and theright parahippocampal gyrus. These correlations also mainlyreflected an actual reversal in the relationship between thetask and its control (see Table 1): areas normally deactivatedby the younger subjects became activated by the oldersubjects, particularly the mesial and lateral portions of thepolar prefrontal cortex, and the parahippocampal gyrus.

Partial correlations showed that the associations betweenage and rCBF activation largely remained significant atthe P , 0.005 level when performance was covaried out.The only exception was the positive correlation in theright polar prefrontal cortex which retained the directionof the relationship but became less robust (P, 0.05).Multiple regression analysis showed that the addition ofperformance level to age as an independent variable couldnot explain a significantly greater portion of the varianceof rCBF activation in these regions than age alone.

RPM paradigmFor the RPM paradigm (Fig. 1B, middle row, and Table2), like the WCST, age-related reductions of rCBF activation(negative correlations) also occurred mainly in areasactivated by the young subjects (Fig. 1A and Table 2):inferolateral temporal cortex including the fusiform gyrusbilaterally and the middle temporal gyrus on the left;portions of the left medial temporal cortex including theparahippocampal gyrus; the left inferior parietal lobule;and the cerebellum. Unlike the WCST, these negativecorrelations mainly represented areas in which oldersubjects showed reduced ability to activate regions normallyinvolved in RPM performance, rather than reversals in thefunctional relationship between the task and control.

Positive correlations between age and activation (Fig.1B, middle row, and Table 2) were found exclusively inareas normally deactivated in the younger group duringRPM performance compared with the RPMc: left and rightsuperior temporal gyri, left posterior cingulate and, as inthe WCST, mesial and lateral portions of the polarprefrontal cortex. These findings reflected age-relateddiminution or disappearance of the relative deactivation.In other words, both the positive and negative correlationsappear to represent, in general, an attenuation in the oldersubjects of the activation/deactivation pattern typically seenin the younger group.

Partial correlation analysis showed that the associationbetween age and rCBF activation remained significant atthe P , 0.005 level when performance was covaried out.The only exception was the right mesiopolar prefrontalcortex, where the direction of the correlation remained thesame but became less robust (P, 0.05). As in the WCST,multiple regression analysis showed that the addition ofperformance level to age as an independent variable did



ownloaded from



Table 2 Raven’s Progressive Matrices

BA Coordinate r P Activation in young Activation in old

x y z Z P Z P

Negative correlations between rCBF change and ageParietal–inferior lobule L 40 –30 –42 28 –0.53,0.001 2.70 ,0.005 –1.04 NSTemporal–fusiform L 37 –50 –52 –24 –0.62,0.0001 5.21 ,0.00001 3.04 ,0.005Temporal–fusiform R 37 54 –58 –16 –0.62,0.0001 5.71 ,0.00001 3.33 ,0.001Temporal–middle gyrus L 21 –52 –32 –4 –0.51,0.001 0.84 NS –3.89 ,0.0001Parahippocampal L 36 –22 –48 –20 –0.51,0.001 4.09 ,0.0001 0.64 NSCerebellum–midline L –4 –82 –28 –0.59 ,0.0001 4.26 ,0.0001 0.96 NSCerebellum–lateral R 28 –82 –28 –0.65 ,0.0001 5.88 ,0.00001 2.50 ,0.01Cerebellum–lateral R –30 –84 –28 –0.62 ,0.0001 5.04 ,0.00001 2.98 ,0.005

Positive correlations between rCBF change and agePosterior cingulate L 23 –12 –54 12 0.64,0.0001 –5.52 ,0.00001 0.93 NSFrontal–lateral polar L 9 –18 48 28 0.58,0.0001 –5.04 ,0.00001 –2.75 ,0.005Frontal–lateral polar L 10 –24 46 0 0.51,0.001 –2.57 0.01 0.48 NSFrontal–mesial polar R 10 10 52 0 0.53,0.001 –4.98 ,0.00001 –2.46 ,0.01Temporal–superior L 22 –50 –36 16 0.58,0.0001 –6.55 ,0.00001 –3.81 ,0.0001Temporal–superior R 22 58 –42 16 0.55,0.001 –5.72 ,0.00001 –3.56 ,0.001

Performance expressed in % correct responses; L5 left; R 5 right; BA 5 Brodmann area;x, y, z 5 stereotaxic coordinates;x 5 medial–lateral distance from the middle (positive5 right); y 5 anterior–posterior distance relative to the anterior commissure(positive5 anterior);z 5 superior–inferior distance from the intercommissural line (positive5 superior); NS5 not significant.

not explain a significantly greater portion of the varianceof rCBF activation in these regions than age alone.

Comparison of age–activation correlation mapsfor WCST and RPM paradigmsTo evaluate statistically whether the neurophysiological cor-relates of ageing were different for the two tasks, weperformed voxel-wise Williams’ tests. This analysis (Fig. 1B,bottom row, and Table 3) revealed a number of regions inwhich the relationship between age and rCBF activationdiffered across the two tasks, suggesting that demonstrableneurophysiological correlates of ageing depend upon theneural systems required for the particular cognitive operationsattempted. The left dorsolateral prefrontal cortex was activ-ated by both tasks in the younger cohort, but activationdecreased during the WCST only, and was actually deactiv-ated in the older group (Fig. 2). In other cases, thesecross-correlational differences occurred in areas where theactivation patterns of the two tasks differed in the youngercohort (cf. Fig. 1A, bottom row, with bottom row of Fig.1B). These included regions differentially activated acrossthe two tasks in young subjects that were not brought on-line as robustly by the older subjects: right inferior occipitalcortex and fusiform gyrus for RPM and left temporoparietalcortex for the WCST. Each of these represents a regionwhose activation is important for the particular task, but isreduced during that task in the older subjects (top two rowsof Fig. 1B). There were also regions normallynot a part ofthe task’s activation circuit (i.e. regions not activated, or evendeactivated by younger subjects), that were brought on-line,or were less deactivated, in a task-specific manner by theolder subjects: the right cuneus and the right parahippocampal

area for the WCST and the right superior temporal cortexfor RPM.

Our analyses also demonstrated some regions in which theeffects of ageing on rCBF activation were similar across theWCST and RPM, as seen in the top and middle rows of Fig.1B. Positive correlations in the polar portions of the prefrontalcortex (Fig. 2) and negative correlations in the cerebellumwere found in both task paradigms, and did not differaccording to Williams’ test (Fig. 1B, bottom; Fig. 2, top andbottom right).

DiscussionIn our cohort, cognitive performance on both the WCST andRPM declined with age, a finding in agreement with previousstudies of these and similar tasks (Burke, 1985; Whehilanand Lesher, 1985; Daigneault and Braun, 1993) and with thegeneral notion that working memory, abstract reasoningand problem-solving abilities deteriorate in older subjects(Botwinick, 1984). Also in this cohort, age-related changesin rCBF activation were observed during both tasks evenafter correction for changes in performance, and significantacross-task differences in the location and direction of therelationship between age and neurophysiological activationwere found. We did not, however, detect a significant correla-tion between global CBF values and age for any of the fourcognitive conditions. The absence of an age effect on globalbrain activity disagrees with some earlier studies demonstrat-ing a reduction of global values with increasing age (Kuhlet al., 1984; Shawet al., 1984; Gradyet al., 1990), but isconsistent with others (Dasturet al., 1963; Duaraet al.,1983; de Leonet al., 1984) in which, like ours, subjectswere screened carefully to exclude those with risk factors for



ownloaded from



Table 3 Statistical comparison of the age–rCBF acivation correlations for the WCST and the RPM (Williams’ test)

Williams’ test Correlation between age and rCBF activation Activation in young (Z) Activation in old (Z)

BA Coordinate r P WCST RPM WCST RPM WCST RPM

x y z r P r P

WCST . RPMOccipital–cuneus R 18 2 –84 12 4.02,0.001 0.45 ,0.005 –0.16 NS –1.30 0.30 1.78 –1.68Occipital–inferior R 18 28 –90 –8 3.92 ,0.001 0.30 0.06 –0.44 ,0.005 2.30 5.82 3.94 3.38Parahippocampal R 36 24 –50 –8 3.57,0.001 0.46 ,0.005 –0.19 NS –1.01 3.32 2.92 1.85Temporal–fusiform R 37 38 –54 –12 3.56,0.001 0.28 0.07 –0.36 ,0.05 0.54 5.87 2.35 4.88

RPM . WCSTFrontal–dorsolateral L 9 –32 4 36 3.76,0.001–0.59 ,0.001 0.02 NS 3.31 1.97 –1.06 1.51Temporal–superior R 22 60 –44 12 4.59,0.001–0.21 NS 0.50 ,0.001 0.08 –5.75 –2.29 –3.49Temporoparietal L 39 –48 –62 28 3.63,0.001–0.32 ,0.05 0.34 ,0.05 3.19 –1.91 0.48 0.09

WCST 5 Wisconsin Card Sorting Test; RPM5 Raven’s Progressive matrices; L5 left; R 5 right; BA 5 Brodmann area;x, y, z 5 stereotaxic coordinates;x 5 medial–lateral distance from the middle (positive5 right); y 5 anterior–posterior distance relative to the anterior commissure (positive5 anterior);z 5 superior–inferior distance from the intercommissural line (positive5 superior); NS5 not significant.

cerebrovascular disease. There is evidence that the presence ofsuch factors accelerates the age-related decline in globalvalues of CBF and glucose metabolism (for a review, seeMeyer et al., 1994).

The most important findings to emerge from this studywere the correlations between age and rCBF activation andthe across-task differences in this relationship. Technicalreasons for these findings, particularly those that could berelated epiphenomenologically to features of the ageing brain,must be considered. Although frank structural atrophy aswell as less apparent neuronal loss are not uncommon inageing, it is unlikely that these neurostructural changes playeda role. First, we carefully excluded subjects with MRIevidence of microvascular disease, significant atrophy orother anatomical abnormality. Secondly, we found similarage–activation relationships in an earlier analysis of thesedata (Espositoet al., 1995) in which stereotaxic normalizationwas not used and individual regions of interest were drawnon each subject’s MRI, thereby minimizing the possibility ofincreased partial volume effects in the older subjects (due tomore CSF-containing voxels being averaged with true braintissue if subclinical atrophy were present). Thirdly, it ishighly unlikely that partial volume effects could explainour several findings ofincreasedactivation with ageing,particularly where these represented actual reversals of therelationship between the task and its control, as seen for theWCST in the hippocampal/parahippocampal region; thisregion would be a likely candidate for partial volume effects,if they exist, because some investigators (de Leon, 1995),although not all (Sullivanet al, 1995), have suggested it isparticularly subject to neuronal loss with ageing. Finally, andperhaps most importantly, is our use of the within-subject,paired task activation method in which baseline sensorimotorand target tasks are compared. With this approach, non-specific effects of atrophy are likely to occur in the baselinetask as well as the task of interest and will, therefore, becontrolled for, an important advantage highlighted below.

Also, the across-paradigm statistical differences between theage–activation relationships that were found in several regionscannot be explained on the basis of atrophy.

The use of rCBF activation data (task minus control) hasthe crucial advantage of controlling for non-specific effectsof ageing and focuses the analysis on changes in regionalactivity that are related to the specific effects of age onhigher cognitive processing. However, this approach maymake it difficult to conclude firmly that the task-specific age-related changes we have reported are related specifically tothe cognitive aspects of the task, rather than the sensorimotoraspects of the experiment. To investigate this further, wetested, in those regions showing significant age–rCBF activa-tion correlations (Tables 1 and 2) or significantly differentage rCBF activation correlations across task (Table 3),whether similar relationships to age could also be found forunsubtracted rCBF measured during the WCST and RPMalone. For the two task paradigms considered separately, inall regions with negative correlations between age and rCBFactivation (Tables 1 and 2), negative correlations at the trendlevel or greater were indeed also seen between age and rCBFduring the WCST and RPM alone—with one exception,the left parahippocampal region during RPM. That thesedecreases with age could be demonstrated may link thesechanges more firmly to the cognitive aspects of the task, butalso may not be surprising given the fact that the relationshipbetween age and the activity of the grey matter as a wholeis dominated by negative trends, as has been reported byothers (for a review, see Meyeret al., 1994). This regionallynon-specific tendency for grey matter activity to decreasewith age, probably due to microscopic structural changesthroughout the brain (e.g. clinically inapparent neuronal loss),made it difficult to demonstrate in unsubtracted data most ofthe increases with age that were apparent in the activationdata (task minus control). These positive correlations betweenage and brain function during the tasks were largely onlyapparent when viewed against the baseline of the sensorimotor



ownloaded from



Fig. 2 Plots of the correlations between rCBF activation (task minus control) and age for the left dorsolateral prefrontal cortex (left) andleft polar prefrontal cortex (right) during the WCST (top) and RPM (bottom) paradigms; insets show the mean activation values (task –control) for the younger and older groups. In the left dorsolateral prefrontal cortex (left), activation decreased with age during the WCST(top left; r 5 –0.59,P , 0.0001; see also Table 1), whereas during RPM activation remained constant across the age span (bottom left;r 5 0.02, NS). The mean values for the older and younger groups indicate that the negative correlation during the WCST represents atendency in the older subjects to deactivate this normally activated region; in contrast, during RPM, there was virtually no difference inactivation between younger and older groups. In the left polar prefrontal cortex (right), positive correlations were found during both theWCST (top left;r 5 0.60,P , 0.0001) and RPM (bottom right;r 5 0.58,P , 0.0001) (see also Tables 1 and 2). For both tasks, theolder subjects showed less suppression of the polar prefrontal cortex than the young. During the WCST, polar prefrontal cortex tended tobecome activated compared with baseline in the older subjects, a reversal of the normal relationship between the task and its control.WCST 5 Wisconsin Card Sorting Test; WCSTc5 Wisconsin Card Sorting Test control; RPM5 Raven’s Progressive Matrices;RPMc 5 Raven’s Progressive Matrices control; rCBF5 regional cerebral blood flow.

controls. In some cases, the relationship between the taskand control actually reversed direction, while in others theneurophysiological differences between task and controlbecame less; these relative changes can, themselves, be seenas another example of the task-dependent changes which wediscuss below in the context of the different neural systemssubserving our two cognitive tasks. We believe that the factthat age-related changes in the control task may also occur[as has been demonstrated for simple perceptual matchingtasks (Gradyet al., 1994)] is another factor in favour of thepaired task approach we used, in which any changes due tosensory or motor aspects of performing our tasks are ‘base-lined’ out. It is thus important to emphasize that the within-task paradigm changes we report here occur relative to low

level sensorimotor tasks, which themselves may have age-related changes. However, because our two sensorimotorcontrol tasks, both being no-delay match-to-sample tasks,require the same cognitive operations and differ only inresponse mode (motor for the WCSTc and verbal for theRPMc), we hypothesized that any age-related changes inthem would have little effect in the across-task differencesin the relationship between age and rCBF activation (Table 3),the crux of our experiment. We thus predicted that thoseresults would be well supported by a Williams’ test foracross-task differences between the correlations of age andrCBF during the WCST itself, and of age and rCBF duringthe RPM itself (i.e. unsubtracted data). This prediction wasborn out: in every case listed in Table 3, except the superior



ownloaded from



temporal cortex, Williams’ test demonstrated differencesbetween the correlations of age and unsubtracted rCBF datafor the WCST and RPM (albeit in some cases at lessersignificance, probably due to better control of non-age-relatedvariance and non-specific between-subject effects when thecontrol task baselines are used). This result anchors the task-dependent nature of our neurophysiological results quitefirmly to differences in the higher cognitive operationsnecessary for the two tasks and proves that this task depend-ence can be demonstrated both in relation to, as well asindependently of, age-related changes in the control condition.

Another potential factor in our results could be that theolder subjects tended to have greater intersubject variabilityin task performance than the young subgroup. However, thiswas not reflected in greater variability in utilizing regionalcerebral resources. Formal statistical comparison of the vari-ability in rCBF activation for the older and younger cohorts(with Levene’s test) revealed no difference in any of theregions in which there were significant correlations betweenrCBF activation and age (Pù 0.05). Figure 2 graphicallydemonstrates one example of the absence of increased neuro-physiological variability with increasing age.

Task-specific changes in overall activationpatterns with ageThe neurofunctional manifestations of the ageing processrevealed by the within- and across-task correlationaldifferences suggest that different pathophysiologicalmechanisms are operational during these two tasks andunderlie the age-related cognitive impairments observed withthem. In general, where age-related changes occurred duringRPM, older subjects exhibited an overall attenuation ofthe activation/deactivation pattern (Table 2): areas normallyactivated relative to the control condition by the youngsubjects (e.g. parahippocampal region, inferior parietal lobuleand fusiform gyrus) were not recruited as robustly by theolder subjects; areas normally significantly suppressed (e.g.polar frontal cortex and superior temporal lobe) were lesssuppressed relative to the control. In other words, there wasfailure to produce a focused neural response by engagingtask-appropriate neural activity patterns and inhibitinginappropriate ones; the neural activity patterns during theRPM and the RPMc (and perhaps the cognitive operationsemployed during them) became more alike in the oldercohort. This finding is similar to electrophysiologicalobservations that older subjects do not produce the expecteddifferential activity patterns when challenged with differenttypes of stimuli, such as novel and target items (Friedmanand Simpson, 1994). Behavioural observations that oldersubjects have difficulty with mental flexibility and cognitivefocus (Hochanadel and Kaplan, 1994) may reflect theseneurobiological underpinnings.

During the WCST, on the other hand, the age-relatedchanges were manifest not only as failure of some normally

activated regions to come on-line relative to the controlcondition (e.g. dorsolateral prefrontal cortex, inferior parietallobule and anterior cingulate), but also as recruitment abovethe control level of regions that typically are not activated,or that are even relatively deactivated, by the young cohort(frontopolar cortex, cuneus and parahippocampal gyrus) aswell as suppression relative to control of regions that typicallyare recruited or unchanged in the younger cohort (Table 1).These results can, like the RPM results, be viewed asfailure to focus neurophysiologically by engaging appropriateregional activity and suppressing non-task-related regions,but in this instance may also represent the use of alternativecircuitry in an attempt to compensate for the older subjects’inability to bring on-line the appropriate neural network.

Regional task-specific changes in activationwith ageAt the level of individual regions, it also appears that manyof the identified neurophysiological correlates of ageing weretask-specific, depending upon the neural circuitry necessaryfor the cognitive behaviour. The prefrontal cortex is a casein point. We entered this investigation with the hypothesisthat prefrontal functional correlates of ageing might be fluid,varying with the role of the region in performance of theparticular task. This proved to be true in one prefrontal area,the left DLPFC (BA 9/45–46). This region was activated byboth tasks in the young subgroup, but the effects of age onthis activation were significantly different across the WCSTand RPM paradigms (Table 3, Figs 1B and 2). DLPFCbecamedeactivated with increasing age during the WCST, butremained constantly activated over the age span during RPM.

In contrast, the age–activation relationships were similarfor both tasks in another prefrontal region, the polar prefrontalcortex (Fig. 1B), which normally is deactivated in both theWCST and RPM paradigms (albeit more so for RPM, Fig.1A). A positive correlation with age was observed in bothtasks (Tables 1 and 2, and top two rows of Fig. 1B). Thisnormally supressed area actually became activated abovebaseline in the older subjects during the WCST (Table 1, topright of Fig. 2); during the RPM, there was less functionalsuppression relative to baseline in the older subjects (Table2, bottom right of Fig. 2), but they did not activate abovethe control task level. Thus, despite the fact that the directionand magnitude of the correlations in this area were similar forthe two tasks, the underlying pathophysiological mechanism,even here, could still be subtly different: failure to inhibit inRPM appropriately versus actual recruitment of alternative,normally suppressed regions in the WCST (see below).

In more posterior portions of the brain, specifically theinferior parietal lobule (BA 39/40) and inferior temporalcortex (BA 21/37), across-task differences in the age–activation relationship were also demonstrated statistically(see bottom row of Fig. 1B). Interestingly, the directions ofthe differences were opposite in these two regions and



ownloaded from



reflected the directions of the activation patterns in youngsubjects. In the inferior parietal lobule (BA 39/40), which isactivated more robustly during the WCST than RPM inyoung subjects (Fig. 1A, bottom row,128 mm), there wassignificantly greater reduction in activation with age duringthe WCST. In inferior temporal areas, which are activatedmore robustly during the RPM than during the WCST inyoung subjects, there was significantly greater reduction withage during RPM than there was during WCST.

These data clearly demonstrate that the age-relatedneurofunctional changes in these regions are context-dependent and reflect (i) the differential normal physiologicalresponse during the tasks and (ii) the related issue ofthe importance of the regions for the particular cognitiveoperations having primacy in the tasks. More specifically,the across-task differences appear to reflect the morecircumscribed dependence of the WCST on working memoryand prefrontal cortical systems, and of RPM on visuospatialprocessing systems, computational problem solving andpostrolandic regions. This hypothesis is suggested by theobservation that, for example, the DLPFC and the inferiorparietal lobule showed significantly greater age-relateddecreases in activation during the WCST than RPM, andthese areas have been linked specifically to working memoryand to the WCST (Friedman and Goldman-Rakic, 1994;Bermanet al., 1995). On the other hand, the inferolateraltemporal cortex, which has been related more to visualprocessing (Haxbyet al., 1991) than to working memoryperse, showed significantly greater age-related decline in functionduring RPM than WCST, and RPM is more highly dependenton visual processing and more sensitive to damage in thisarea than in prefrontal cortex (Bassoet al., 1973).

Our data also provide evidence that the functionalrelationship between regions may be altered with ageing.The parahippocampal/hippocampal region is activatedconsistently in young subjects during RPM, but normally isnot activated or is even deactivated during the WCST (Ostremet al., 1993; Bermanet al., 1995; Mattayet al., 1996). Wefound that the relationship between age and rCBF activationin this region was significantly different across our two taskparadigms, with an age-related increase in parahippocampal/hippocampal activation during the WCST and a decreaseduring the RPM. These findings, together with those in theDLPFC (i.e. decreased activation with age for the WCSTand no change during RPM), suggest that alteration ofthe normal prefrontal–hippocampal functional relationshipneeded for good performance during the WCST and theRPM (Mattayet al., 1996) is altered in a task-specific manneras age increases. To test this, we performed apost hocthree-way ANOVA with one independent measure (group: oldversus young) and two repeated measures (task: RPM versusWCST; and region: hippocampal area versus DLPFC). Thegroup 3 task 3 region interaction was highly significant[F(1,38) 5 13.3, P , 0.0008, Fig. 3]. Furthermore, duringthe WCST, a negative correlation between left DLPFC andright parahippocampal/hippocampal gyrus was found in the

Fig. 3 Task-specific changes in the prefrontal–hippocampalfunctional interaction with age assessed by three-way analysis ofvariance (age group3 task3 region):F(1,38) 5 13.3,P , 0.0008. In the young group (continuous line), activation ofthe left dorsolateral prefrontal cortex was associated with righthippocampal suppression during the WCST, and with righthippocampal activation during the RPM. In the old group (dashedline), the pattern was largely reversed: reduced activation in theleft dorsolateral prefrontal cortex was associated with increasedright hippocampal activation during the WCST, whereas duringRPM, preserved left dorsolateral prefrontal cortical activation wasassociated with reduced hippocampal activation. WCST5Wisconsin Card Sorting Test; RPM5 Raven’s ProgressiveMatrices; LDLPFC5 left dorsolateral prefrontal cortex; RHipp5right hippocampal region.

young subjects (r5 –0.64; P 5 0.005), while no suchrelationship was found in the old subjects (r5 –0.33;P . 0.16). Interestingly, during RPM, there was no correlationbetween these two regions in the young subjects, but anegative interaction emerged in the older cohort (r5 –0.48;P , 0.05), resembling the one that was operational in theyoung subjects during the WCST. One interpretation of theseresults is that some functional relationships that are importantfor task performance are lost with ageing, while newfunctional links not typically present may be generated,perhaps in compensation. These results are consistent withprevious findings that these two regions are particularlysensitive to the effects of ageing (Gradyet al., 1995; Schacteret al. 1996; for reviews, see Kemper, 1994; de Leon, 1995).Our data extend those findings by demonstrating that thefunctional relationship between these two regions and theage-related changes in this relationship are context-dependentand depend upon the neural systems most necessary forthe tasks.

To address in a more global sense whether neural networksfunctional in young subjects are also present in the oldersubjects, we expanded thispost hocexploratory covarianceanalysis to examine a larger set of potential network nodes.Using the method of Cabezaet al. (1997b), we selected fromSPM activation analyses a set of voxels for each task activatedor deactivated most robustly by the younger subjects. Thesecandidate constituents of the neural networks engaged during



ownloaded from



WCST and RPM performance in young subjects were chosenbilaterally from DLPFC, polar prefrontal cortex, anteriorcingulate, inferior parietal lobule, parahippocampal/hippocampal region, inferolateral temporal cortex andoccipital cortex. From these voxels, we then extracted foreach of the two tasks the patterns of regional covariance inthe younger subjects, as indicators of interregional functionalinteractions present during cognitive performance (Horwitzet al., 1991; Cabezaet al., 1997b). The same exploratoryanalysis for the same voxels was applied to the older groupto identify age-related changes in the utilization of the neuralnetworks engaged during our two tasks. During the WCST,anterior–posterior positive covariances were found in theyoung subjects between the left DLPFC and the left andright inferior parietal lobule (P, 0.02), while inverserelationships were found between the left DLPFC and theleft polar prefrontal cortex (P, 0.05) and, as mentionedabove, between the left DLPFC and the right parahippocampalregion (P 5 0.005). However, in the old subjects, therewas a dissolution of this covariance pattern, with only therelationship between left DLPFC and right inferior parietallobule remaining. During the RPM, as expected, youngsubjects showed a different pattern of functional interactions:the right parahippocampal region positively correlated withthe right inferolateral temporal cortex and with the rightoccipital cortex (P, 0.01), as did the right inferolateraltemporal cortex with the right inferior parietal lobule (P50.05), while negative interactions between left polar prefrontalcortex and right inferior parietal lobule and between rightpolar prefrontal cortex and right lateral prefrontal cortex(P , 0.05) were observed. In the older subjects, theseinteractions disappeared, with the exception of that betweenthe right parahippocampal gyrus and the right inferolateraltemporal cortex. Interestingly, in addition to the newinverse relationship between left DLPFC and the rightparahippocampal region mentioned above, new positiverelationships not seen in the young group arose in the oldercohort: right DLPFC activation became correlated with rightinferior parietal lobule and with right occipital cortexactivation (P, 0.05). Thesepost hocanalyses suggest thatprefrontal–parietal functional interactions within theworking memory system and temporal–parietal–hippocampalfunctional interactions within posterior visuospatialprocessing systems are altered in older subjects. These resultsalso lend further support to the more general notion thattask-specific changes in neural systems occur with ageing.

The context-dependence of these age-related changes andthe fact that the functional relationship between regions alsochanges in a task-specific way, even in the setting of twotask paradigms with considerable cognitive overlap, are mostconsistent with the notions of disconnectivity and systemsfailure. Disconnectivity, i.e. abnormal functional integrationacross a distributed network of regions, has been consideredas a basis for cognitive dysfunction in a variety of clinicalsettings (Friston and Frith, 1995). The related term, systemfailure, refers to functional abnormalities that may become

apparent when cognitive or other demands call for a networkof regions to be brought on-line in a coordinated manner;therefore, such abnormalities may, indeed, be contextdependent. Whether the prefrontal cortex has a special rolein such disconnectivity and system failure, as could besurmized from its central executive function, is a topic formore extensive analysis in future research, perhaps usingstructural equation modelling.

Possible mechanismsThere is considerable speculation in the literature aboutthe interpretation and source of age-related neurofunctionalchanges. At the cognitive level, our finding of context-dependent decreases in neural recruitment in areas importantfor cognitive performance could represent the primaryneurobiological underpinnings of reduced mental capacity.This concept may be particularly relevant in the prefrontalworking memory system which is, by definition, of finitecapacity (Baddeley, 1992) and which capacity may be limitedfurther with ageing. Our findings of increased activation withage in areas not usually activated could represent secondarychanges, i.e. the use of alternative cognitive strategies in theface of those primary limitations (e.g. hippocampal activationduring the WCST because of the use of long-term insteadof prefrontal short-term working memory mechanisms).Alternatively, these increases could themselves representprimary neurophysiological changes, such as functionalinability to suppress activity in regions not important for thecognitive operation (e.g. in polar prefrontal cortex for bothtasks), as has been suggested by a number of evoked potentialstudies (e.g. Friedman and Simpson, 1994; Nielsen-Bohlmanand Knight, 1995). This possibility is particularly intriguinggiven recent suggestions that suppression of rCBF in areasunrelated to the task may be an important mechanism forfocusing attention (Roland, 1982; Haxbyet al., 1994; Mattayet al., 1996). Hasher and Zacks (1988) argue that inhibitionof specific regions in order to suppress processing of irrelevantstimuli is defective in the elderly and is responsible forreduced cognitive performance when the frontal lobeparticipates in the cognitive operation. Dysfunction of theseinhibitory mechanisms would admit extraneous informationthat would engage the prefrontal system and use neural time,attention and memory resources that should be focused onprocessing relevant information. The outcome would be arelative reduction of the capacity of the central executive ofthe working memory system and of the ability of the prefrontalcortex to coordinate and regulate the activity of other regionsduring cognition, in other words, system failure.

It is instructive to view in light of the capacity-limitednature of working memory the various neurophysiologicalage effects that occur when DLPFC, itself, is faced withdifferent cognitive demands. In the present study, DLPFCactivation is decreased with ageing in the context of a taskin which DLPFC is activated in young subjects, workingmemory plays a primary role and working memory capacity



ownloaded from



is exceeded in ageing (i.e. the WCST, top left of Fig. 2). Inour other task, RPM, DLPFC is also activated in youngsubjects, but working memory is not primary to the task andplays a role which may remain within its capacity withageing; in that context, DLPFC activation remained constantover the age span (bottom left of Fig. 2). In other studies oftasks in which working memory and related prefrontalsystems normally play no role and in which DLPFC is notactivated in young subjects [visuospatial processing of facesand location (Gradyet al., 1994) and memory recall (Cabezaet al., 1997a)], DLPFC activation actually increased withageing; in that context, working memory and prefrontalsystems may have been called upon as a compensatorycognitive strategy in the face of other extra-frontal deficitsmore primary to the tasks (Gradyet al., 1994).

The dopamine system is an attractive candidate in whichto search for a common mechanism for these findings becausealterations in dopamine receptors, particularly of the D2 type(Volkow et al., 1996a; Wonget al., 1997), and of presynapticdopamine transporters (van Dycket al., 1995; Volkowet al.,1996b) have been documented in studies of human ageingand because dopamine plays a special role in prefrontalcortex. In both animals and humans, monoamines modulatecortical signal-to-noise, suppressing spontaneous backgroundneural firing and specifically enhancing cortical neuralresponses to relevant stimuli (Footeet al., 1975; Sawaguchi,1987; Arnsten and Goldman-Rakic, 1990; Danielet al., 1991;Mattay et al., 1996). The context-dependent changes inactivation pattern we observed with ageing are wellconceptualized as diminished signal-to-noise. In the prefrontalcortex, an optimal range of dopaminergic tone is required foroptimal function (Desimone, 1995; Williams and Goldman-Rakic, 1995). Moreover, in a previous study (Mattayet al.,1996), we augmented monoaminergic function withdextroamphetamine while young subjects performed the sametwo tasks as in the present study and found that during theWCST, activation increased in DLPFC (the ‘signal’) anddecreased in hippocampus (the ‘noise’), while exactly theopposite pattern of changes occurred during RPM.Collectively, those task-specific changes in the functionalrelationship between DLPFC and hippocampus withamphetamine are, in turn, opposite to those that occurredwith ageing in the present study (Fig. 3). In a recentstudy of rodents, D1 receptors were shown to modulatehippocampal–prefrontal cortical circuits (Seamanset al.,1998). Further neuroimaging studies in older subjects withpharmacological manipulation of monoaminergic neuro-transmission may clarify the role of this system in cognitivedeficits in ageing.

ConclusionsWe used two tasks with considerable cognitive overlap, butsome important differences in cognitive features and neuralsubstrate. Those differences were reflected in differentialneurophysiological manifestations of ageing during

performance of the two tasks. The context dependency ofthe neurofunctional findings is most consistent with thenotions of system failure and disconnectivity, but prefrontalimpairment may play a special role. The older subjects’impaired neurophysiological focus—their inability to bringthe appropriate systems on-line while suppressing others—appears to underly their impaired performance on both tasks.In this study, we took advantage of our cohort’s even agedistribution to search for linear relationships between ageand rCBF activation. It is possible that non-linear analyseswould yield additional information. New statisticalapproaches, such as that employed in the present study,may help to explore further the important role that alteredinterregional functional connectivity appears to have in thedecline of some higher cognitive functions with age(Wickelgren, 1996). The present study has the limitations ofany cross-sectional investigation on the effects of normalageing. Within-subject longitudinal studies are likely to haveincreased sensitivity to age-related neurofunctional changes.Finally, this work emphasizes that tasks accessing relevantneural systems must be used to measure neurofunctionalchanges with ageing and that those changes must beinterpreted in light of the role of the particular neural structurein the cognitive operation being studied.

References

Arnsten AFT, Goldman-Rakic PS. Stress impairs prefrontal cortexcognitive function in monkeys: role of dopamine [abstract]. SocNeurosci Abstr 1990; 16: 164.

Baddeley A. Working memory. Science 1992; 255: 556–9.

Basso A, De Renzi E, Faglioni P, Scotti G, Spinnler H.Neuropsychological evidence for the existence of cerebral areascritical to the performance of intelligence tasks. Brain 1973; 96:715–28.

Belleville S, Peretz I, Malenfant D. Examination of the workingmemory components in normal aging and in dementia of theAlzheimer type. Neuropsychologia 1996; 34: 195–207.

Berman KF, Illowsky BP, Weinberger DR. Physiological dysfunctionof dorsolateral prefrontal cortex in schizophrenia. IV. Furtherevidence for regional and behavioral specificity [see comments].Arch Gen Psychiatry 1988; 45: 616–22. Comment in: Arch GenPsychiatry 1989; 46: 758–60.

Berman KF, Ostrem JL, Randolph C, Gold J, Goldberg TE,Coppola R, et al. Physiological activation of a cortical networkduring performance of the Wisconsin Card Sorting Test: a positronemission tomography study. Neuropsychologia 1995; 33: 1027–46.

Botwinick J. Aging and behavior: a comprehensive integration ofresearch findings. New York: Springer; 1984.

Burke HR. Raven’s Progresssive Matrices (1938): more on norms,reliability, and validity. J Clin Psychol 1985; 41: 231–5.

Cabeza R, Grady CL, Nyberg L, McIntosh AR, Tulving E, Kapur S,et al. Age-related differences in neural activity during memory



ownloaded from



encoding and retrieval: a positron emission tomography study.J Neurosci 1997a; 17: 391–400.

Cabeza R, McIntosh AR, Tulving E, Nyberg L, Grady CL. Age-related differences in effective neural connectivity during encodingand recall. Neuroreport 1997b; 8: 3479–83.

Carpenter PA, Just MA, Shell P. What one intelligence test measures:a theoretical account of the processing in the Raven ProgressiveMatrices Test. Psychol Rev 1990; 97: 404–31.

Craik FI, Jennings JM. Human memory. In: Craik FI, SalthouseTA, editors. The handbook of aging and cognition. Hillsdale (NJ):Lawrence Erlbaum; 1992. p. 51–110.

Daigneault S, Braun CM. Working memory and the Self-OrderedPointing Task: further evidence of early prefrontal decline in normalaging. J Clin Exp Neuropsychol 1993; 15: 881–95.

Daniel DG, Weinberger DR, Jones DW, Zigun JR, Coppola R,Handel S, et al. The effect of amphetamine on regional cerebralblood flow during cognitive activation in schizophrenia. J Neurosci1991; 11: 1907–17.

Dastur DK, Lane MH, Hansen DB, Kety SS, Butler RN, Perlin S,et al. Effects of aging on cerebral blood circulation and metabolismin man. In: Birren JE, Butler RN, Greenhouse SW, Sokoloff L,Yarrow MR, editors. Human aging: a biological and behavioralstudy. Bethesda (MD): U.S. Department of Health, Education, andWelfare; 1963. p. 59–76.

Daube-Witherspoon ME, Chon KS, Green SL, Carson RE,Herscovitch P. Factors affecting dispersion correction for continuousblood withdrawal and counting systems [abstract]. J Nucl Med1992; 33: 1010.

de Leon MJ, George AE, Ferris SH, Christman DR, Fowler JS,Gentes CI, et al. Positron emission tomography and computedtomography assessments of the aging human brain. J Comput AssistTomogr 1984; 9: 88–94.

de Leon MJ, Convit A, DeSanti S, Golomb J, Tarshish C, Rusinek H,et al. The hippocampus in aging and Alzheimer’s disease. [Review].Neuroimaging Clin N Am 1995; 5: 1–17.

Desimone R. Neuropsychology. Is dopamine a missing link? [news;comment]. Nature 1995; 376: 549–50. Comment on: Nature 1995;376: 572–5.

Devous MD Sr, Stokely EM, Chehabi HH, Bonte FJ. Normaldistribution of regional cerebral blood flow measured by dynamicsingle-photon emission tomography. J Cereb Blood Flow Metab1986; 6: 95–104.

Duara R, Margolin RA, Robertson-Tchabo EA, London ED,Schwartz M, Renfrew JW, et al. Cerebral glucose utilization, asmeasured with positron emission tomography in 21 resting healthymen between the ages of 21 and 83 years. Brain 1983; 106: 761–75.

Esposito G, Kirkby BS, Van Horn JD, Ostrem JL, Weinberger DR,Berman KF. Task-specific neurophysiological changes in normalaging: PET rCBF measurements during cognition [abstract]. J NuclMed 1995; 36: 21p.

Esposito G, Van Horn JD, Weinberger DR, Berman KF. Genderdifferences in cerebral blood flow as a function of cognitive statewith PET. J Nucl Med 1996; 37: 559–64.

Folstein MF, Folstein SE, McHugh PR. ‘Mini-mental state’. Apractical method for grading the cognitive state of patients for theclinician. J Psychiatr Res 1975; 12: 189–98.

Foote SL, Freedman R, Oliver AP. Effects of putativeneurotransmitters on neuronal activity in monkey auditory cortex.Brain Res 1975; 86: 229–42.

Friedman HR, Goldman-Rakic PS. Coactivation of prefrontal cortexand inferior parietal cortex in working memory tasks revealed by2DG functional mapping in the rhesus monkey. J Neurosci 1994;14: 2775–88.

Friedman D, Simpson G. Amplitude and scalp distribution of targetand novel events: effects of temporal order in young, middle agedand older adults. Cogn Brain Res 1994; 2: 49–63.

Friston KJ, Frith CD. Schizophrenia: a disconnection syndrome?[Review]. Clin Neurosci 1995; 3: 89–97.

Friston KJ, Frith CD, Liddle PF, Dolan RJ, Lammertsma AA,Frackowiak RS. The relationship between global and local changesin PET scans [see comments]. J Cereb Blood Flow Metab 1990;10: 458–66. Comment in: J Cereb Blood Flow Metab 1993; 13:1038–40.

Friston KJ, Holmes AP, Worsley KJ, Poline JB, Frith CD,Frackowiak RS. Statistical parametric maps in functional imaging:a general linear approach. Hum Brain Mapp 1995; 2: 189–210.

Goldman-Rakic PS. Circuitry of primate prefrontal cortex andregulation of behavior by representational memory. In: MountcastleV, Plum F, editors. Handbook of physiology, Sect. 1, Vol. 5, Pt. 1.Bethesda (MD): American Physiological Society; 1987. p. 373–417.

Grady CL, Horwitz B, Schapiro MB, Rapoport SI. Changes in theintegrated activity of the brain with healthy aging and dementia ofthe Alzheimer type. In: Battistin L, Gerstenbrand F, editors. Agingbrain and dementia: new trends in diagnosis and therapy. New York:John Wiley; 1990. p. 355–70.

Grady CL, Maisog JM, Horwitz B, Ungerleider LG, Mentis MJ,Salerno JA, et al. Age-related changes in cortical blood flowactivation during visual processing of faces and location. J Neurosci1994; 14: 1450–62.

Grady CL, McIntosh AR, Horwitz B, Maisog JM, Ungerleider LG,Mentis MJ, et al. Age-related reductions in human recognitionmemory due to impaired encoding. Science 1995; 269: 218–21.

Hasher L, Zacks RT. Working memory, comprehension and aging:a review and a new view. In: Bower GH, editor. The psychologyof learning and motivation, Vol. 22. San Diego: Academic Press;1988. p. 193–225.

Haxby JV, Grady CL, Horwitz B, Ungerleider LG, Mishkin M,Carson RE, et al. Dissociation of object and spatial visual processingpathways in human extrastriate cortex. Proc Natl Acad Sci USA1991; 88: 1621–5.

Haxby JV, Horwitz B, Ungerleider LG, Maisog JM, Pietrini P,Grady CL. The functional organization of human extrastriate cortex:a PET–rCBF study of selective attention to faces and locations. JNeurosci 1994; 14: 6336–53.

Hochanadel G, Kaplan E. Neuropsychology of normal aging. In:Albert ML, Knoefel JE, editors. Clinical neurology of aging. 2nded. New York: Oxford University Press; 1994. p. 231–44.



ownloaded from



Horwitz B. Functional interactions in the brain: use of correlationsbetween regional metabolic rates. J Cereb Blood Flow Metab 1991;11: A114–20.

Jensen AR. The chronometry of intelligence. In: Sternberg RJ,editor. Advances in the psychology of human intelligence, Vol. I.Hillsdale (NJ): Lawrence Erlbaum; 1982. p. 255–310.

Kemper TL. Neuroanatomical and neuropathological changes duringaging and dementia. In: Albert ML, Knoefel JE, editors. Clinicalneurology of aging. 2nd ed. New York: Oxford University Press;1994. p. 3–67.

Koeppe RA, Holden JE, Ip WR. Performance comparison ofparameter estimation techniques for the quantitation of local cerebralblood flow by dynamic positron computed tomography. J CerebrBlood Flow Metab 1985; 5: 224–34.

Kuhl DE, Metter EJ, Riege WH, Hawkins RA. The effect of normalaging on patterns of local cerebral glucose utilization. Ann Neurol1984; 15 Suppl: S133–7.

Light LL. Adult age differences in reasoning from new information.J Exp Psychol Learn Mem Cognit 1982; 8: 435–47.

Luria AR. The working brain. Harmondsworth (UK): Penguin; 1973.

Martin AJ, Friston KJ, Colebatch JG, Frackowiak RS. Decreases inregional cerebral blood flow with normal aging. J Cereb BloodFlow Metab 1991; 11: 684–9.

Mattay VS, Berman KF, Ostrem JL, Esposito G, Van Horn JD,Bigelow LB, et al. Dextroamphetamine enhances ‘neural network-specific’ physiological signals: a positron emission tomographyrCBF study. J Neurosci 1996; 16: 4816–22.

Meyer JS, Kawamura J, Terayama Y. Cerebral blood flow andmetabolism with normal and abnormal aging. In: Albert ML,Knoefel JE, editors. Clinical neurology of aging. 2nd ed. New York:Oxford University Press; 1994. p. 214–34.

Milner B. Effects of different brain lesions on card sorting. ArchNeurol 1963; 9: 90–100.

Morris RG, Craik FI, Gick ML. Age differences in working memorytasks: the role of secondary memory and the central executivesystem. Q J Exp Psychol 1990; 42A: 67–86.

Mountain MA, Snow WG. Wisconsin Card Sorting test as a measureof frontal pathology. Clin Psychol 1992; 7: 108–18.

Nagahama Y, Fukuyama H, Yamauchi H, Katsumi Y, Magata Y,Shibasaki H, et al. Age-related changes in cerebral blood flowactivation during a Card Sorting Test. Exp Brain Res 1997; 114:571–7.

Naritomi H, Meyer JS, Sakai F, Yamaguchi F, Shaw T. Effects ofadvancing age on regional cerebral blood flow. Studies in normalsubjects and subjects with risk factors for atherothrombotic stroke.Arch Neurol 1979; 36: 410–6.

Neill JJ, Dunn OJ. Equality of dependent correlation coefficients.Biometrics 1975; 31: 531–43.

Nielsen-Bohlman L, Knight RT. Prefrontal alterations duringmemory processing in aging. Cereb Cortex 1995; 5: 541–9.

Offenbach SI. A developmental study of hypothesis testing and cueselection strategies. Dev Psychol 1974; 10: 484–90.

Ostrem JL, Berman KF, Mattay VS, Weinberger DR. Cerebralactivation during problem solving and abstract reasoning asdemonstrated with PET: a neural network subserving intelligence[abstract]. Soc Neurosci Abstr 1993; 19: 792.

Pantano P, Baron JC, Lebrun-Grandie P, Duquesnoy N, Bousser MG,Comar D. Regional cerebral blood flow and oxygen consumption inhuman aging. Stroke 1984; 15: 635–41.

Prabhakaran V, Smith JA, Desmond JE, Glover GH, Gabrieli JD.Neural substrates of fluid reasoning: an fMRI study of neocorticalactivation during performance of the Raven’s Progressive MatricesTest. Cognit Psychol 1997; 33: 43–63.

Raven JC. Guide to progressive matrices. London: H. K. Lewis;1938.

Richardson JT. Evolving concepts of working memory. In:Richardson JT, Engle RW, Hasher L, Logie RH, Stoltzfus ER, ZacksRT, editors. Working memory and human cognition. New York:Oxford University Press; 1996. p. 15–7.

Roland PE. Cortical regulation of selective attention in man. Aregional cerebral blood flow study. J Neurophysiol 1982; 48:1059–78.

Salthouse TA. Working memory as a processing resource in cognitiveaging. Dev Rev 1990; 10: 101–24.

Salthouse TA. Reasoning and spatial abilities. In: Craik FI, SalthouseTA, editors. Handbook of aging and cognition. Hillsdale (NJ):Lawrence Erlbaum; 1992. p. 167–211.

Sawaguchi T. Catecholamine sensitivities of neurons related to avisual reaction time task in the monkey prefrontal cortex. JNeurophysiol 1987; 58: 1100–22.

Schacter DL, Savage CR, Alpert NM, Rauch SL, Albert MS. Therole of hippocampus and frontal cortex in age-related memorychanges: a PET study. Neuroreport 1996; 7: 1165–9.

Seamans JK, Floresco SB, Phillips AG. D1 receptor modulation ofhippocampal–prefrontal cortical circuits integrating spatial memorywith executive functions in the rat. J Neurosci 1998; 18: 1613–21.

Shaw TG, Mortel KF, Meyer JS, Rogers RL, Hardenberg J,Cutaia MM. Cerebral blood flow changes in benign aging andcerebrovascular disease. Neurology 1984; 34: 855–62.

Shimosegawa E, Kanno I, Hatazawa J, Fujita H, Iida H, Miura S,et al. Photic stimulation study of changing the arterial partialpressure level of carbon dioxide. J Cereb Blood Flow Metab 1995;15: 111–4.

Steiger JH. Tests for comparing elements of a correlation matrix.Psychol Bull 1980; 87: 245–51.

Sullivan EV, Marsh L, Mathalon DH, Lim KO, Pfefferbaum A.Age-related decline in MRI volumes of temporal lobe gray matterbut not hippocampus. Neurobiol Aging 1995; 16: 591–606.

Talairach J, Tournoux P. Co-planar stereotaxic atlas of the humanbrain. Stuttgart: Thieme; 1988.

van Dyck CH, Seibyl JP, Malison RT, Laruelle M, Wallace E,Zoghbi SS, et al. Age-related decline in striatal dopamine transporterbinding with iodine-123-beta-CITSPECT [see comments]. J NuclMed 1995; 36: 1175–81. Comment in: J Nucl Med 1996; 37: 1911–2.



ownloaded from



Volkow ND, Wang GJ, Fowler JS, Logan J, Gatley SJ, MacGregorRR, et al. Measuring age-related changes in dopamine D2 receptorswith 11C-raclopride and 18F-N-methylspiroperidol. Psychiatry Res1996a; 67: 11–16.

Volkow ND, Ding YS, Fowler JS, Wang GJ, Logan J, Gatley SJ,et al. Dopamine transporters decrease with age. J Nucl Med 1996b;37: 554–9.

Waldemar G, Hasselbalch SG, Andersen AR, Delecluse F, PetersenP, Johnsen A, et al. 99mTc-d,l-HMPAO and SPECT of the brain innormal aging. J Cereb Blood Flow Metab 1991; 11: 508–21.

Whelihan WM, Lesher EL. Neuropsychological changes in frontalfunction with aging. Dev Neuropsychol 1985; 1: 371–80.

Wickelgren I. For the cortex, neuron loss may be less than thought[new] [see comments]. Science 1996; 273: 48–50. Comment in:Science 1996; 274: 484–5.

Williams EJ. The comparison of regression variables. J R StatistSoc Ser B 1959; 21: 396–9.

Williams GV, Goldman-Rakic PS. Modulation of memory fields bydopamine D1 receptors in prefrontal cortex [see comments]. Nature1995; 376: 572–5. Comment in: Nature 1995; 376: 549–50.

Wong DF, Young D, Wilson PD, Meltzer CC, Gjedde A.Quantification of neuroreceptors in the living human brain: III. D2-like dopamine receptors: theory, validation, and changes duringnormal aging. J Cereb Blood Flow Metab 1997; 17: 316–30.

Woods RP, Cherry SR, Mazziotta JC. Rapid automated algorithmfor aligning and reslicing PET images. J Comput Assist Tomogr1992; 16: 620–33.

Yoshii F, Barker WW, Chang JY, Loewenstein D, Apicella A,Smith D, et al. Sensitivity of cerebral glucose metabolism to age,gender, brain volume, brain atrophy, and cerebrovascular risk factors.J Cereb Blood Flow Metab 1988; 8: 654–61.

Received November 30, 1998. Accepted January 12, 1999



ownloaded from


Context-dependent, neural system-specific neurophysiological concomitants of ageing: mapping PET...

Documents

Transcript of Context-dependent, neural system-specific neurophysiological concomitants of ageing: mapping PET...