Neuropsychologia 46 (2008) 2109–2121

Contents lists available at ScienceDirect

Neuropsychologia

journa l homepage: www.e lsev ier .com/ locate /neuropsychologia

Hippocampal activation during episodic and semantic memory retrieval:Comparing category production and category cued recall

Lee Ryana,∗, Christine Coxa, Scott M. Hayesb, Lynn Nadela

a Cognition & Neuroimaging Laboratories, Department of Psychology, University of Arizona, United Statesb Center for Cognitive Neuroscience, Duke University, United States

campowevd epormay taskpal acclearthe sing thwe counted

of thr aut

egoryinctio

a r t i c l e i n f o

Article history:Received 3 October 2007Received in revised form 25 February 2008Accepted 25 February 2008Available online 18 March 2008

Keywords:HippocampusSemantic memoryCategory productionEpisodic memoryfMRIVerbal fluency

a b s t r a c t

Whether or not the hippodebate in the literature. Hvation during semantic ansource of the retrieved infa classic semantic memorcued recall. Left hippocamother regions of the brainretrieval strategies duringtion; for example, visualizutensils. In Experiment 2,strategies could have accopresented that elicited oneand nonspatial, and neitheobserved for all three catWe conclude that the dist

memory models suggest.

1. Introduction

It is clear that the hippocampus plays an essential role inthe formation and retrieval of episodic memories. In fact, recentfMRI evidence suggests that the hippocampus participates in thesuccessful retrieval of episodic information and autobiographi-cal events, especially detailed contextual information, even forevents that occurred over 20 years ago (Ryan et al., 2001; forreview, see Nadel, Campbell, & Ryan, 2007; Nadel, Winocur, Ryan, &Moscovitch, 2007). In contrast, debate continues regarding whetheror not the hippocampus is critical for the retrieval of semanticmemories, including personal semantics and world knowledge.Much of the evidence on both sides of this debate comes frompatients with medial temporal lobe (MTL) damage. In a recentreview, Moscovitch, Nadel, Winocur, Gilboa, and Rosenbaum (2006)concluded that retrograde amnesia for semantic memory is either

∗ Corresponding author at: Department of Psychology, University of Arizona, P.O.Box 210068, Tucson, AZ 85721-0068, United States. Tel.: +1 520 621 7443;fax: +1 520 621 9306.

E-mail address: ryant@email.arizona.edu (L. Ryan).

0028-3932/$ – see front matter © 2008 Elsevier Ltd. All rights reserved.doi:10.1016/j.neuropsychologia.2008.02.030

us participates in semantic memory retrieval has been the focus of mucher, few neuroimaging studies have directly compared hippocampal acti-

isodic retrieval tasks that are well matched in all respects other than thetion. In Experiment 1, we compared hippocampal fMRI activation during, category production, and an episodic version of the same task, categorytivation was observed in both episodic and semantic conditions, althoughly distinguished the two tasks. Interestingly, participants reported usingemantic retrieval task that relied on autobiographical and spatial informa-emselves in their kitchen while producing items for the category kitchennsidered whether the use of these spatial and autobiographical retrievalfor the hippocampal activation observed in Experiment 1. Categories wereree retrieval strategy types, autobiographical and spatial, autobiographicalobiographical nor spatial. Once again, similar hippocampal activation wastypes, regardless of the inclusion of spatial or autobiographical content.n between semantic and episodic memory is more complex than classic

© 2008 Elsevier Ltd. All rights reserved.

spared completely or confined to a period of about 10 years prior

to the head injury, providing that the damage is limited primar-ily to the hippocampal formation. In contrast, Squire and others(Luo & Niki, 2002; Manns, Hopkins, & Squire, 2003; Squire & Zola,1998; Squire, Stark, & Clark, 2004) emphasize that at least someamnesics appear to have significant deficits in semantic mem-ory retrieval, even for well-established world knowledge. Semanticmemory impairment tends to be more extensive if the damageincludes other medial temporal lobe and neocortical structures andcan reach the same level of deficit as autobiographical memoryloss, or even exceed it, in some patients (Bayley, Hopkins, & Squire,2003; Bayley and Squire, 2005; but see Maguire, Frith, Rudge, andCipolotti, 2005).

Neuroimaging evidence is even less consistent than lesionstudies regarding hippocampal involvement in episodic, but notsemantic, retrieval. A growing number of studies have reported MTLactivity during tasks that require access to semantic knowledge,including hippocampal activation during retrieval of public events(Maguire, Vargha-Khadem, and Mishkin (2001)) and famous faces(Kapur, Friston, Young, Frith, & Frackowiak, 1995; Leveroni et al.,2000; Bernard et al., 2004), and parahippocampal gyrus activationfor famous faces (Haist, Bowden Gore, & Mao, 2001) and famous

cholog

2110 L. Ryan et al. / Neuropsy

names (Douville et al., 2005). A handful of neuroimaging studiesfocusing on semantic spatial knowledge have also found activationin hippocampal and adjacent MTL structures. For example, Maguire,Frackowiak, and Frith (1997) reported activation in parahippocam-pal gyrus when experienced London taxi-drivers were required tofind novel routes from one location to another when familiar routeswere blocked. Few neuroimaging studies, however, have made adirect comparison between episodic and semantic retrieval tasksthat are well matched in other respects, including the type of stim-uli presented, the familiarity of the stimuli, and the responses madeby the participant, while varying only the source of the retrievedinformation. In one such study, Maguire and Mummery (1999)compared yes/no recognition for autobiographical events and pub-lic events and found hippocampal activation during both semanticand episodic retrieval, although the level of activation was greaterfor episodic events. Duzel et al. (1999) also matched conditionscarefully in a 2 × 2 design crossing semantic living/nonliving judg-ments with recognition old/new judgments. In contrast to Maguireand Mummery’s (1999) results, however, they found activation inmedial temporal lobe only when comparing recognition judgments(old > new) but not semantic judgments.

The discussion of whether or not the hippocampus is involvedpreferentially in retrieval of semantic and episodic memoryassumes that these two types of memory are independent ofone another, and this may not be the case. Tulving’s recent work(Tulving, 2005) has outlined the many similarities across these twotypes of memories. Barsalou and coworkers (Barsalou, 1983, 1988;Barsalou & Sewell, 1985) have suggested that semantic memoryis contextually bound to autobiographical information, such thatepisodic memory is frequently used to generate semantic informa-tion. For example, Vallee-Tourangeau, Anthony, and Austin (1998)asked participants to generate category exemplars to such commoncategories as kitchen utensils or food items, and then asked them todescribe the strategies they used to generate the items. In over 70%of the common categories, participants reported using a strategyinvolving a personally familiar context, such as imagining their ownkitchen, or walking through the aisles of their neighborhood gro-cery story (see also Walker & Kintsch, 1985; Williams & Hollan,1981). Category production tasks are used widely in neuropsy-chological evaluations to assess the integrity of semantic memory(Andrewes, 2001; Lezak, Howieson, & Loring, 2004). If episodicmemory is preferentially used by neurologically normal individ-uals to generate semantic information in such classic semantictasks as category production, then patients with hippocampal dam-

age should show impairment on this task. Contrary to this notion,patients with MTL amnesia are most often reported as normal oncategory production tasks (Schmolck, Kensinger, Corkin, & Squire,2002). However, one study (Gleissner & Elger, 2001) has reportedthat patients with lesions restricted to the hippocampal complexgenerate fewer exemplars during category production than nor-mal individuals or patients with non-hippocampal temporal lobelesions. In addition, Pihlajamaki et al. (2000) reported fMRI activa-tion in the left hippocampus and adjacent parahippocampal gyruswhen normal young subjects generated typical category items.

Clearly, sufficient ambiguity exists in the literature to warrantfurther investigation of whether or not the hippocampus is engagedduring semantic retrieval, particularly given the critical importanceof this issue for theories of memory consolidation (for review anddiscussion, see Moscovitch et al., 2006). In the present study, weused a typical category production task where participants gener-ated exemplars to category names such as modes of transportationor kitchen utensils. In Experiment 1, we compared episodic andsemantic versions of the task that differed only in the source ofthe exemplars that were generated—semantic memory (categoricalworld knowledge) or episodic memory (a single learning episode

ia 46 (2008) 2109–2121

that occurred 24 h earlier). Note that we define episodic memoryhere narrowly, referring only to the ability to recall informationfrom a prior event that is bounded by a single spatial–temporalcontext, and not in the broader sense of requiring an assessmentof autonoetic consciousness, or the sense of recollecting, as Tulving(2005) has recently described.

In a second experiment, we compared retrieval within seman-tic memory for categories that were designed to elicit differenttypes of retrieval strategies. Previous research (Barsalou, 1988;Vallee-Tourangeau et al., 1998) has shown that some, but not all,categories elicit the use of personally familiar spatial contexts (suchas their garage or a kitchen) when participants are asked to gener-ate category items. Other categories (such as things that are red orexpensive things) do not elicit these contextual strategies. We exam-ined whether or not the presence of spatial and autobiographicalstrategies during category production would determine whetheror not activation is observed in the hippocampus and other MTLregions.

2. Experiment 1

2.1. Methods

2.1.1. ParticipantsParticipants were healthy University of Arizona undergraduate and graduate

students who gave written informed consent and received course credit as com-pensation. All experimental procedures were approved by the University of Arizonainstitutional review board. Exp. 1 included 10 participants (5 males, 5 females; meanage 24.5, range 19–36 years; mean years of education 14.2, S.D. 3). All participantswere screened for prior significant head injury, drug or alcohol abuse, psychiatricdisorder, and contraindications to MRI.

2.1.2. MaterialsThirty categories were selected from the Battig and Montague (1969) norms and

a more recent category normative dataset (Van Overschelde, Rawson, & Dunlosky,2004). Common categories with at least 20 exemplars, such as animals, furniture,and kitchen utensils were chosen. For the episodic memory task, seven relativelytypical exemplars were chosen from each category excluding the one or two mostcommon responses. The 30 categories were randomly split into two lists, with eachlist presented equally often in the episodic and semantic retrieval conditions.

2.1.3. ProcedureThe experiment included a study phase and a retrieval phase. The study phase

occurred 24 h prior to the fMRI scanning session. Participants came to the laboratoryand were required to learn 7 exemplars from each of 15 categories to a criterion of 10perfect repetitions. We set a strict learning criterion in order to ensure that partici-pants could recall the items without difficulty 24 h later. The study phase proceededas follows. The experimenter read the first category name aloud followed by the

7 exemplars at 2 s intervals. Participants were instructed to repeat the exemplarsaloud, in any order. The experimenter read each subsequent category and exemplarlist, and the participant repeated them in any order. No feedback was given aftereach list. After one complete presentation of the 15 categories, the learning trialswere repeated, with two exceptions. First, on each repetition, the categories werepresented in a new random order, although the exemplars were always presentedin the same order for all learning trials. Second, when the category was named bythe experimenter, participants were instructed to recall the list of seven exemplarsin any order. If recall was not perfect, the list of exemplars was read again by theexperimenter, and the participant repeated the exemplars aloud. The learning pro-cess continued until participants correctly recalled all the exemplars from the 15categories perfectly, 10 consecutive times. The study session took approximately 1 hto complete.

Twenty-four hours later, participants underwent fMRI scanning. While in thescanner, participants were shown 30 category names, presented in random orderusing high resolution goggles (Resonance Technologies, Inc.). Each category was pre-ceded by a 1.5-s cue “OLD” or “NEW”, which was then replaced with a category namethat remained on the screen for 12 s. OLD categories were category names for whichparticipants had learned a list of exemplars 24 h earlier. For OLD categories, partici-pants were instructed to recall the seven exemplars they had learned the previousday (referred to as the Recall condition). NEW categories were category names thatparticipants had not been exposed to during the study session the previous day.For NEW categories, participants were instructed to generate the first seven itemsbelonging to the category that came to mind (referred to as the Generate condi-tion). In each case, they pressed a mouse button whenever they recalled/generateda word in the category. Because we did not obtain verbal responses in the scan-

hologia 46 (2008) 2109–2121 2111

L. Ryan et al. / Neuropsyc

ner, the button presses provided some indication that responses for OLD and NEWcategories were similar in terms of number of responses and response times. Partic-ipants were given 12 s to recall/generate items from each category. The 12 s responsewindow was chosen because pilot data indicated this provided sufficient time forparticipants to recall the seven items from the memorized list, and also to generate asimilar number of exemplars for non-studied categories. A visual-motor control con-dition was also included 15 times, randomly interspersed between categories. Forthe control condition, the letter “X” was presented seven times within the 12 s win-dow. The timing of the X’s was randomly jittered across the 12 s, so that each X waspresented for 800–1200 ms, with interstimulus intervals of 400–600 ms. This pat-tern better approximated the button presses made by participants during categoryproduction and cued recall conditions. Participants pressed the mouse button eachtime another X appeared on the screen. Finally, 15 rest periods were interspersedrandomly between category and control trials, each lasting 16 s during which theword “REST” appeared on the computer screen.

Immediately following the fMRI session, participants were once again asked

to recall the lists of category exemplars to ensure that they had recalled the correctitems in the scanner. For each category, they were also asked to describe any specificthoughts or strategies that they were aware of as they generated and recalled items.

2.1.4. Image acquisitionImages were acquired on a 1.5-T GE full-body echo speed magnet. A set of 3-plane

localizer images were first collected in order to align the functional images. Func-tional data were acquired using a single-shot spiral pulse sequence (Glover & Law,2001; Glover & Lee, 1995), matrix 64 × 64, TR = 2000 ms, TE = 40 ms, flip angle = 90,19 sections, 6 mm, no skip. Sections were placed obliquely, perpendicular to the longaxis of the hippocampus in order to minimize partial voluming of the hippocam-pus, and covered the anterior two-thirds of the brain including the posterior extentof the hippocampus (see Fig. 1 showing the placement of the sections). Becausethe focus of the study was on MTL functioning, we chose to maximize fully vol-umed data obtained from this region while keeping a reasonably short TR, but at acost of covering posterior regions including the precuneus, some occipital cortex,and superior parietal cortex. A whole brain high resolution SPGR 3D volume wasalso obtained for co-registration of the functional dataset, matrix = 256 × 256, flipangle = 30, TR = 22 ms, TE = 8 ms, FOV = 25 cm, 1.5 mm sections, no skip.

2.1.5. Image analysisData were analyzed as a blocked design using SPM99 (Wellcome Dept. of Cog-

nitive Neurology; http://www.fil.ion.ucl.ac.uk/spm). ROI analyses were completedusing MarsBaR (Brett, Anton, Valabregue, & Poline, 2002) and mean effect sizes wereoutput for further analysis to SPSS, version 14. Images were reconstructed offlineand then realigned to the third volume for motion correction. Spatial normalization

Table 1Regions of activation in the medial temporal lobe for Experiment 1 (cluster p < .0001) and

L/R MNI

x

Experiment 1Hippocampus

Generatea L −23Recalla L −21

Parahippocampal gyrusGeneratea L −19

R 21Recalla L −20

R 24

Experiment 2Hippocampus

Autobiographical Nonspatiala L −23R –

Autobiographical Spatiala L −23R 24

Nonautobiographical Nonspatiala L −25R 25

Parahippocampal gyrusAutobiographical Nonspatiala L –

R 19Autobiographical Spatiala L −24

R 21Nonautobiographical Nonspatiala L −18

R 20

Laterality (L = left, R = right), MNI coordinates, extent of activation (in voxels; k), and peakin Experiment 1 and for each retrieval strategy in Experiment 2.

a Task.

Fig. 1. Slice selection for Experiment 1. Note that coverage did not extend to poste-rior parietal and occipital regions.

parameters were estimated by warping each participant’s mean functional imageto the standard MNI (Montreal Neurological Institute) EPI template (Ashburner &Friston, 1999). The normalized images were resliced to 3 mm × 3 mm × 3 mm voxelsand smoothed with an isotropic 8 mm FWHM Gaussian kernel. The time series ineach voxel was highpass-filtered to 1/128 Hz and scaled to a grand mean of 100,averaged over all voxels and scans within the session.

Statistical analyses were performed in two stages. In the first stage, neural activ-ity was modeled by a delta function at stimulus onset. The ensuing BOLD responsewas modeled by convolving these delta functions with a canonical hemodynamicresponse function (HRF; Friston et al., 1995). The resulting time courses were downsampled to form covariates in a General Linear Model. Covariates were modeled forthe canonical HRFs of the episodic and semantic retrieval conditions described ear-lier, the control condition, and a single covariate representing the mean (constant)over scans (only canonical HRF covariates were used to make contrasts and move tosecond level analyses). Contrasts of parameter estimates comprised the data for the

Experiment 2 (cluster p < .005)

k T

y z

−19 −16 163 7.43−19 −15 53 4.11

−26 −17 210 8.07−18 −19 51 6.29−25 −19 146 5.95−24 −18 40 3.82

−29 −8 32 2.86– – – –

−30 −7 101 3.90−26 −10 45 3.05−30 −7 73 3.29−26 −10 37 2.99

– – – –−29 −13 21 3.07−35 −12 99 4.04−32 −13 84 4.28−36 −9 18 3.14−31 −13 78 3.66

voxel t-statistics are provided for each region of interest for the two retrieval tasks

cholog

rate coe) anaiguou

glia, and bilateral thalamus.

2.3. Discussion

The results provide evidence that the left hippocampus andbilateral parahippocampal gyrus are activated during the retrievalof both episodic and semantic information. Importantly, these taskswere similar in terms of the cues presented and the responsesmade by the participants. Category production is considered aclassic tool for evaluating semantic memory. When patients areimpaired on this task, as is often the case for individuals with earlyAlzheimer’s disease, the result is taken as an indication of a break-down in the semantic knowledge network (Chan, Salmon, Butters,& Johnson, 1995; Monsch et al., 1994; Welsh, Butters, Hughes, Mohs,& Heyman, 1992). The simplest interpretation of the results wouldbe that the hippocampus is equally involved in the retrieval of bothepisodic and semantic information. However, given that this findingis inconsistent with the literature on MTL amnesia, we consid-ered an alternative hypothesis focusing on the retrieval strategiesadopted by participants during the Generate condition. Following

2112 L. Ryan et al. / Neuropsy

Fig. 2. Left (L) and right (R) medial temporal lobe activation for the Recall and Genefrom activation crossed with hippocampus (red) and parahippocampal gyrus (bluactivation in the right hippocampus did not survive our extent threshold of 10 cont

second-stage analyses, which treated participants as a random effect. Using meth-ods described by Forman et al. (1995), the statistical threshold was determined byconsidering both voxel-wise alpha and the extent of contiguous activation. A clus-ter of 10 contiguous voxels thresholded at p < .01 was required in order to obtaina cluster-level criterion of p < .0001. Stereotactic coordinates were generated in thestandard Montreal Neurological Institute (MNI) brain by SPM, and are reported herein MNI space. Mean effect sizes for regions of interest and anatomical boundariesfor MTL regions (hippocampus and parahippocampal gyrus) were extracted usinganatomical masks from MarsBaR (Brett et al., 2002).

2.2. Results

2.2.1. Behavioral resultsParticipants were able to recall all the category items perfectly.

The response time data suggest that the two conditions were wellmatched in terms of number of responses and timing of responses.Within the 12 s window, they generated an average of 8.0 (S.D. .3)items in the Generate condition compared to seven items in theRecall condition, but this difference was not statistically signifi-cant. The reaction times were also similar across the two conditionswith responses ranging on average between 1606 (S.D. 78) and8900 (S.D. 108) ms for the semantic condition, and between 1733(S.D. 61) and 8216 (S.D. 94) ms for the episodic condition (allpairwise t’s < 1).

2.2.2. Imaging results

Table 1 shows the MNI coordinates and cluster extents for

regions of activation within MTL for the Recall and Generate con-ditions. Fig. 2 shows the strikingly similar placement and extentof activation within MTL for the two conditions. Both conditionselicited activation in left hippocampus and bilateral parahippocam-pal gyrus.

In order to further contrast the activation between conditions,we compared mean effect sizes for Recall and Generate conditionsin MTL regions that overlapped across conditions. We crossed thecombined activation maps for the two conditions with anatomicalregion of interest maps from MarsBaR (Brett et al., 2002), obtainingmean effect sizes for each participant for the following regions: lefthippocampus and left and right parahippocampal gyrus. The meansfor each region were compared with paired t-tests and showedsimilar mean effect sizes across conditions (all p’s > .05, see Fig. 3).

Other brain regions showed some similarity of activation acrossthe two retrieval conditions, but were differentiated by severalnotable exceptions (see Table 2). Recall elicited activation in leftposterior caudate nucleus and the right reticular formation of themidbrain, while Generate showed activation in bilateral inferiorfrontal gyrus, a more posterior and lateral region of dorsolateral

ia 46 (2008) 2109–2121

nditions contrasted with the control task in Experiment 1, cluster p < .0001. Resultstomical regions of interest are presented for each condition separately. Note thats voxels.

prefrontal cortex, more extensive activation in bilateral basal gan-

Fig. 3. Mean effect sizes for the Recall (Rec) and Generate (Gen) conditions inExperiment 1 contrasted with control within medial temporal lobe (MTL) combinedactivation regions of interest (ROI), cluster p < .0001. ROIs included all active voxelswithin left hippocampus and left and right parahippocampal gyrus for both the Recand Gen conditions. There were no significant differences in mean magnitude ofactivation between conditions in these regions.

L. Ryan et al. / Neuropsychologia 46 (2008) 2109–2121 2113

Table 2Brain regions outside the medial temporal lobe activated by Recall and Generate conditions in Experiment 1 (cluster p < .0001)

Condition L/R BA MNI k T

x y z

Recall and generateMedial prefrontal cortexa B 32 −4 34 22

Recallb 1665 7.40Generateb 2134 8.82

Ventrolateral prefrontal cortexa L 47 −29 30 −10Recallb 62 7.92Generateb 27 3.91

Dorsolateral prefrontal cortexa L 6/8 −26 16 56Recallb 340 6.11Generateb 56 3.79

Basal gangliaa L – −16 −1 9Recallb 95 3.68Generateb 415 4.15

Midbraina L – −6 −22 −20Recallb 208 9.17Generateb 117 5.39

Posterior cingulatea L 29 −4 −50 3Recallb 44 6.38Generateb 230 8.86

R 29 7 −47 3Recallb 38 7.61

−

21

−7

33−14−

nt of arior c

rus ex; midbamus

Generateb

Fusiform gyrusa L 19Recallb

Generateb

R 19Recallb

Generateb

Recall onlyPosterior caudatea L –Midbraina R –

Generate onlyVentrolateral prefrontal cortexa R 47Dorsolateral prefrontal cortexa L 8Basal gangliaa R –Thalamusa B –

Laterality (L = left, R = right, B = bilateral), Brodmann Area (BA), MNI coordinates, extewell as regions specific to each condition. Medial prefrontal cortex includes the anteinferior frontal gyrus; dorsolateral prefrontal cortex includes the superior frontal gyand globus pallidus; fusiform gyrus extents into lingual gyrus and occipital cortexcerebral and cerebellar peduncles; and posterior caudate extends into superior thal

a Region.b Task.

the fMRI session, we asked participants to describe what they were

thinking about as they generated and recalled items from the var-ious categories. Consistent with Vallee-Tourangeau et al. (1998),during the Generate condition, participants often reported thatthey retrieved autobiographical and spatial contextual informationthat helped them to generate the category exemplars. For example,for the category kitchen utensils, all of the participants describedpicturing themselves in their own kitchen, looking across the coun-tertops and even opening drawers, as they named items they “saw”in their kitchen. For the category family members, the majorityof participants visualized the faces of their family members asthey generated exemplars. In contrast, during the Recall condi-tion, participants did not describe such autobiographical and spatialstrategies. Instead, they either reported that they were unaware ofany particular strategy, or in some cases, they described the seman-tic organization of the list. For example, one participant said thatthey recalled the list of animals ordered from smallest to largest.Few participants reported that the episodic lists were organizedinto a cohesive spatial context.

Although the debriefing information was not collected or codedsystematically, these self-reports raised an interesting hypothesis,namely, that the hippocampal activation during the Generate con-

62 5.0530 −73 −21

34 4.3680 5.12

−80 −1836 4.1422 3.47

10 −5 18 16 3.91−20 −23 45 5.50

26 1 21 4.6045 8 47 63 4.98

7 1 78 5.375 −10 3 90 4.29

ctivation (in voxels; k), and peak voxel t-statistics are given for regions of overlap asingulate extending to medial frontal gyrus; ventrolateral prefrontal cortex includestending to middle frontal gyrus; basal ganglia includes regions of both the putamenrain refers to the dorsal brainstem extending from the reticular formation to the

.

dition may have been driven by the inclusion of autobiographical

and spatial contextual information into the retrieval process forcategory exemplars. To test this hypothesis, we developed cate-gories for an additional study that would reliably elicit differenttypes of strategies. We began with a normative study to identifycategories that would or would not elicit autobiographically rele-vant information and spatial contextual information. Ideally, thiswould provide four types of category strategies: autobiographi-cal with spatial context (visualizing my kitchen), autobiographicalwithout spatial context (the faces of my family members), non-autobiographical with spatial context (visualizing a jail cell or thespace shuttle), and non-autobiographical without spatial context(either no discernable strategy or relying on semantic strategiesalone such as largest to smallest).

We piloted over 60 categories based on work by Barsalou andcoworkers (Barsalou, 1983, 1988; Vallee-Tourangeau et al., 1998).Forty undergraduates were asked to produce the first seven itemsthat came to mind for each category, and then subsequently wereasked to describe what they were thinking about as they producedthe items. Our goal was to obtain categories for which 80% of par-ticipants reported using a similar type of strategy. Three of the fourcategory retrieval strategies were clearly distinguishable from one

cholog

2114 L. Ryan et al. / Neuropsy

another using this criterion, including autobiographical with spatialcontext (AS; things found in a garage, or kitchen utensils), autobi-ographical without spatial context (AN; things worn on the feet,high school classes), and neither autobiographical nor spatial con-text (NN; precious stones, things that are red). However, we couldnot elicit categories where participants used strategies that wereclearly not autobiographical but contained a single spatial context.For example, such categories as things found in a jail cell would pre-sumably draw on knowledge of a context that most undergraduateshad not yet experienced personally. While participants regularlydescribed visualization of a spatial context, they often describedquasi-autobiographical experiences such as imagining a jail cell ina movie that they had watched recently. In order to keep the strat-egy types clearly separated from one another, we compared onlythree types, AS, AN, and NN. In order to further ensure that the ASand AN categories drew on autobiographical strategies to differ-

entiate them from the NN condition, the category names includedthe pronouns “your” and “you” where appropriate, such as thingsin your garage, or things worn on your feet.

If the use of a spatial context is driving the hippocampal activa-tion observed in Experiment 1, we would expect to see activationonly for the AS condition, but not AN or NN categories. Alter-natively, any autobiographical content, spatial or otherwise, mayresult in hippocampal activation. By this view, hippocampal acti-vation should be observed for AS and AN, but not NN categories.Finally, if the hippocampus is important for any semantic retrievaltask, then activation in MTL should be evident for all three categoryconditions, AS, AN, and NN.

3. Experiment 2

3.1. Methods

3.1.1. ParticipantsParticipants included University of Arizona undergraduate and graduate

students who gave written informed consent and received course credit as com-pensation. Exp. 2 included 14 participants (5 males, 9 females; mean age 24.8, range18–40 years; mean years of education 15.7, S.D. 2.9). All participants were screened

Fig. 4. Left (L) and right (R) medial temporal lobe activation for each retrieval strategy cNonspatial (AN), Autobiographical Spatial (AS), and Nonautobiographical, Nonspatial (NNgyrus (blue) anatomical regions of interest are presented for each condition separately. Nour extent threshold of 12 contiguous voxels.

ia 46 (2008) 2109–2121

for prior significant head injury, drug or alcohol abuse, psychiatric disorder, andcontraindications to MRI.

3.1.2. MaterialsA total of 39 categories obtained from the pilot study described earlier were

included in the study. These categories were likely to draw on the three types ofstrategies described earlier, autobiographical with spatial context (AS), autobio-graphical without spatial context (AN), and neither autobiographical nor spatial(NN). Thirteen categories were included in each condition.

3.1.3. ProcedurePrior to scanning, participants were given several practice categories and were

instructed simply to generate as many items belonging to the category as they couldthink of, and to press the mouse button in their right hand each time they thoughtof a new category item. While in the scanner, participants were presented visuallywith a category name and were given 12 s to generate items from each category. All39 categories were presented in random order during a single fMRI scan. Thirteenblocks of a visual-motor control condition, described in Experiment 1 (seven X’s

jittered within a 12-s window), were also randomly inserted between categories.Each category and control block was separated by a 4-s rest period in which theword “REST” appeared.

After the functional scan was completed, participants were debriefed in orderto obtain information about the strategies that they used to generate items for eachcategory. Although the categories were developed to elicit particular strategy types,individual differences did occur, and we therefore identified the specific strategytype used for each category based on the participants’ own responses, rather thanrelying on the normative data. In addition, occasionally a category was excludedwhen a participant could not describe an unambiguous retrieval strategy. Thus, foreach participant, the number of categories included in the AN, AS, and NN conditionsvaried. However, no participant had less than eight categories in each condition. Themean numbers of categories across participants included in each condition wereas follows: AN mean = 10.5 (S.D. 2.1), AS mean = 14.3 (S.D. 2.4), and NN mean = 13(S.D. 2.9). When compared using paired t-tests, significantly fewer categories wereincluded in the AN condition compared to either AS or NN conditions (p’s < .05).

3.1.4. Image acquisition and analysisImages were acquired using the same acquisition parameters as Experiment

1, with the exception that images were collected horizontally aligned to theanterior–posterior commissural plane, covering the whole brain with 27 sections,3.5 mm no skip. Image preprocessing and data analyses were carried out with thesame methods as described in Experiment 1. Whole brain analyses were carriedout at a cluster threshold of p < .0001 which was determined following methodsdeveloped by Forman et al. (1995) using a voxel-wise p < .01 and an extent of 10 ormore contiguous voxels. However, because our hypotheses were based on finding

ontrasted with the control task in Experiment 2, cluster p < .005; Autobiographical). Results from activation crossed with hippocampus (red) and parahippocampal

ote that for the AN condition, activation in the right hippocampus did not survive

holog

L. Ryan et al. / Neuropsyc

a lack of activation in one or more category conditions and because we includedfewer categories in each condition, we employed a more liberal statistical thresholdwhen assessing effect sizes within the medial temporal lobe regions of interest. Athreshold of p < .05 with an extent of 12 or more contiguous voxels provided a clusteralpha of p < .005, consistent with previous imaging studies on medial temporal loberegions that have also used a less stringent statistical criterion (for example, Addis,Wong, & Schacter, 2007; Dobbins, Rice, Wagner, & Schacter, 2003).

3.2. Results

3.2.1. Medial temporal lobe regionsTable 1 shows the MNI coordinates and cluster extents for

regions of activation within the MTL for the three strategy condi-tions. Fig. 4 shows the placement and extent of activation for eachof the three strategy conditions compared to the control conditionwithin the MTL. The distribution of activation appears similar forthe three conditions and consistent with the pattern of activationobserved in Experiment 1 during category production. Activationwas observed in both hippocampus and parahippocampal gyrusfor all three category conditions, even in the NN category conditionwhere participants reported that they are unaware of any strategybeyond the semantic characteristics of the items.

In order to make direct comparisons between the three categoryconditions within the MTL, activation maps combining activationacross all three conditions were crossed with anatomical ROI masksobtained from MarsBaR, providing the mean effect size for eachparticipant from left and right hippocampus, left and right pos-terior parahippocampal gyrus extending posteriorly to fusiformgyrus and lingual gyrus, and an additional cluster in anterior rightparahippocampal gyrus (see Fig. 5 for ROIs). Effect sizes were com-pared across the three strategy conditions in each region usinga one-way ANOVA with followed paired t-tests. The results aredepicted in Fig. 6. First, significant activation was obtained for allthree conditions in right and left hippocampus (but activation inright hippocampus for the AN condition did not survive our extentthreshold of 12 contiguous voxels), and the mean effect size in theseregions did not differ across the category types, F’s < 1. This was alsotrue in the anterior right parahippocampal gyrus, F < 1. In the moreposterior regions incorporating posterior parahippocampal gyrusand fusiform gyrus, significantly greater activation was observedbilaterally for the AS categories compared to either of the two otherstrategy conditions, AN or NN, as indicated by a main effect of cat-egory type, F(2, 81) = 4.99, p < 01. Follow-up paired t-tests betweenconditions are included in Fig. 6.

3.2.2. Other regions of activationThe similarity of activation observed in MTL regions also

extended to most, but not all, other brain regions. Table 3Ahighlights other brain regions showing common clusters ofactivation across all three strategy conditions, including bilat-eral dorsomedial prefrontal cortex, left dorsolateral prefrontalcortex, left ventrolateral prefrontal cortex, left thalamus,bilateral retrosplenial/posterior cingulate cortex, and bilateralfusiform/visual cortex extending from the lingual gyrus tofusiform gyrus and cuneus. Voxels of maximal activation withineach overlapping cluster were extracted in SPM99.

In addition to regions of overlap, there were notable regionsof activation which distinguished the three strategy conditions.Table 3B highlights regions that differentiated the strategy con-ditions, showing specific activation for either one or two of thecategory types, but not all three. These regions were compareddirectly using mean effect sizes. As discussed previously, the AScondition, in which participants reported visualizing spatial con-texts, was associated with greater activation in bilateral posteriorparahippocampal gyrus extending into the fusiform gyrus and lin-gual gyrus. AS also resulted in specific activation in lateral and

ia 46 (2008) 2109–2121 2115

superior parietal cortex and superior precuneus compared to ANand NN (see Figs. 7 and 8). The AN condition elicited activationin this same region of superior precuneus as well, but did notexceed our extent threshold of 10 contiguous voxels. Interestingly,a large region of activation in the ventromedial prefrontal cortexwas uniquely observed in the AN condition, with greater activationfor AN compared to either AS or NN. The NN condition showed oneregion within the midbrain, in the dorsal brainstem, that elicitedgreater activation compared to AS and AN.

3.3. Discussion

In summary, the results from Experiment 2 showed a simi-lar pattern of activation for semantic retrieval as observed duringcategory production in Experiment 1. Activation was again evi-dent within the medial temporal lobes as participants generatedexemplars for categories. Whether or not the three categorytypes elicited differential activation depended upon the specificregion within MTL. Within the hippocampus proper, activationdid not differ as a function of whether participants used auto-biographical or spatial contextual information as a strategy forretrieval. The posterior parahippocampal gyrus, however, showedsignificantly greater activation for categories that elicited an auto-biographically relevant spatial context, such as the participant’skitchen, garage, or local supermarket. AS categories also acti-vated fusiform and lingual gyrus, along with superior parietalcortex. Collectively, these posterior regions have been associatedwith the processing of complex scenes (Burgess, Maguire, Spiers,& O’Keefe, 2001; Hayes, Ryan, Schnyer, & Nadel, 2004), episodicmemory retrieval (Gilboa, Winocur, Grady, Hevenor, & Moscovitch,2004; Ryan et al., 2001; Shannon & Buckner, 2004), spatial naviga-tion tasks (Maguire, Frackowiak, & Frith), and retrieval of remotespatial information in amnesics (Rosenbaum et al., 2004), andare consistent with the self-reports of participants of visualiz-ing and sometimes even navigating through personally familiarspaces.

Activation in the superior precuneus was present for both con-ditions that elicited autobiographical content, AS and AN (but didnot meet extent thresholds in the AN condition). This region alsohas been associated with episodic memory retrieval (Gilboa etal., 2004; Ryan et al., 2001; Shannon & Buckner, 2004), and hasbeen implicated in tasks that relate to judgments regarding the self(Cavanna & Trimble, 2006; Northoff et al., 2006). Ventromedial pre-frontal cortex, observed only during retrieval of the AN categories,

is also a region that has been associated with self-referential tasks(Johnson et al., 2002; Northoff et al., 2006; Saxe, Moran, & Scholz,2006), consistent with the notion that participants generated auto-biographically relevant information in order to generate items forcategories such as family members.

Of particular interest is the pattern of activation associatedwith NN categories, where participants are unaware of any spe-cific autobiographical or spatial strategy used to generate theitems. Of the three category types, NN categories should rep-resent the “purest” semantic memory task, uncontaminated byepisodic or autobiographical content. NN categories were mostclearly differentiated from AS and AN in posterior cortical regionsincluding the precuneus, posterior parietal, and fusiform gyrus,and anteriorly in ventromedial prefrontal cortex. Nevertheless,NN categories shared many common regions with both AS andAN categories, including the hippocampus, dorsomedial prefrontalcortex, and retrosplenial cortex, all regions that are also com-monly activated during episodic memory retrieval tasks. It appearsthat even for these categories, exemplar retrieval is mediated,at least in part, by regions in common with episodic memoryretrieval.

2116 L. Ryan et al. / Neuropsychologia 46 (2008) 2109–2121

from a

Fig. 5. Combined activation region of interest (ROI) masks for Experiment 2, created cluster p < .005, which were then crossed with anatomical masks of medial temporal lobe(R) hippocampus = yellow, left posterior cluster = green, a separate cluster in right parahiextend across parahippocampal gyrus, lingual gyrus, and fusiform gyrus.

4. General discussion

In studies of verbal fluency, including category production, theliterature tends to focus primarily on the critical role of the leftinferior frontal cortex and whether or not this frontal region issingular or further separable into phonological and semantic sub-components (see Costafreda et al., 2006, for review). Discussionof temporal lobe function has emphasized the importance of lat-eral temporal cortex, rather than medial temporal lobe regions,in category fluency tasks (Baldo, Schwartz, Wilkins, & Dronkers,2006; Schmolck et al., 2002). Medial temporal lobe involvementin category production has garnered less attention. However, sev-eral studies have found MTL activation during category production.For example, Pihlajamaki et al. (2000) found activation in the leftmedial temporal lobe including the hippocampus and posteriorparahippocampal gyrus during generation of exemplars for typi-cal categories such as clothing, food, animals, and plants (see also

Fig. 6. Mean effect sizes for each retrieval strategy contrasted with control in Experimentp < .005; Autobiographical Nonspatial (AN), Autobiographical Spatial (AS); Nonautobiogrlingual gyrus, and fusiform gyrus. The AS condition had a significantly higher mean magni

combination of activation in all three retrieval strategies contrasted with control at

regions. Five separate ROIs were used for analysis: left (L) hippocampus = red, rightppocampal gyrus = light blue, right posterior cluster = dark blue. Posterior clusters

Hirshorn & Thompson-Schill, 2006, Exp. 2; Mummery, Patterson,Hodges, & Wise, 1996). One reason for fewer reports of MTL acti-vation may be that scans are often limited to the frontal lobe, andthus do not assess all brain regions (Crosson et al., 2001; Paulesu etal., 1997).

Nevertheless, the consistent finding in the literature is thatamnesics with MTL damage are generally not impaired on categoryproduction, or for that matter, on many other tasks of seman-tic retrieval. In a recent paper, Schmolck et al. (2002) tested sixamnesic patients with damage either restricted to the MTL or dam-age that extended past MTL into anterior and lateral temporal lobe.Although some individual patients did poorly on specific semanticretrieval tasks, as a group, deficits were only obvious in the amnesicpatients with damage extending into lateral temporal areas.

Given the ability of individuals with hippocampal damage toperform adequately on this task, why do we see hippocampal acti-vation in normal participants performing category production? One

2 within medial temporal lobe combined activation regions of interest (ROI), clusteraphical, Nonspatial (NN). Posterior clusters extend across parahippocampal gyrus,tude of activation than AN and NN conditions in the left and right posterior clusters.

L. Ryan et al. / Neuropsychologia 46 (2008) 2109–2121 2117

Fig. 7. Brain regions outside the medial temporal lobes showing overlapping and strategy-specific activation in Experiment 2, cluster p < .0001. Areas of overlap are shown inred, areas specific to Autobiographical Nonspatial (AN) in yellow, areas specific to Autobiographical Spatial (AS) in green, and areas specific to Nonautobiographical, Nonspatial(NN) in blue. All three retrieval strategies activated the dorsomedial prefrontal cortex (DMPF), left dorsolateral and ventrolateral prefrontal cortex (DLPF, VLPF), thalamus(T), retrosplenial cortex (R), and fusiform/visual cortex (VC). AN retrieval uniquely activated the ventromedial prefrontal cortex (VMPF); AS retrieval uniquely activated leftsuperior lateral parietal cortex (LP) and left superior precuneus (P). The superior precuneus was also activated in the AN condition, but activation did not meet our extentcriteria of 10 contiguous voxels (only 8 voxels were observed). The midbrain (MB), a region of dorsal brainstem with activation extending to the inferior colliculus, wasactivated by NN retrieval only. Left superior precentral gyrus (Precentral) was activated by both AS and NN strategies.

Fig. 8. Mean effect sizes for each retrieval strategy in Experiment 2 contrasted with control within strategy-specific cluster regions of interest (ROI), cluster p < .0001;Autobiographical Nonspatial (AN), Autobiographical Spatial (AS); Nonautobiographical, Nonspatial (NN). The AN condition had a significantly higher magnitude of activationthan AS or NN in the ventromedial prefrontal cluster; the AS condition had a significantly higher magnitude of activation than AN or NN in the lateral parietal cluster; theAN and AS conditions had significantly higher magnitudes of activation than NN in the superior precuneus cluster (but note that AN activation did not exceed our extentthreshold of 10 contiguous voxels); the NN condition had a significantly higher magnitude of activation than AN or AS in the midbrain cluster, a region of dorsal brainstem;and the NN condition had a marginally significant higher magnitude of activation than AN in the precentral gyrus cluster.

2118 L. Ryan et al. / Neuropsycholog

possibility is that the hippocampus is utilized when and if it isavailable and intact, providing a more efficient and flexible wayof searching for and retrieving semantic information. By this view,the presentation of the category acts as a cue that activates relatedinformation, semantic and episodic alike. Related episodic infor-mation, such as familiar places or people, may be used by theparticipant to generate the semantic material in a sort of boot-strapping of semantic retrieval. The hippocampus may providesupport for retrieval and increase efficiency of search, particularlywith categories that are less well established, such as categoriesthat are based on organizing principles other than typical seman-tic relatedness, such as a particular location (things in a garage)or a shared perceptual feature (things that are red). These types

Table 3(A) Brain regions outside the medial temporal lobe activated by all three retrieval strategiep < .0001

Region L/R BA M

x

ADorsomedial prefrontal cortex B 32 −

ANa

ASa

NNa

Dorsolateral prefrontal cortex L 46 −ANa

ASa

NNa

Ventrolateral prefrontal cortex L 47 −ANa

ASa

NNa

Thalamus L – −ANa

ASa

NNa

Retrosplenial/posterior cingulate L 29 −ANa

ASa

NNa

R 29ANa

ASa

NNa

Fusiform/visual cortex B 19 −ANa

ASa

NNa

Strategy L/R BA MNI

x

BAN only

Ventromedial prefrontalb B 11 −4AS only

Lateral parietalb L 7 −26Superior precuneusb L 7 −4

NN onlyMidbrainb L – −4

AS & NNPrecentral gyrusb L 6 −50

ASa

NNa

Laterality (L = left, R = right, B = bilateral), Brodmann Area (BA), MNI coordinates, extent of afor each retrieval strategy (AN = Autobiographical Nonspatial, AS = Autobiographical Spatcingulate gyrus extending to medial prefrontal gyrus; dorsolateral prefrontal cortex includcortex includes inferior frontal gyrus extending to insula and superior temporal gyrus; fand cerebellum; ventromedial prefrontal cortex includes a region of medial frontal gyrussuperior precuneus regions extend to superior occipital cortex and the cuneus; midbrainto periaqueductal gray; and regions of precentral gyrus extend to middle frontal gyrus.

a Strategy.b Region.

ia 46 (2008) 2109–2121

of categories have been referred to by Barsalou (1983) as ad hoccategories.

What is special about ad hoc categories? During these tasks, onemust use semantic knowledge in a flexible and perhaps relativelynovel way, emphasizing characteristics of an item and similaritiesacross items that would otherwise be irrelevant. We hypothesizethat, under these conditions, amnesics with medial temporal lobedamage may do less well in generating category members, com-pared to generation tasks where the organizing principle amongstitems is common and overlearned. Amnesics might be relativelyintact on classic semantic categories, such as animals or fruitsand vegetables, relying primarily on intact non-medial temporalsystems, but may have more difficulty when asked to generate

s and (B) brain regions showing strategy-specific activation in Experiment 2, cluster

NI k T

y z

2 24 42294 4.89966 7.45633 4.01

50 24 2643 3.30

915 4.79990 4.97

38 24 250 3.71

413 5.55543 6.48

2 −11 1012 2.8640 3.6627 3.23

5 −58 894 4.97

162 7.1543 4.91

6 −48 619 4.3869 5.19

– –8 −66 −2

699 11.20749 14.96807 12.27

k T

y z

50 −10 70 3.93

−78 44 31 2.82−82 46 16 4.50

−30 −14 125 5.25

−1 5058 3.62

119 3.58

ctivation (in voxels; k), and peak voxel t-statistics are given for clusters of activationial, NN = Nonautobiographical Nonspatial). Dorsomedial prefrontal cortex includeses middle frontal gyrus extending to inferior frontal gyrus; ventrolateral prefrontalusiform/visual cortex includes lingual gyrus extending to cuneus, fusiform gyrus,

; lateral parietal cortex includes superior parietal lobule adjacent to the precuneus;refers to activation in the dorsal brainstem extending from the reticular formation

holog

L. Ryan et al. / Neuropsyc

exemplars to labels such as things that are expensive or things youtake on a camping trip. Our argument here is similar to the notion ofdegeneracy, discussed by Friston and Price (2003) and Price andFriston (2002), which emphasizes that there are multiple waysin which a complex cognitive task can be accomplished. We arenot suggesting that the hippocampus is merely “involved but notnecessary” for semantic retrieval. Instead, we propose that multi-ple retrieval systems are engaged cooperatively, depending uponthe requirements of the task, providing complementary routes forretrieval. Understanding the interaction of these systems will prob-ably require the integration of data from studies of both normal andbrain-injured individuals.

The concept of semantic and episodic memory as interactive andgenerative in the normal brain appears to have support not onlyfrom cognitive psychology (for example, Barsalou, Huttenlocher,& Lamberts, 1998), but also from evidence demonstrating over-lapping brain networks involved in the retrieval of semantic andepisodic memory. In a recent paper, Rajah and McIntosh (2005)used structural equation modeling to identify networks that mightdifferentiate semantic and episodic memory retrieval. Instead, sep-arate episodic and semantic models failed to differentiate fromone another, with similar patterns of path coefficients for the tworetrieval models across tasks. The authors argue that the samememory network was engaged across tasks, and that differencesbetween episodic and semantic retrieval may reflect variation alonga continuum of processing during task performance, rather than theoutput of two independent memory systems. A similar outcomewas recently reported by Burianova and Grady (2007), showingoverlapping networks of activation during semantic and episodicretrieval, which included regions of left medial temporal lobe.

According to Barsalou, “there are no invariant knowledge struc-tures in memory. Instead, people continually construct uniquerepresentations from loosely organized generic and episodicknowledge to meet the constraints of particular contexts”(Barsalou, 1988, p. 236). Instead of focusing on abstracted concepts,Barsalou emphasizes the critical role of instances for generatingsemantic knowledge. This is an interesting idea, because it presentssemantic memory as something that is not simply a stable andaccurate record of past learning, but something that is genera-tive, flexible, contextually bound, and subject to revision throughnovel experience. Semantic memory is generated anew each timeit is required, in much the same way as Bartlett (1932) and others(e.g., Nadel, Campbell, et al., 2007) have noted that episodic memo-ries are reconstructed and revised over time and through multiple

retrievals. This highlights the similarities between episodic andsemantic memories (Tulving, 2005) and stands in contrast to theassumption that semantic memory, at least, is a faithful record ofprior knowledge.

Other related memory research has recently demonstrated howsemantic and episodic memories can interact with one another.Using a very different paradigm, Westmacott and Moscovitch(2003) showed that episodic memory can contribute to per-formance on tests of semantic memory. They reported thatcategorization judgments regarding professions for famous peopleis faster and more accurate if the name is associated with a recollec-tion that has personal significance to the participant. In the examplethey cite, Elvis Presley might be associated with a personal visit toGraceland, whereas Frank Sinatra holds no such personal associa-tion. Performance therefore favors Elvis Presley, even though bothpeople are equally famous. For another individual, the oppositemight well hold true, depending upon their own unique episodicexperiences.

These considerations suggest that debates in the field aboutwhether or not episodic and semantic memory are part of thesame memory system, subserved by the same neural substrate,

ia 46 (2008) 2109–2121 2119

miss the point (see Moscovitch et al., 2006, for discussion). Differ-ences observed between episodic and semantic memory tasks maybe better understood in terms of the information required for anyparticular task, given that individuals will use whatever means areavailable to them, episodic or semantic, in order to solve the prob-lem. Episodic memory tasks, by definition, require the retrieval ofcontextual information including time and place. Even a recogni-tion task requires that the individual determine not only whether anitem has been experienced before (which would include all wordsin a list) but whether the particular word occurred at a particulartime and place. No such requirement is inherent in most semanticmemory tasks, where often the task requires a judgment about thesemantic qualities of words (such as in a living/nonliving judgmenttask). Engagement of the hippocampus may depend on the degreeto which performance requires (or can benefit from) the retrieval ofcontextual, spatial, and temporal information. In the case of normalindividuals, category production tasks can be solved by accessinga mix of episodic, autobiographic, and semantic knowledge. Forexample, if a subject is asked to retrieve items that are found in arestaurant, they may think about restaurants in general, or a restau-rant that they go to frequently, but they may be equally likely tovisualize the specific restaurant that they were at last night. As weand others have suggested, the hippocampus may best be viewed asa system that automatically uses inputs (in this case, a category cue)to generate linked information, thereby retrieving any appropriaterelated information.

One important implication of this formulation is that it suggeststhat the connections between hippocampal (episodic) represen-tations and extrahippocampal stores of semantic information arenot lost altogether over time. Rather, some of these connectionsremain, and can be used in the normal case to retrieve semanticknowledge connected to specific contexts. The fact that there arealternative routes for accessing semantic memory is not surprising.It remains to be seen how flexible these alternative routes are, andunder what circumstances they may or may not be sufficient for atask. This formulation is consistent with recent work exploring thefate of contextual coding in animals. A number of studies, mostlyusing fear conditioning, have shown that shortly after training, fearis relatively restricted to the original training context (Wiltgen &Silva, 2007; Winocur, Moscovitch, & Sekeres, 2007). With the pas-sage of time, and no further training, fear will be evinced in a widerrange of contexts. This “loss” of context-specificity has typicallybeen viewed as involving the loss of some hippocampal trace thatwould resist generalizing fear across contexts, but such an interpre-

tation is not mandated by the data. In fact, it is equally possible thata separate, “semantic” trace is created (outside the hippocampus)that connects fear to generic features of the original training con-text, such as grid floors, or a small box with four walls. This tracewould support fear in multiple contexts, but its presence need notinvolve the loss of the more specific trace supporting fear in theoriginal context. Much as in the human case, rats would have mul-tiple ways to access fear memory, some via the original trainingepisode, others via semantic knowledge about that experience.

The broader implication of this way of thinking concerns thenotion of a “memory system”. Perhaps it is no longer sensible totalk about self-contained memory systems, either singular or plu-ral. Rather, we should be talking about the acquisition of variouskinds of knowledge (see also Lee, Barense, & Graham, 2005; Nadel,2008), and the subsequent deployment of particular aspects of thatknowledge in the service of “memory” tasks. Within this view,all forms of knowledge would be subject to transformation (viaconsolidation) and updating (via reactivation and reconsolidation),and memory retrieval would involve accessing the appropriateknowledge to fit the task demands. There is neither one normultiple “memory” systems, only a variety of systems that both

cholog

2120 L. Ryan et al. / Neuropsy

process and store different types of information, to be deployedas required.

References

Addis, D. R., Wong, A. T., & Schacter, D. L. (2007). Remembering the past and imagin-ing the future: Common and distinct neural substrates during event constructionand elaboration. Neuropsychologia, 45, 1363–1377.

Andrewes, D. (2001). Neuropsychology: From theory to practice. New York: Taylor &Francis.

Ashburner, J., & Friston, K. J. (1999). Nonlinear spatial normalization using basisfunctions. Human Brain Mapping, 7(4), 254–266.

Baldo, J. V., Schwartz, S., Wilkins, D., & Dronkers, N. F. (2006). Role of frontalversus temporal cortex in verbal fluency as revealed by voxel-based lesionsymptom mapping. Journal of International Neuropsychological Society, 12(6),896–900.

Barsalou, L. W. (1983). Ad hoc categories. Memory & Cognition, 11(3), 211–227.Barsalou, L. W. (1988). The content and organization of autobiographical memories.

In U. Neisser & E. Winograd (Eds.), Remembering reconsidered: Ecological and tra-ditional approaches to the study of memory (pp. 193–243). Cambridge UniversityPress.

Barsalou, L. W., & Sewell, D. R. (1985). Contrasting the representation of scripts andcategories. Journal of Memory and Language, 24, 646–665.

Barsalou, L. W., Huttenlocher, J., & Lamberts, K. (1998). Basing categorization onindividuals and events. Cognitive Psychology, 36(3), 203–272.

Bartlett, F. C. (1932). Remembering. Cambridge: Cambridge University Press.Battig, W. F., & Montague, W. E. (1969). Category norms for verbal items in 56 cate-

gories: A replication and extension of the Connecticut category norms. Journalof Experimental Psychology, 80, 1–46.

Bayley, P. J., & Squire, L. R. (2005). Failure to acquire new semantic knowledge inpatients with large medial temporal lobe lesions. Hippocampus, 15(2), 273–280.

Bayley, P. J., Hopkins, R. O., & Squire, L. R. (2003). Successful recollection of remoteautobiographical memories by amnesic patients with medial temporal lobelesions. Neuron, 38(1), 135–144.

Bernard, F. A., Bullmore, E. T., Graham, K. S., Thompson, S. A., Hodges, J. R., & Fletcher,P. C. (2004). The hippocampal region is involved in successful recognition of bothremote and recent famous faces. Neuroimage, 22(4), 1704–1714.

Brett, M., Anton, J. L., Valabregue, R., & Poline, J. B. (2002). Region of interest analysisusing an SPM toolbox. In Paper presented at the 8th international conference onfunctional mapping of the human brain

Burgess, N., Maguire, E. A., Spiers, H. J., & O’Keefe, J. (2001). A temporoparietal andprefrontal network for retrieving the spatial context of lifelike events. Neuroim-age, 14(2), 439–453.

Burianova, H., & Grady, C. L. (2007). Common and unique neural activations in auto-biographical, episodic, and semantic retrieval. Journal of Cognitive Neuroscience,19(9), 1520–1534.

Cavanna, A. E., & Trimble, M. R. (2006). The precuneus: a review of its functionalanatomy and behavioural correlates. Brain, 129, 564–583.

Chan, A. S., Salmon, D. P., Butters, N., & Johnson, S. A. (1995). Semantic network abnor-mality predicts rate of cognitive decline in patients with probable Alzheimer’sdisease. Journal of International Neuropsychological Society, 1(3), 297–303.

Costafreda, S. G., Fu, C. H., Lee, L., Everitt, B., Brammer, M. J., & David, A. S. (2006).A systematic review and quantitative appraisal of fMRI studies of verbal flu-ency: Role of the left inferior frontal gyrus. Human Brain Mapping, 27(10),799–810.

Crosson, B., Sadek, J. R., Maron, L., Gokcay, D., Mohr, C. M., Auerbach, E. J., et al.(2001). Relative shift in activity from medial to lateral frontal cortex during inter-nally versus externally guided word generation. Journal of Cognitive Neuroscience,13(2), 272–283.

Dobbins, I. G., Rice, H. G., Wagner, A. D., & Schacter, D. L. (2003). Memory orien-tation and success: Separable neurocognitive components underlying episodicrecognition. Neuropsychologia, 41, 318–333.

Douville, K., Woodard, J. L., Seidenberg, M., Miller, S. K., Leveroni, C. L., Nielson, K. A.,et al. (2005). Medial temporal lobe activity for recognition of recent and remotefamous names: An event-related fMRI study. Neuropsychologia, 43(5), 693–703.

Duzel, E., Cabeza, R., Picton, T. W., Yonelinas, A. P., Scheich, H., Heinze, H. J., et al.(1999). Task-related and item-related brain processes of memory retrieval. Pro-ceedings of the National Academy of Sciences of the United States of America, 96(4),1794–1799.

Forman, S. D., Cohen, J. D., Fitzgerald, M., Eddy, W., Mintun, M. A., & Noll, D. C. (1995).Improved assessment of significant activation in functional magnetic resonanceimaging (fMRI): Use of a cluster-size threshold. Magnetic Resonance in Medicine,33, 636–647.

Friston, K. J., & Price, C. J. (2003). Degeneracy and redundancy in cognitive anatomy.Trends in Cognitive Sciences, 7, 151–152.

Friston, K. J., Holmes, A. P., Worsley, K. J., Poline, J. B., Frith, C. D., & Frackowiak, R.S. J. (1995). Statistical parametric maps in functional imaging: A general linearapproach. Human Brain Mapping, 2, 189–210.

Gilboa, A., Winocur, G., Grady, C. L., Hevenor, S. J., & Moscovitch, M. (2004). Remem-bering our past: Functional neuroanatomy of recollection of recent and veryremote personal events. Cerebral Cortex, 14(11), 1214–1225.

Gleissner, U., & Elger, C. E. (2001). The hippocampal contribution to verbal fluencyin patients with temporal lobe epilepsy. Cortex, 37(1), 55–63.

ia 46 (2008) 2109–2121

Glover, G. H., & Law, C. S. (2001). Spiral-in/out BOLD fMRI for increased SNRand reduced susceptibility artifacts. Magnetic Resonance in Medicine, 46(3),515–522.

Glover, G. H., & Lee, A. T. (1995). Motion artifacts in fMRI: Comparison of 2DFTwith PR and spiral scan methods. Magnetic Resonance in Medicine, 33(5),624–635.

Haist, F., Bowden Gore, J., & Mao, H. (2001). Consolidation of human memory overdecades revealed by functional magnetic resonance imaging. Nature Neuro-science, 4(11), 1139–1145.

Hayes, S. M., Ryan, L., Schnyer, D. M., & Nadel, L. (2004). An fMRI study of episodicmemory: Retrieval of object, spatial, and temporal information. Behavioral Neu-roscience, 118(5), 885–896.

Hirshorn, E. A., & Thompson-Schill, S. L. (2006). Role of the left inferior frontal gyrusin covert word retrieval: Neural correlates of switching during verbal fluency.Neuropsychologia, 44(12), 2547–2557.

Johnson, S. C., Baxter, L. C., Wilder, L. S., Pipe, J. G., Heiserman, J. E., & Prigatano, G. P.(2002). Neural correlates of self-reflection. Brain, 125(Pt 8), 1808–1814.

Kapur, N., Friston, K. J., Young, A., Frith, C. D., & Frackowiak, R. S. (1995). Activation ofhuman hippocampal formation during memory for faces: A PET study. Cortex,31(1), 99–108.

Lee, A. C. H., Barense, M. D., & Graham, K. S. (2005). The contribution of thehuman medial temporal lobe to perception: Bridging the gap between animaland human studies. The Quarterly Journal of Experimental Psychology, 58B(3/4),300–325.

Leveroni, C. L., Seidenberg, M., Mayer, A. R., Mead, L. A., Binder, J. R., & Rao, S. M.(2000). Neural systems underlying the recognition of familiar and newly learnedfaces. Journal of Neuroscience, 20(2), 878–886.

Lezak, M. D., Howieson, D. B., & Loring, D. W. (2004). Neuropsychological assessment(4th ed.). New York: Oxford University Press.

Luo, J., & Niki, K. (2002). Role of medial temporal lobe in extensive retrieval of task-related knowledge. Hippocampus, 12(4), 487–494.

Maguire, E. A., & Mummery, C. J. (1999). Differential modulation of a common mem-ory retrieval network revealed by positron emission tomography. Hippocampus,9(1), 54–61.

Maguire, E. A., Frackowiak, R. S., & Frith, C. D. (1997). Recalling routes around Lon-don: Activation of the right hippocampus in taxi drivers. Journal of Neuroscience,17(18), 7103–7110.

Maguire, E. A., Vargha-Khadem, F., & Mishkin, M. (2001). The effects of bilateral hip-pocampal damage on fMRI regional activations and interactions during memoryretrieval. Brain, 124(Pt 6), 1156–1170.

Maguire, E. A., Frith, C. D., Rudge, P., & Cipolotti, L. (2005). The effect of adult-acquiredhippocampal damage on memory retrieval: An fMRI study. NeuroImage, 27(1),146–152.

Manns, J. R., Hopkins, R. O., & Squire, L. R. (2003). Semantic memory and the humanhippocampus. Neuron, 38(1), 127–133.

Monsch, A. U., Bondi, M. W., Butters, N., Paulsen, J. S., Salmon, D. P., Brugger, P., et al.(1994). A comparison of category and letter fluency in Alzheimer’s disease andHuntington’s disease. Neuropsychology, 8, 25–30.

Moscovitch, M., Nadel, L., Winocur, G., Gilboa, A., & Rosenbaum, R. S. (2006). The cog-nitive neuroscience of remote episodic, semantic and spatial memory. CurrentOpinion in Neurobiology, 16(2), 179–190.

Mummery, C. J., Patterson, K., Hodges, J. R., & Wise, R. J. (1996). Generating ‘tiger’as an animal name or a word beginning with T: Differences in brain activation.Proceedings of the Royal Society of London Series B, Biological Sciences, 263(1373),989–995.

Nadel, L. (2008). Multiple memory systems: A new view. In J. Byrne (Ed.), Learningand memory: A comprehensive reference. Elsevier Press Ltd.

Nadel, L., Campbell, J. L., & Ryan, L. (2007). Autobiographical memory retrieval andhippocampal activation as a function of repetition and the passage of time. Neu-ral Plasticity [online]. Available: http://www.hindawi.com/, Article ID 90472, 14pp.

Nadel, L., Winocur, G., Ryan, L., & Moscovitch, M. (2007). Systems con-solidation and hippocampus: Two views. Debates in Neuroscience, 1(2–4), 55–66.

Northoff, G., Heinzel, A., de Greck, M., Bermpohl, F., Dobrowolny, H., & Panksepp,J. (2006). Self-referential processing in our brain—a meta-analysis of imagingstudies on the self. Neuroimage, 31(1), 440–457.

Paulesu, E., Goldacre, B., Scifo, P., Cappa, S., Gilardi, M. C., Castiglioni, I., et al. (1997).Functional heterogeneity of left inferior frontal cortex as revealed by fMRI. Neu-roReport, 8, 2011–2016.

Pihlajamaki, M., Tanila, H., Hanninen, T., Kononen, M., Laakso, M., Partanen, K., etal. (2000). Verbal fluency activates the left medial temporal lobe: A functionalmagnetic resonance imaging study. Annals of Neurology, 47(4), 470–476.

Price, C. J., & Friston, K. J. (2002). Degeneracy and cognitive anatomy. Trends inCognitive Sciences, 6, 416–421.

Rajah, M. N., & McIntosh, A. R. (2005). Overlap in the functional neural systemsinvolved in semantic and episodic memory retrieval. Journal of Cognitive Neuro-science, 17(3), 470–482.

Rosenbaum, R. S., Winocur, G., Ziegler, M., Hevenor, S. J., Grady, C. L., & Moscovitch,M. (2004). fMRI studies of remote spatial memory in an amnesic person. Brain& Cognition, 54(2), 170–172.

Ryan, L., Nadel, L., Keil, K., Putnam, K., Schnyer, D., Trouard, T., et al. (2001). Hippocam-pal complex and retrieval of recent and very remote autobiographical memories:Evidence from functional magnetic resonance imaging in neurologically intactpeople. Hippocampus, 11(6), 707–714.

L. Ryan et al. / Neuropsycholog

Saxe, R., Moran, J. M., & Scholz, J. (2006). Overlapping and non-overlapping brainregions for theory of mind and self reflection in individual subjects. Social Cog-nitive and Affective Neuroscience, 1(3), 229–234.

Schmolck, H., Kensinger, E. A., Corkin, S., & Squire, L. R. (2002). Semantic knowledgein patient H.M. and other patients with bilateral medial and lateral temporallobe lesions. Hippocampus, 12(4), 520–533.

Shannon, B. J., & Buckner, R. L. (2004). Functional-anatomic correlates of mem-ory retrieval that suggest nontraditional processing roles for multiple distinctregions within posterior parietal cortex. Journal of Neuroscience, 24(45),10084–10092.

Squire, L. R., & Zola, S. M. (1998). Episodic memory, semantic memory, and amnesia.Hippocampus, 8(3), 205–211.

Squire, L. R., Stark, C. E., & Clark, R. E. (2004). The medial temporal lobe. Annual Reviewof Neuroscience, 27, 279–306.

Tulving, E. (2005). Episodic memory and autonoesis: Uniquely human? In H. S. Ter-race & J. Metcalfe (Eds.), The missing link in cognition: Origins of self-reflectiveconsciousness (pp. 3–56). Oxford University Press.

Vallee-Tourangeau, F., Anthony, S. H., & Austin, N. G. (1998). Strategies for generatingmultiple instances of common and ad hoc categories. Memory, 6(5), 555–592.

ia 46 (2008) 2109–2121 2121

Van Overschelde, J. P., Rawson, K. A., & Dunlosky, J. (2004). Category norms: Anupdated and expanded version of the Battig and Montague (1969) norms. Journalof Memory and Language, 50(3), 289–335.

Walker, W. H., & Kintsch, W. (1985). Automatic and strategic aspects of knowledgeretrieval. Cognitive Science, 9, 261–283.

Welsh, K. A., Butters, N., Hughes, J. P., Mohs, R. C., & Heyman, A. (1992). Detectionand staging of dementia in Alzheimer’s disease. Use of the neuropsychologicalmeasures developed for the Consortium to Establish a Registry for Alzheimer’sDisease. Archives of Neurology, 49(5), 448–452.

Westmacott, R., & Moscovitch, M. (2003). The contribution of autobiographical sig-nificance to semantic memory. Memory & Cognition, 31(5), 761–774.

Williams, M. D., & Hollan, J. D. (1981). The process of retrieval from very long-termmemory. Cognitive Science, 5, 87–119.

Wiltgen, B. J., & Silva, A. J. (2007). Memory for context becomes less specific withtime. Learning & Memory, 14(4), 313–317.

Winocur, G., Moscovitch, M., & Sekeres, M. (2007). Memory consolidation or trans-formation: Context manipulation and hippocampal representations of memory.Nature Neuroscience, 10(5), 555–557.