thalamo-cortical disruption contributesto short-term · 2019. 1. 20. · for this,...

OR I G INA L ART I C L E

Thalamo-Cortical Disruption Contributes to Short-TermMemory Deficits in Patients with Medial TemporalLobe DamageNatalie L. Voets1,2, Ricarda A. L. Menke1, Saad Jbabdi1, Masud Husain1,2,3,Richard Stacey4, Katherine Carpenter5, and Jane E. Adcock1,2

1FMRIB Centre, Nuffield Department of Clinical Neurosciences, 2Epilepsy Research Group, Nuffield Department ofClinical Neurosciences, University of Oxford, John Radcliffe Hospital, Oxford OX3 9DU, UK, 3Department ofExperimental Psychology, University of Oxford, Oxford OX1 3UD, UK, 4Department of Neurosurgery and 5RussellCairns Unit, John Radcliffe Hospital, Oxford OX3 9DU, UK

Address correspondence to Natalie L. Voets, FMRIB Centre, University of Oxford, John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK.Email: [email protected]

AbstractShort-term (STM) and long-term memory (LTM) have largely been considered as separate brain systems reflecting fronto-parietal and medial temporal lobe (MTL) functions, respectively. This functional dichotomy has been called into question byevidence of deficits on aspects of working memory in patients with MTL damage, suggesting a potentially direct hippocampalcontribution to STM. As the hippocampus has direct anatomical connections with the thalamus, we tested the hypothesis thatdamage to thalamic nuclei regulating cortico-cortical interactionsmay contribute to STM deficits in patients with hippocampaldysfunction.We used diffusion-weightedmagnetic resonance imaging-based tractography to identify anatomical subdivisionsin patientswithMTL epilepsy. From these, wemeasured resting-state functional connectivity with detailed cortical divisions ofthe frontal, temporal, and parietal lobes. Whereas thalamo-temporal functional connectivity reflected LTM performance,thalamo-prefrontal functional connectivity specifically predicted STM performance. Notably, patients with hippocampalvolume loss showed thalamic volume loss, most prominent in the pulvinar region, not detected in patients with normalhippocampal volumes. Aberrant thalamo-cortical connectivity in the epileptic hemispherewasmirrored in a loss of behavioralassociation with STM performance specifically in patients with hippocampal atrophy. These findings identify thalamo-corticaldisruption as a potential mechanism contributing to STM deficits in the context of MTL damage.

Key words: epilepsy, functional connectivity, hippocampus, memory, thalamus

IntroductionSince the discovery that medial temporal lobe (MTL) damage caninduce profound amnesia while leaving immediate recall intact(Corkin 1984), short-term and long-term memory (STM and

LTM) have been largely considered as separate functional brainsystems (Wickelgren 1968; Shallice and Warrington 1970). Com-plementing a LTM circuit in the temporal lobe, the short-term re-tention and manipulation of information in “working” memory

© The Author 2015. Published by Oxford University Press.This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited.For commercial re-use, please contact [email protected]

Cerebral Cortex, November 2015;25: 4584–4595

doi: 10.1093/cercor/bhv109Advance Access Publication Date: 24 May 2015Original Article

http://creativecommons.org/licenses/by-nc/4.0/

http://creativecommons.org/licenses/by-nc/4.0/

http://www.oxfordjournals.org

has been allocated to a primarily pre-fronto-parietal network(Goldman-Rakic 1988; Owen et al. 1999; Todd and Marois 2004).

More recently, a direct contribution ofMTL structures towork-ing memory has been advanced, based on MTL neural activityduring STM tasks (Ranganath and D’Esposito 2001; Oztekinet al. 2010) and memory failures even across short delays inMTL-damaged patients (Olson, Moore, et al. 2006; Olson, Page,et al. 2006; Hartley et al. 2007; Pertzov et al. 2013). These data con-flict with other demonstrations of preserved STM despite exten-sive hippocampal lesions in human patients (Jeneson et al. 2010;Baddeley et al. 2011) and animal models (Alvarez et al. 1994). Toreconcile these findings, it is argued that relative hippocampalinvolvement might reflect how much task manipulations (i) ex-ceed immediate memory capacity (Jeneson and Squire 2012) or(ii) require complex stimulus associations (Yonelinas 2013). How-ever, uncertainty concerning the extent of MTL damage (Badde-ley et al. 2010) poses major interpretational challenges inhippocampal damage patients.

Indeed, the MTL has extensive connections including withretrosplenial cortex, thalamus, and prefrontal cortex (Aggleton2012). The thalamus in particular holds a privileged position incortico-subcortical and cortico-cortical bidirectional informationflow (Sherman andGuillery 2011).Markedmemory deficits followisolated thalamic damage (Dagenbach et al. 2001; Kubat-Silmanet al. 2002; Van derWerf et al. 2003) or electrically induced disrup-tion (Ojemann et al. 1971) in human patients. Experimental le-sion studies have identified key contributions of nuclear groupswithin the thalamus to both long-term (Aggleton and Mishkin1983; Sziklas and Petrides 1999) and working memory (Isseroffet al. 1982), paralleling dissociated functions of the frontal (Gold-man-Rakic 1988) and temporal (Mishkin 1982) cortex with whichthey are connected. Pertinently, postmortem studies in epilepticpatients with hippocampal damage reveal cell loss affectingmediodorsal and lateral thalamic nuclei (Sinjab et al. 2013),now detectable in vivo using advanced brain-imaging ap-proaches (Bernhardt et al. 2012; Keller et al. 2014). Such evidencefor thalamic atrophy raises the possibility that, in the context ofMTL injury, altered integrity of thalamic nuclei, and disruption oftheir associated respective thalamo-cortical functional circuits,may contribute to aspects of STM impairment.

Here, we used MRI techniques sensitive to thalamo-corticalconnectivity to examine whether altered thalamic integrity im-pacts on commonly usedmeasures of LTMand STMperformancein patients with temporal lobe epilepsy (TLE), selected to haveelectrophysiological evidence of hippocampal seizures accom-panied by normal clinical MRI or hippocampal atrophy. A well-

established anatomical connectivity-based approach based ondiffusion tensor MRI enabled us to quantify volumes of majorthalamic subregions and their relation to hippocampal atrophy.We used a complementary resting fMRI-based functional con-nectivity approach to correlate neural signals from thalamicsubregionswith detailed cortical divisions in the prefrontal, tem-poral, and parietal lobes. Finally, we related anatomical and func-tional measures of thalamic integrity to standardized clinicalmeasures of LTM and STM function. For this, neuropsychologicaltest scores were used in hierarchical multiple regression modelsto test the prediction that aberrant extra-MTL thalamo-corticalconnectivity contributes to impairments in STM performancein patients with MTL dysfunction. While not informing thenature of computations performed by the hippocampus or thal-amus in respect of LTM or STM, our aim was to elucidatepotentially differential contributions of dissociated thalamic net-works to memory function, as a step towards informing bothcomplexmemory systems organization and the potential contri-bution of extra-hippocampal disruption to STM deficits followinghippocampal injury.

Materials and MethodsParticipants

Eighteen right-handed patientswith TLE aged 31 ± 8.4 years (range18–49) were recruited through the Oxford Epilepsy SurgeryProgramme. Patients were selected from a larger cohort to haveseizures arising unilaterally from the left MTL based on compre-hensive clinical assessment including video-telemetry, high-reso-lution clinical diagnostic MRI, and neuropsychological evaluation.Clinical MRI revealed reduced hippocampal volumes accompan-ied by high signal on T2-weighted sequences, consistent withhippocampal sclerosis (Falconer et al. 1964), in 10 patients. MR im-aging was normal or equivocal in 6 patients and revealed subtleMTL focal cortical dysplasia without atrophy in two others.Patientswith seizures arising fromgross lesions (tumors, caverno-ma) were excluded. All patients had drug-resistant seizures (esti-mated frequency of typical complex partial seizures ranging from3 to 4 per week to clusters every 2 weeks approximately) and weretaking varied combinations of antiepileptic medications (seeTable 2). The mean age at onset of chronic seizures was 15.5 ± 9.3years (range: 2–36) and average duration of epilepsywas 16.5 ± 10.2years (range 5–42). Controls were 25 right-handed healthy volun-teers with no neurological or psychiatric history, age-matchedto patients (mean 32.3 ± 6.5 years, range: 20–49, independent

Table 1 Demographic and volumetric data

Measure Controls,Mean ± SD

Hippocampal atrophypatients, Mean ± SD

Normal hippocampal volumepatients, Mean ± SD

Age (years) 32.2 ± 6.5 34.5 ± 7.8 26.6 ± 7.2Epilepsy onset (years) — 15.5 ± 10.7 15.5 ± 8Epilepsy duration (years) — 19.3 ± 11.3 13 ± 8Left hippocampus (mm3) 4027 ± 422 2515 ± 332* (P < 0.001) 3772 ± 499 (n.s.)Right hippocampus (mm3) 4087 ± 365 3551 ± 518* (P = 0.001) 4126 ± 408 (n.s.)Left thalamus (mm3) 10 374 ± 745 9141 ± 1381* (P = 0.002) 9930 ± 558 (n.s.)Right thalamus (mm3) 9954 ± 643 8996 ± 1346* (P = 0.007) 9659 ± 647 (n.s.)Digit span (normalized z-score) — 9.1 ± 3.5 8.7 ± 3.7List learning (normalized z-score) — −0.76 ± 0.9 −0.79 ± 0.9

Note: Asterisks denote significant volume difference between patients and healthy controls. Hippocampal and thalamic volumes in the normal hippocampal volume

patients did not differ from healthy controls.

n.s., not significant.

Thalamo-Cortical Contributions to Memory Deficits Voets et al. | 4585

samples t-test, P = 0.58). Demographic data are presented inTable 1. Informed written consent was obtained from all partici-pants. The study was approved by the South London ResearchEthics committee.

Neuropsychology

Each patient had a comprehensive neuropsychological assess-ment as part of surgical evaluation including tests of general in-tellectual ability and attentional/executive function from theWechsler Adult Intelligence Scale (WAIS-IV), and tests of verbaland nonverbalmemory abilities from the Brain Injury Rehabilita-tion Trust Memory and Information Processing Battery (BMIPB).From the WAIS-IV, we selected a composite, age-scaled DigitSpan score, integrating performance over forward, backward,and sequence ordering of a list of aurally presented numbers.We chose this composite score rather than simple forward spanto capture both temporary maintenance of verbal information aswell as activemanipulation of the individual digits inmemory asan index of working memory function. Additionally, we selectedlist learning performance from the BMIPB as a measure of LTMacquisition sensitive toMTL integrity (Baxendale et al. 2012). Dur-ing this task, patients were read a list of 15 unrelated commonnouns, then asked to repeat back all items they could recall.This process was repeated a further 5 times with the same list re-peated to the patient followed by spontaneous recall. The totalnumber of words recalled over the five trials was converted tonormalized z-scores as an index of verbal learning. Digit spanwas notmeasured in 1 patient. Two patients weremissing verballearning scores. Individual patient scores are presented inTable 2.

Magnetic Resonance Imaging

MRI datawere acquired ona 3TSiemensVerio scanner using a 32-channel head coil. Anatomical T1-weighted structural imageswere acquired using a 3D MPRAGE sequence, providing isotropicvoxels of 1 × 1 × 1 mm3. Diffusion MRI datasets were obtainedusing an echo-planar imaging sequence (TE = 87 ms, TR = 9600ms, voxel size 2 × 2 × 2 mm3, 65 slices, GRAPPA acceleration

factor = 2) consisting of 8 nondiffusion-weighted and 60 diffu-sion-weighted images acquired with a b-value of 1000 s ×mm−2.T2*-weighted blood oxygen level dependent (BOLD) data were ac-quired during a 5-min resting fMRI scan using an echo-planar im-aging sequence (TR = 3.5 s, TE = 30 ms, slice thickness = 2 mm, 54slices, voxel size 2 × 2 × 2 mm3). Participants were asked to lie stillwith their eyes closed but remain awake.

Image Analysis

Subcortical Structural Segmentation: Thalamus and HippocampusTo determine overall subcortical volumes, the thalamus in eachhemispherewas segmented automatically fromeachparticipant’sT1-weighted structural image using FMRIB’s Integrated Registra-tion and Segmentation Tool (FIRST) (Patenaude et al. 2011). Eachsegmentation was carefully inspected and manually correctedwhere necessary to exclude any voxels encroaching on the hippo-campus posteriorly or the fornix medially. As hippocampal scler-osis reduces the accuracy of automated segmentation of MTLstructures in patients (Pardoe et al. 2009), hippocampiweremanu-ally outlined for every participant on their intensity-normalizedT1-weighted structural. Each hippocampus was manually deli-neated twice in every subject, at least 2 weeks apart, by an experi-enced rater (N.L.V.), naïve to radiological diagnosis in patients. Theintra-class correlation coefficient showed excellent intra-rater re-liability for manually defined volumes of both the left (0.96, 95%confidence interval: 0.92–0.98) and right (0.87, 95% confidenceinterval: 0.77–0.93) hippocampus. Therefore, the first and secondhippocampal segmentations were averaged for every subject.Finally, thalamic and hippocampal volumes were normalized forhead size bymultiplying themwith a volumetric scaling factor de-rived from the automated tool SIENAX as previously described(Menke et al. 2014).

Hippocampal SubdivisionsIncreasing evidence supports functional and anatomical divisionswithin the hippocampus (Aggleton 2012). The relative impact ofhippocampal damage on working memory performance maytherefore depend on the extent of atrophy affecting subregionsof the hippocampus. To divide each person’s hippocampus into

Table 2 Individual patient data

Patient Age Gender Age atonset

ClinicalMRI

Antiepileptic medications Digit span(normalized z-score)

List learning(normalized z-score)

1 24 F 15 HS CARB, LEV, PREGAB 10 0.442 18 F 9 HS LAM, CLOB, LEV 6 −0.53 40 F 30 FCD CARB, CLOB 9 −1.424 23 F 12 HS PHEN, LEV, CLOB, LAM 10 0.135 42 M 36 HS LAM, LEV 14 0.396 19 F 5 NV GAB, CLOB, CARB 5 N/A7 38 M 19 NV OXCARB, LACOS, CLOB, ACET 7 0.538 23 F 2 HS LEV, OXCARB, ZON 5 −1.239 35 F 10 HS CARB, TOP 6 −0.7210 24 F 17 NV LEV, CARB, CLOB N/A N/A11 32 M 22 FCD CARB, LEV, CLOB 13 −1.8312 29 F 7 NV LAC, CLOB, LEV, TOP 6 −1.3313 30 M 18 NV LAM, CARB 6 −0.8114 35 F 18 HS LAM, LEV 8 −1.4215 31 M 26 NV LEV, VAL 16 0.2316 49 M 7 HS LAM, GAB 8 −2.0917 37 M 5 HS LEV, CARB, CLOB 14 −0.8618 29 F 21 HS TOP, LAM, CLOB 9 −1.85

Note: CARB, carbamazepine; LEV, levetiracetam; PREGAB, pregabalin; LAM, lamotrigine; CLOB, clobazam; PHEN, phenytoin; GAB, gabapentin; OXCARB, oxcarbazepine;

LAC, lacosamide; ACET, acetazolamide; ZON, zonisamide; TOP, topiramate; VAL, sodium valproate.

4586 | Cerebral Cortex, 2015, Vol. 25, No. 11

anterior and posterior subregions, we used population averagemasks from our recent study in healthy controls and patientswith TLE, some of whom are also included in the present study(Voets et al. 2014). In brief, hippocampal functional divisionswere previously identified from resting fMRI data based on prefer-ential patterns of functional connectivity between the hippocam-pus and cortical regions including the temporal pole, entorhinalcortex, orbitofrontal cortex, posterior parahippocampal gyrus,lingual + fusiform cortex, thalamus, dorsolateral prefrontal cortex,and precuneus + posterior cingulate cortex. Patterns of hippocam-po-cortical functional connectivity were summed across patientsand healthy controls and thresholded at 50% to create a popula-tion average map revealing consistent anterior and posterior hip-pocampal subregions. Here, we nonlinearly aligned the previouslygenerated population averagemaps to the T1-weighted anatomic-al scan of each individual in the current study to obtain corre-sponding hippocampus subregion volumes, scaled to eachindividual’s high-resolution brain scan.

Cortical Region-of-Interest Parcellation Using FreesurferThemultiple nuclear divisions of the thalamus cannot be reliablyidentified on conventional anatomical MRI scans, but they aredissociable using diffusion-tractography methods sensitive todistinct anatomical connections of thalamic nuclei with large-scale cortical lobes of the brain (Behrens et al. 2003). To identifythese thalamic subdivisions in our patients and controls basedon anatomical connections with the cortex, we, therefore, firstcreated anatomical masks of the cortical lobes by combining in-dividual cortical parcellations obtained using FreeSurfer (v5.2).Individual subjects’ T1 volumes were linearly aligned to theMNI 305 average brain template, bias corrected, skull-stripped,and segmented into tissue types. The segmented white mattervolume was used to derive a surface representing the gray–white matter boundary, which was automatically corrected fortopology defects and carefully inspected in each participant for

accurate tissue classification, especially in the anterior temporallobes. The gray–white surface was inflated to form a sphere andwarped to match curvature features across subjects (Dale et al.1999; Fischl et al. 1999). After alignment to the spherical-spacestandard curvature template, the cortex was partitioned basedon gyral and sulcal structure using an automated segmentationprocedure (Desikan et al. 2006). The resulting hemisphere-latera-lized cortical parcellations were reformatted and converted intobinary masks for compatibility with FMRIB’s Software Library.

For diffusion-based classifications, the detailed FreeSurfer-de-rived cortical parcellations for each individual were combined atthe “lobe” level to obtain large-scale cortical masks representingthe occipital lobe, temporal lobe, prefrontal cortex, parietal lobe,precentral, and postcentral regions, for eachhemisphere separate-ly (Fig. 1a). However, for functional connectivity analyses, each ofthe detailed FreeSurfer-derived cortical parcellations within thefrontal, temporal, and parietal regions were considered separately(see below) foranatomically fine-grained analyses. To restrict rest-ing fMRI partial correlation analyses to regions most likely to berelevant for memory processing, we selected a subset of the de-tailed cortical Freesurfer parcellations. For the frontal lobe, this in-cluded the superior, rostral middle frontal and caudal middlefrontal gyri, themedial and lateral orbitofrontal regions, and front-al pole but excluded the 3 Broca’s area FreeSurfer parcellations foreach hemisphere. For the parietal lobe, we included all four Free-surfer parcels (the superior and inferior parietal lobules, supra-marginal gyrus, and precuneus). For the temporal lobe, weselected the superior, middle, inferior temporal, and parahippo-campal gyri, the fusiform and entorhinal cortices and the tem-poral pole, but omitted the “banks of the superior temporalsulcus” and “transverse” parcels (Fig. 1c).

Diffusion-Based Segmentation of the ThalamusWe used a well-characterized diffusion-based classificationapproach (Behrens et al. 2003) to isolate thalamic subregions in

Figure 1. Structural and functional thalamo-cortical connectivity analyses. (a) An automated diffusion-based classification approach was used to identify probabilistic

anatomical connections between large-scale cortical lobes (top left) and every voxel in the thalamus (thalamus shown in black). This approach segments the

thalamus into subregions showing distinct anatomical connectivity profiles (b) (red = voxels connected to prefrontal cortex, green = parietal lobe, yellow = temporal

lobe) shown in an example healthy control and a patient. (c) Resting-state functional MRI signal correlation analysis between the anatomical connectivity-defined

parcels of the thalamus (red = voxels connected to prefrontal cortex, green = parietal lobe, yellow = temporal lobe) and detailed FreeSurfer-derived cortical

parcellations within each respective lobe, as well anterior and posterior hippocampus subregions from previously generated population maps (Voets et al. 2014). TLE,

temporal lobe epilepsy.


each individual. Diffusion datawere skull-stripped and affine-re-gistered to the first, nondiffusion-weighted volume to correct foreddy current-induced distortions and head motion. Voxel-wiseestimates of fiber orientations (with up to 2 fibers per voxel)and their uncertainty were calculated using FMRIB’s DiffusionToolbox (FDT). The 6 Freesurfer-derived large-scale cortical“lobe” masks were then used in a diffusion-based probabilistictractography analysis, to classify every voxel in the thalamus ac-cording to the cortical “lobe” to which it is preferentially con-nected (Behrens et al. 2003; Keller et al. 2014). The resulting sixclassification parcels for each individual were converted into vo-lumes of thalamic regions preferentially connected to each lobe.Thalamic subregion volumes were first scaled by intracranialvolume and then converted to ratios (volume of thalamic voxelspreferentially connected to a cortical lobe divided by volume ofthat cortical lobe) to minimize group differences in cortical lobesizes impacting on structural connectivity.

Functional Connectivity Analysis Between Thalamus Subregionsand Cortical Regions of InterestsNext, to quantify functional connectivity (FC) between each ofthe diffusion-defined segments within the thalamus, and fine-grained cortical regions most likely to be involved in memoryfunction, we measured fMRI signal correlations between the de-tailed Freesurfer-based subparcellations of the prefrontal, tem-poral, and parietal lobe and the respective thalamic segmentsfrom the diffusion-based classification analysis. Analyses wereperformed separately for the left and right hemispheres.

Resting fMRI data preprocessing included correction for headmotion, correction for geometric distortions at air tissue boundar-ies using fieldmaps, spatial smoothing using a 5 mm full-widthhalf-maximum Gaussian kernel and high-pass filtering (100 s) toreduce low frequencyartefacts. A seed-based correlationapproach(SBCA) (O’Reilly et al. 2010) was used to measure thalamo-corticalFC. First, every participant’s prefrontal, temporal, and parietalthalamic diffusion-based segments (see Fig. 1b for representativesegmentations in a healthy control and TLE patient) were regis-tered to their resting fMRI data using a boundary-based optimizedregistrationmethod (Greve and Fischl 2009). Next, individuals’ de-tailed FreeSurfer 1 mm pial parcellations (Fig. 1c) were smoothedwith a Gaussian filter of 2 mm, registered to their 2 mm fMRIdata, and then masked to exclude voxels classified as cerebro-spinal fluid (CSF) from automated individual subject tissue seg-mentations generated using FMRIB’s Automated SegmentationTool (FAST). Finally, 3 SBCAswere performed in each hemisphere.

Resting fMRI signal from each individual’s “prefrontal” thal-amus segment was correlated with the characteristic time courseof each of their 6 prefrontal cortical parcels (superior, rostral mid-dle frontal and caudal middle frontal gyri, medial and lateral orbi-tofrontal regions and frontal pole). This was repeated, correlatingfMRI signal from the thalamic “parietal” segmentwith that of eachof 4 parietal cortical parcels [superior and inferior parietal lobules,supramarginal gyrus and precuneus as defined on the Desikan–Killianyatlas (Desikanet al. 2006)], andfinally, fromthe “temporal”thalamus segment with each of 7 temporal lobe cortical parcels(superior, middle, inferior temporal and parahippocampal gyri,fusiformand entorhinal cortices, temporal pole), aswell as the an-terior and posterior hippocampus. The time courses representinghead motion, CSF, and white matter were regressed out to reducethe influence of structured noise.

Thus, for every participant, we obtained 19 correlationmaps ineachhemisphere: 9measuring FCof the “temporal” thalamus seg-ment with each temporal lobe subregion, 6 measuring FC of the“frontal” thalamus segment with each frontal lobe parcel, and 4

measuring FC between the “parietal” thalamic segment and par-ietal lobe subregions. Each partial correlation map represented,for every voxel in the relevant thalamus segment, its signal correl-ation magnitude (−1 to 1) with a given cortical subregion in thesame hemisphere (accounting for the magnitude of correlationwith every other cortical parcel in that lobe). From this, wecalculated for every participant the average FC of a given thalamicsegment with a respective cortical parcel.

Statistical AnalysesImaging measures were compared between patients and controlsand related to neuropsychological scores using statistical testsimplemented in SPSS Statistics (v21). Structural volumetric anddiffusion-based measures were compared between groups usingmultivariate ANCOVAs, co-varying for age. Thalamo-cortical FCfrom the three thalamic segments in each hemisphere was com-pared between groups throughmultivariate ANCOVAs, co-varyingforageaswell as sizes of the relevant thalamic segment. FC resultswere Bonferroni-corrected for multiple comparisons separatelyfor each thalamus segment with its corresponding set of correl-ation maps (9 in each temporal lobe, 6 in each frontal lobe, and 4in each parietal lobe). Relationships between behavioral and im-aging measures in patients were established first using 2-tailedPearson’s partial correlations, removing variability associatedwith age. Subsequently, brain-imaging measures associated withSTM performance were entered into a hierarchical linear regres-sion model to assess their unique predictive contribution toperformance variability in this memory domain.

ResultsGlobal Thalamus Volumes

Multivariate ANCOVAs revealed a significant effect of group onoverall thalamus sizes. Relative to healthy controls, TLE patientshad smaller thalamic volumes, both in the ipsilateral left (F = 9.55,P = 0.004) and contralateral right hemisphere (F = 6.15, P = 0.017)(Fig. 2b). Across patients and controls, as well as within each sub-group, hippocampal volume correlated with thalamus volume inboth hemispheres (left hemisphere: R = 0.58, P < 0.001; right hemi-sphere: R = 0.54, P < 0.001) (Fig. 2c). Duration of epilepsy did notcorrelate with either thalamic or hippocampal volumes (P > 0.1).

Volumes of Connectivity-Defined Thalamus Subregions

Multivariate ANCOVAs revealed an effect of group also on “ipsilat-eral,” but not contralateral, thalamic subregional volumes for TLEpatients compared with controls (F = 3.51, P = 0.024). This was dueto a significant reduction in thalamic voxels preferentiallyconnected to the parietal lobe in the epileptic (left) hemisphere(F = 6.40, P = 0.015) (Fig. 3b). There was no difference between pa-tients and controls in thalamo-prefrontal or thalamo-temporalvolume ratios.

Thalamo-Cortical FC

The TLE patient group showed abnormally increased FC relativeto controls between the contralateral (right) but not the ipsilateral(left) thalamic “parietal segment” and precuneus after Bonferro-ni-correction for four parietal cortical parcels (F = 7.52, correctedto P = 0.036). However, there was no significant difference be-tween the patients and controls in FC between the thalamus dif-fusion-defined prefrontal segment and any cortical subregionswithin the prefrontal lobe, either ipsi- or contralaterally.


Similarly, FC between the thalamus “temporal segment” andtemporal lobe subregions in patients remained within the nor-mal range.

Impact on Neuropsychological Performance

To determine the extent to which observed changes in thalamicvolumes and connection patterns impacted onmemory integrityin TLE patients, we first explored associations between our struc-tural and functional MRI measures and neuropsychological per-formance on test of STM and LTM. Seven patients achievedcomposite digit span scores of 7 or less (mean score for TLEgroup: 8.94 ± 3.5, range: 5 to 16), while 5 patients had list learningscore that fell in the borderline/impaired range (z-scores of −1.34or lower;mean for TLE group:−0.77 ± 0.89, range0.53 to−2.09).Nei-ther duration of epilepsy nor age at onset of habitual seizures cor-related significantly with either neuropsychological test measure.

STM (composite digit span) performance in patients was notsignificantly correlated with global hippocampal volumes. How-ever, smaller volumes of the posterior ipsilateral hippocampuswere associated with poorer digit span (R = 0.55, P = 0.027). Neitherglobal nor subregional thalamus volumes correlated with digitspan. Instead, lower digit span in patients was associated withmeasures of both thalamo-prefrontal and thalamo-parietal FC.Specifically, lower FC between the contralateral thalamicprefrontal segment and caudal middle frontal gyrus (R = 0.70,P = 0.003) (Fig. 4a) was associated with lower digit span; this wasalso a trend ipsilaterally (R = 0.46, P = 0.075). Similarly, lower FCbetween the contralateral thalamic “parietal segment” and thesupramarginal gyrus reflected lower digit span performance(R = 0.53, P = 0.034).

In contrast to these STM associations with thalamo-prefront-al and thalamo-parietal FC, LTM performance reflected thalamo-temporal FC. Lower list learning performance was seen inpatients with lower ipsilateral thalamo-entorhinal FC (R = 0.57,P = 0.028) (Fig. 4b). In addition, list learning was negatively corre-lated with volume ratios of the thalamic “prefrontal segment”both ipsi- (R = −0.67, P = 0.006) and contralaterally (R = −0.72,P = 0.002) in patients. However, this was not reflected in accom-panying associations between list learning performance andeither thalamo-prefrontal or thalamo-parietal FC.

Finally, to directly assess the specificity of this apparent dis-sociation in thalamo-cortical pathways associated with STM,we performed a hierarchical multiple regression analysis. Forthis analysis, we included age in the first step of the multiple

regression, and in the second step, we modeled the unique pre-dictive value of each measure associated with digit span (fromthe analysis above), to identify the magnitude and order inwhich each contribute to STM performance. We also includedthalamo-entorhinal resting connectivity (found to correlatewith list learning), to help interpret the specificity of our correl-ation findings. This model as a whole was a significant predictorof STM performance, accounting for 86% of variance in digit span(F = 10.2, P = 0.001). Age explained 10.4% variance in digit spanperformance. When controlling for age, our imaging measuresaccounted for an additional 75.6% variance (significant F change:0.001). Of the individual variables, volume of the posterior ipsilat-eral hippocampus (β = 0.47, P = 0.007) and FC between the contra-lateral thalamus and caudal middle frontal gyrus (β = 0.99,P = 0.013)made significant unique contributions to themodel. Ip-silateral thalamo-caudal middle frontal (P = 0.09), contralateralthalamo-supramarginal (P = 0.16), and, importantly, ipsilateralthalamo-entorhinal (P = 0.26) resting FC did not contributesignificantly to digit span performance. Removing ipsilateral tha-lamo-caudal middle frontal connectivity (correlated with contra-lateral thalamo-caudal middle frontal connectivity) from themodel did not significantly impact these findings.

Finally, we repeated the hierarchical linear regression ana-lysis, controlling for age and modeling each of the measures as-sociated with list learning to identify the most significantpredictors of LTM performance. As above, we included in theLTM model also those variables correlated with digit span, to as-sess the specificity of thalamo-cortical measures for both mem-ory domains. The model as a whole was a significant predictor ofLTM performance (F = 5.16, P = 0.022), explaining 92.5% of vari-ance in list learning. Age explained 22.9% variance in list learningperformance. After controlling for age, the neural measures ex-plained an additional 80.2% variance (significant F change =0.019). Of the individual measures, only ipsilateral thalamo-en-torhinal cortex resting connectivity contributed significantly tolist learning performance (β = 0.72, P = 0.049). Neither ipsilateralnor contralateral thalamic–prefrontal connectivity volume ratios,nor any of the measures correlated with digit span performance,reached significance as predictors of list learning in this model.

Effect of Hippocampal Atrophy on Thalamo-CorticalInteraction

Both secondary deafferentation of thalamic connections from anatrophic hippocampus andpropagation of epileptic activity could

Figure 2. Global thalamic volumes in epilepsy and relation to hippocampal atrophy. (a) Example automated segmentation of the thalamus in a representative healthy

control. (b) Overall thalamus volumes were reduced in TLE patients (light grey bars) relative to healthy controls (white bars) (boxplots depict mean [± 2 standard

deviations]) in both hemispheres. (c) Hippocampal and thalamus volumes were significantly correlated in all subjects (R = 0.58, P < 0.001).


in theory contribute to functional thalamic disruption in TLE(Sinjab et al. 2013). If thalamo-cortical dysfunction reflects thethalamus’ role in propagation of epileptic activity, similar pat-terns of disruption might be expected across all patients withchronic TLE, regardless of the size of the hippocampus. Con-versely, if hippocampal sclerosis compounds thalamic atrophy,greater thalamo-cortical disruption would be expected in pa-tients with smaller hippocampal volumes. To investigate thesepossibilities, we performed a preliminary analysis, dividing ourpatients into 2 groups: Those with ipsilateral hippocampal vo-lumes 2 standard deviations or more below the range of ourhealthy controls (“atrophy” group, n = 10) and those with ipsilat-eral hippocampal volumes within the normal range (n = 8).

Despite the reduced sample sizes, patients with hippocampalatrophy—but not those with normal hippocampal volumes—showed bilateral reductions in global thalamic volumes (left:F = 10.91, P = 0.002; right: F = 7.73, P = 0.009) (Fig. 2c), as well as involume ratios of connections with the ipsilateral parietal lobe(F = 4.34, P = 0.045).

Patients with and without hippocampal atrophy showed cor-respondingly divergent associations between thalamo-corticalcommunication and digit span. In normal volume patients,digit span performance correlated bilaterally with both thala-mo-caudal middle frontal FC (right hemisphere: R = 0.792,P = 0.034, left hemisphere: R = 0.788, P = 0.036) and thalamo-supramargical FC (right hemisphere: R = 0.915, P = 0.004, lefthemisphere: R = 0.742, P = 0.056).

Conversely, in patients with left hippocampal volume loss,only FC between the contralateral (right) thalamus and caudalmiddle frontal gyrus remained significantly correlated withdigit span (R = 0.744, P = 0.014). The loss of ipsilateral thalamo-prefrontal FC association with digit span may reflect aberrant,heightened thalamo-caudal middle frontal gyrus FC seen in theaffected (left) hemisphere (F = 4.04, P = 0.053, corrected for thal-amic prefrontal segment volume) of patients with hippocampalatrophy, but not patients with normal volumes, relative tocontrols.

We did not contrast list learning between the patient sub-groups as both patientsmissing these valueswere from the smal-ler normal volume group.We did not repeat hierarchicalmultipleregression analyses based on the reduced sample size in thesesubgroups.

DiscussionDeficits in amnestic patients on some tests of STM (Nichols et al.2006; Olson, Page, et al. 2006; Hartley et al. 2007; Pertzovet al. 2013), but not others (Drachman and Arbit 1966; Zarahnet al. 2005; Jeneson et al. 2010; Baddeley et al. 2011), have sparkeddebate about a potentially direct role for thehippocampus in STM(Cashdollar et al. 2011). Here, we tested an alternative possibilitythat deficits on aspects of STM reflect disruption of thalamo-cor-tical communication associated with MTL damage. MultimodalMRI was used to quantify thalamo-cortical connectivity in

Figure 3. Anatomical connectivity of the thalamus in temporal lobe epilepsy. (a) Volume ratios were calculated of tractography-defined voxels within the thalamus

showing preferential anatomical connections with the temporal lobe (yellow), frontal lobe (red), and parietal lobe (green). (b) Asterisks denote significant reductions

specifically in anatomical connections between the thalamus and parietal lobe in patients relative to controls (P < 0.05), that was driven by the subgroup of patients

with hippocampal atrophy. Boxplots depict volume ratios (± 2 standard deviations) of thalamic connectivity-defined regions in controls (white bars) and patients with

TLE (gray bars). TLE, temporal lobe epilepsy; PFC, prefrontal cortex; TL, temporal lobe; PL, parietal lobe.


patients with epilepsy arising from the left MTL, with and with-out measurable hippocampal atrophy. Resting-state FC analysesrevealed a unique contribution of extra-MTL thalamo-corticalpathways to STM performance. Furthermore, thalamic micro-structure and cortical FC reflected magnitudes of MTL damagein patients, with the greatest thalamic structural and functionaldisruption seen in patients with abnormally reduced hippocam-pal volumes. These results provide evidence that thalamo-cortical disruption might contribute to STM impairments inMTL-damaged patients.

It is increasingly recognized that thalamic nuclei do not simplyrelay information, but actively contribute to cognitive processes inways consistent with—though perhaps qualitatively distinct from—their associated neocortical areas (Hunt and Aggleton 1991; deBourbon-Teles et al. 2014). Although the role of thalamic nucleiin memory processes remains incompletely understood, lesionand electrophysiological data suggest that LTM symptoms mostcommonly arise after anterior thalamic nucleus damage, whileSTM deficits may involve ventral mediodorsal, midline, andventral anterior structures (Van der Werf et al. 2003). Verbal STMdeficits have also been elicited with electrical stimulation of theleft pulvinar nucleus (Ojemann and Fedio 1968) and ventrolateralsites (Ojemann et al. 1971), lesions to which produce deficits ofvisual attention orientation (de Bourbon-Teles et al. 2014).

Thalamic contributions to memory are thought to reflect par-tially independent routes of information flow between thalamic

subregions and functionally distinct neocortical areas (Bentivo-glio et al. 1997). Within the mediodorsal nucleus, regarded asthe prefrontal cortex “gateway” (Tanibuchi and Goldman-Rakic2003), anatomical connections point to at least 3 neural circuitspotentially supporting memory processing in primates (Mitchelland Chakraborty 2013). These include a medial system intercon-nected with orbitofrontal and ventromedial prefrontal cortex aswell as MTL structures; a central system preferentially connectedwith the dorsolateral prefrontal cortex (but not with the MTL);and a lateral system consisting of intralaminar nuclei with dif-fuse prefrontal, anterior cingulate, and basal ganglia projections.Conversely, the anterior thalamic nuclei form a separate circuitinterconnected with the hippocampal formation, mammillarybodies, anterior cingulate, and retrosplenial cortex (Aggletonet al. 2010). Finally, lateral posterior nuclei, including the “associ-ation” pulvinar complex, show extensive connections includingwith prefrontal, posterior parietal, and limbic regions (Romanskiet al. 1997). Differential disruption to these parallel thalamo-cor-tical pathwayswould therefore be predicted to explain variabilityin the nature and extent of memory deficits.

Consistent with this prediction, in our study, STM and LTMdeficits reflected FC within distinct thalamo-cortical networks.Lower digit span performance in patients was associated withlower resting-state fMRI signal correlations between the regionswithin the thalamus preferentially connected to the prefrontalcortex (based on anatomical connectivity) and the caudal middle

Figure 4.Association between thalamo-cortical connectivity and STM and LTM performance. (a) In patients, thalamo-prefrontal FC (with the caudal middle frontal gyrus)

correlated with measures of STM (R = 0.70, P = 0.003) but not list learning. (b) Conversely, thalamo-temporal FC (with the entorhinal cortex) correlated with LTM (R = 0.57,

P = 0.028) but not STM performance. FC, functional connectivity.


frontal gyri. Tract tracing studies in nonhuman primates indicatepreferential connections from the central mediodorsal region tothis cortical area (Mitchell and Chakraborty 2013), which hasbeen attributed a specific role in ordering information in STM(Henson et al. 2000).

In contrast, impaired list learning performance reflected FCbetween the entorhinal cortex and the connectivity-defined thal-amic “temporal” segment, grossly corresponding to the anteriornucleus and parts of the medial mediodorsal nucleus anatomic-ally interconnected with the MTL. The entorhinal cortex consti-tutes the major communication route of the hippocampus andhas been shown to both interact with the hippocampus during(Igarashi et al. 2014) and independently contribute to (Gaskinand White 2013; Yang et al. 2014) aspects of learning.

These findings support fMRI and electrophysiological datademonstrating key roles for thalamo-cortical communication inSTM processing. Working memory load-related neural activityin dorsolateral prefrontal and parietal cortices is also seen inthe thalamususing fMRI (Callicott et al. 1999). Electrophysiologic-al recordings further demonstrate coordinated prefrontal, par-ietal, and thalamic neuronal firing during task delay periods innonhuman primates (Fuster and Alexander 1971; Tanibuchiand Goldman-Rakic 2003), disruption of which induces workingmemory deficits in rodents (Parnaudeau et al. 2013).

Interestingly, digit span reflected contralateral, right hemi-sphere thalamo-cortical FC, while list learning was predicted byipsilateral, left hemisphere thalamo-temporal FC. The most par-simonious explanation for this hemispheric dissociation is thatlateralized left temporal lobe damage is known to produce defi-cits in verbal learning and LTM (Milner 1971). Our finding of re-duced list learning with reduced left thalamo-entorhinal FC isconsistent with longitudinal observations showing that magni-tudes of verbal LTM loss reflect extents of functional left MTL tis-sue damage (Powell et al. 2008). Conversely, imaging studiesoften reveal bilateral fronto-parietal activity during tasks involv-ing STM, thought to reflect stimulus-independent complex pro-cessing demands (Nystrom et al. 2000; Wager and Smith 2003;Chein et al. 2011) within widely distributed circuits subservingworking memory (Goldman-Rakic 1988). We therefore speculatethat in the context of unilateral disruption to bilaterally repre-sented STM networks, patients may rely on remaining (perhapsless efficient) thalamo-cortical connections in the unaffectedhemisphere.

This interpretation is supported by disrupted anatomical andfunctional thalamo-cortical connectivity observed in patientswith—but not without—hippocampal atrophy. We found thathippocampal atrophy patients showed fewer parietal lobe con-nections and abnormally elevated thalamo-prefrontal FC in theepileptic hemisphere. Concurrently, hippocampal atrophy pa-tients showed a loss of functional correlations with STM withinthe corresponding networks. Indeed, while in hippocampus-in-tact patients, digit span reflected bilateral thalamo-prefrontaland thalamo-parietal FC, in hippocampal-atrophy patients,digit span performance correlated only with contralateral thala-mo-prefrontal communication. A loss of parietal anatomical con-nections could possibly disrupt functional associations reliant oncommunication between medial pulvinar/lateral dorsal and su-pramarginal neurons reportedly involved in phonological storage(Henson et al. 2000) and numbermanipulations (Price et al. 2013).The lack of indication for abnormal thalamo-supramarginal FC inpatients implies potentially independent contributions of thesethalamic nuclei to aspects of verbal STM (Ojemann and Fedio1968). Additionally, hippocampal atrophy patients showed ab-normally heightened FC between the thalamus and caudal

middle frontal gyrus in the epileptic hemisphere, which no long-er correlatedwith digit span. Although themechanism for this FCdisruption is unclear, these results mirror selective limbic fiberdegeneration seen in patients with—but not without—hippo-campal sclerosis, especially along the fornix (Concha et al.2009) which connects the hippocampus with the thalamus andmammillary bodies (Aggleton et al. 1986).

What about the role of the hippocampus in STM? This studywas not specifically designed to disentangle mnemonic computa-tions performed by the hippocampus (or indeed the thalamus), butsome findings might be pertinent. Although global hippocampusvolumes were not related to digit performance, smaller volumeof specifically the posterior epileptic hippocampus correlated withand uniquely predicted variance in digit span performance. Thisregion is considered to form part of a posterior memory networkcomposed of the retrosplenial cortex, dorsolateral prefrontal cor-tex, andposterior sensory cortices, in contrast toananterior hippo-campal memory network including the amygdala, orbitofrontalcortex, and temporal pole (Aggleton 2012). However, the associ-ation between digit span and posterior hippocampal volumesdoes not allow us to dissociate the role of direct connections todorsolateral prefrontal cortex via the cingulum or fronto-occipitalfasciculus (Goldman-Rakic et al. 1984) from indirect thalamo-cor-tical projections via the fornix, since chronic epilepsy patientsshow structural damage along the entire limbic circuit (Fockeet al. 2008).

Which experimental task demands/manipulations unveil amemory deficit across short delay intervals in patients with hip-pocampal lesions remains actively debated. While many taskshave focused on relational binding, deficits in associative mem-ory are not the only reported findings. For example, impairmentshave also been identified for faces (Nichols et al. 2006; Olson,Moore, et al. 2006). Conversely, not all associative tasks are im-paired in MTL-damaged patients. Simple associative memoryperformance appears intact when trials involve a small numberof associations, with impairments emerging only on larger setsizes (Jeneson et al. 2010; Pertzov et al. 2013). These findingsecho early observations of markedly reduced general storagecapacity in patients with hippocampal lesions on an extendeddigit span task, most pronounced on digit set sizes exceeding 6(Drachman and Arbit 1966).

An alternative conceptual framework to associative/non-associative processing for these previously observed STM deficitsis the level towhich stimulus computations exceed the spanof im-mediate/STM and begin to call upon LTM “supraspan” resources(Yonelinas 2013). Drachman and Arbit already in 1966 proposedthat complex information, including difficult item associationsbut perhaps also feature-dependent stimuli such as faces (Nicholset al. 2006), may rely on LTM stores – just like supraspan informa-tion – depending on how many item/information features can beheld in immediate memory (Drachman and Arbit 1966). Conse-quently, it has been proposed that instead of the MTL being critic-ally involved in certain types of STM, some types of task call uponLTM processes (Jeneson and Squire 2012). This distinction mightpotentially explain conflicting neuroimaging findings of hippo-campal activity during the delay period for novel face (Ranganathand D’Esposito 2001) but not letter (Zarahn et al. 2005) stimuli.

Our results lend support to the notion that thalamic nucleimay pivotally contribute to deficient STM performance in pa-tients with MTL damage, perhaps by mediating complex compu-tational demands between immediate and long-term stores(Vertes et al. 2007). Thalamic neurons contribute to informationprocessing and gating (Fuster and Alexander 1971; Vertes et al.2007; Rotshtein et al. 2011) as well as associative learning


(Winocur 1985; Hunt and Aggleton 1991; Gibb et al. 2006), seeMitchell and Chakraborty 2013 for a review), and perhaps stimu-lusmaintenance or selection (Parnaudeau et al. 2013) in animals,and have been shown to modulate neocortical synchrony basedon attentional demands, at least in the visual domain, consistentwith a putative role in cortical information distribution (Saal-mann et al. 2012).

In humans, recent fMRI evidence shows differential thalamicnuclei activate during learning and retrieval phases on associa-tive learning tasks in healthy volunteers (Pergola et al. 2013),while relational memory deficits are reported in patients withcertain thalamic lesions (Soei et al. 2008). Further studies areclearly needed to refine categories of computations and/or stimu-lus features fundamentally reliant on the hippocampus, howprocessing demands alter the requirements for STM versusLTM, and the extent to which this balance may hinge upon thal-amic modulation of activity within wider neural networks sup-porting memory function. In this respect, patients with focalthalamic lesions might provide additional unique insight intoconsequences of lesions disrupting distinct thalamo-corticalmemory networks on hippocampal processes.

Although our correlational and regression results support in-dependent contributions of specific thalamo-prefrontal cortexpathway disruption to STM deficits, such evidence does not, ofcourse, inform causal directions or the etiology of thalamo-cor-tical disruption. Secondary degeneration following hippocampalatrophy should preferentially affect the anterior thalamic nu-cleus, whereas postmortem (Sinjab et al. 2013) and imagingdata (Barron et al. 2012; Bernhardt et al. 2012; Duzel et al. 2006)identify primarily mediodorsal, pulvinar, and lateral nucleardamage. Since midline thalamic nuclei have been implicated inseizure spread in animal models of temporal lobe (Bertramet al. 2001), but also generalized, epilepsy (Avoli and Gloor1982), hippocampal atrophy alone may not fully account for tha-lamo-cortical disruption in our patients. Furthermore, subtledamage in our normal volume subgroup cannot be excludedwithout histopathology. Longitudinal and developmental studieswill be necessary to shed further light onto the mechanisms as-sociated with thalamic dysfunction in this epileptic model ofMTL damage.

We defined hippocampal atrophy as volumes falling 2 standarddeviations below those of age-matched controls. To determine theinfluence of this classification, we repeated our patient subgroupanalyses with a more liberal cutoff of hippocampal volumes <1.5SD from those of controls. According to this revised classification,1 patientwas reclassified to the “atrophic” group. Aside fromslight-ly reduced statistical significance for correlations between imagingmarkers and digit span performance in the downsized “normalrange” group, defining atrophy using a 1.5 SD cutoff did not alterthe overall pattern of findings in the subgroup comparisons.

Certain uncontrollable clinical variables potentially contributeto our findings. All patients were taking different combinations ofantiepileptic drugs. The extent to which these influence fMRI sig-nals is not knownand could not be assessed in ourmodest samplesize. Additionally, some antiepileptic medications affect levels ofmotivation, arousal, and/or attention that could impact on cogni-tive performance, although their exact mechanism of action re-mains poorly understood. It seems unlikely that medicationalone selectively influenced thalamo-temporal versus thalamo-prefrontal circuits in a way that might produce the dissociatedanatomical-behavioral relationships we observed, especiallygiven the range of drugs anddoses takenbypatients in this cohort.

In conclusion, patients with epilepsy-related MTL dysfunctionshow thalamic atrophy co-varying with extents of hippocampal

damage. Pathological extents of individual subcortical structuresalone are not sensitive markers of performance, highlighting theimportance of widespread functional circuits to STM and LTM. Ex-tending previous reports linking global thalamic volumes withmemory deficits in patients with MTL dysfunction (Seidenberget al. 2008; Stewart et al. 2009), these findings demonstrate thatspatially distinct thalamo-cortical functional networks are asso-ciated with STM and LTM deficits and are specifically affected inpatients with hippocampal atrophy rather than those with pre-served hippocampal volumes. We propose that beyond the directimpact of MTL lesions, associated effects on specific thalamo-cor-tical circuits might play an important role in the etiology of STMdeficits and may aid to reconcile previous disparate findings inMTL-damaged patients.

FundingThis work was supported by a Vera Down grant from the BritishMedical Association. We gratefully acknowledge personal fund-ing support from the Medical Research Council, and the NationalInstitutes of Health Research Oxford Biomedical Research Centreto N.L.V.

NotesWe thank Drs Giovanna Zamboni and David Cole for helpful feed-back on a draft of thismanuscript.Conflict of Interest: None declared.

ReferencesAggleton JP. 2012. Multiple anatomical systems embeddedwithin

the primate medial temporal lobe: implications for hippo-campal function. Neurosci Biobehav Rev. 36(7):1579–1596.

Aggleton JP, Desimone R,MishkinM. 1986. The origin, course, andtermination of the hippocampothalamic projections in themacaque. J Comp Neurol. 243(3):409–421.

Aggleton JP, Mishkin M. 1983. Memory impairments following re-strictedmedial thalamic lesions inmonkeys. Exp Brain Res. 52(2):199–209.

Aggleton JP, O’Mara SM, Vann SD, Wright NF, Tsanov M,Erichsen JT. 2010. Hippocampal-anterior thalamic pathwaysfor memory: uncovering a network of direct and indirect ac-tions. Eur J Neurosci. 31(12):2292–2307.

Alvarez P, Zola-Morgan S, Squire LR. 1994. The animal model ofhuman amnesia: long-term memory impaired and short-termmemory intact. Proc Natl Acad Sci USA. 91(12):5637–5641.

Avoli M, Gloor P. 1982. Interaction of cortex and thalamus in spikeand wave discharges of feline generalized penicillin epilepsy.Exp Neurol. 76(1):196–217.

Baddeley A, Allen R, Vargha-Khadem F. 2010. Is the hippocampusnecessary for visual and verbal binding in working memory?Neuropsychologia. 48(4):1089–1095.

Baddeley A, Jarrold C, Vargha-Khadem F. 2011. Working memoryand the hippocampus. J Cogn Neurosci. 23(12):3855–3861.

Barron DS, Fox PM, Laird AR, Robinson JL, Fox PT. 2012. Thalamicmedial dorsal nucleus atrophy in medial temporal lobe epi-lepsy: a VBM meta-analysis. Neuroimage Clin. 2:25–32.

Baxendale S, Thompson PJ, Duncan JS. 2012. Neuropsychologicalfunction in patients who have had epilepsy surgery: a long-term follow-up. Epilepsy Behav. 23(1):24–29.

Behrens TE, Johansen-Berg H,WoolrichMW, Smith SM,Wheeler-Kingshott CA, Boulby PA, Barker GJ, Sillery EL, Sheehan K,Ciccarelli O, et al. 2003. Non-invasivemapping of connections


between human thalamus and cortex using diffusion im-aging. Nat Neurosci. 6(7):750–757.

Bentivoglio M, Aggleton JP, Mishkin M. 1997. The thalamus andmemory formation. In: Steriade M, Jones EJ, editors.Thalamus, Vol. 2: experimental and clinical aspects. Amster-dam: Elsevier. p. 689–721.

Bernhardt BC, Bernasconi N, KimH, Bernasconi A. 2012. Mappingthalamocortical network pathology in temporal lobe epilepsy.Neurology. 78(2):129–136.

Bertram EH, Mangan PS, Zhang D, Scott CA, Williamson JM. 2001.Themidline thalamus: alterations and a potential role in lim-bic epilepsy. Epilepsia. 42(8):967–978.

Callicott JH, Mattay VS, Bertolino A, Finn K, Coppola R, Frank JA,Goldberg TE, Weinberger DR. 1999. Physiological characteris-tics of capacity constraints in working memory as revealedby functional MRI. Cereb Cortex. 9(1):20–26.

Cashdollar N, Duncan JS, Duzel E. 2011. Challenging the classicaldistinction between long-term and short-term memory: re-considering the role of the hippocampus. Future Neurol. 6(3):351–362.

Chein JM, Moore AB, Conway AR. 2011. Domain-general mechan-isms of complex working memory span. Neuroimage. 54(1):550–559.

Concha L, Beaulieu C, Collins DL, Gross DW. 2009. White-matterdiffusion abnormalities in temporal-lobe epilepsy with andwithout mesial temporal sclerosis. J Neurol NeurosurgPsychiatry. 80(3):312–319.

Corkin S. 1984. Lasting consequences of bilateralmedial temporallobectomy - clinical course and experimental findings in Hm.Seminars Neurol. 4(2):249–259.

Dagenbach D, Kubat-Silman AK, Absher JR. 2001. Human verbalworking memory impairments associated with thalamicdamage. Int J Neurosci. 111(1–2):67–87.

Dale AM, Fischl B, Sereno MI. 1999. Cortical surface-basedanalysis. I. Segmentation and surface reconstruction.Neuroimage. 9(2):179–194.

de Bourbon-Teles J, Bentley P, Koshino S, Shah K, Dutta A,Malhotra P, Egner T, Husain M, Soto D. 2014. Thalamic controlof human attention drivenbymemory and learning. Curr Biol.24(9):993–999.

Desikan RS, Segonne F, Fischl B, Quinn BT, Dickerson BC,Blacker D, Buckner RL, Dale AM, Maguire RP, Hyman BT,et al. 2006. An automated labeling system for subdividingthe human cerebral cortex on MRI scans into gyral based re-gions of interest. Neuroimage. 31(3):968–980.

Drachman DA, Arbit J. 1966. Memory and the hippocampal com-plex. II. Is memory a multiple process? Arch Neurol. 15(1):52–61.

Duzel E, Schiltz K, Solbach T, Peschel T, Baldeweg T, Kaufmann J,Szentkuti A, Heinze HJ. 2006. Hippocampal atrophy in tem-poral lobe epilepsy is correlated with limbic systems atrophy.J Neurol. 253(3):294–300.

Falconer MA, Serafetinides EA, Corsellis JA. 1964. Etiology andpathogenesis of temporal lobe epilepsy. Arch Neurol.10:233–248.

Fischl B, Sereno MI, Dale AM. 1999. Cortical surface-based ana-lysis. II: Inflation, flattening, and a surface-based coordinatesystem. Neuroimage. 9(2):195–207.

Focke NK, Yogarajah M, Bonelli SB, Bartlett PA, Symms MR,Duncan JS. 2008. Voxel-based diffusion tensor imaging in pa-tients with mesial temporal lobe epilepsy and hippocampalsclerosis. Neuroimage. 40(2):728–737.

Fuster JM, Alexander GE. 1971. Neuron activity related to short-term memory. Science. 173(3997):652–654.

Gaskin S, White NM. 2013. Parallel processing of informationabout location in the amygdala, entorhinal cortex and hippo-campus. Hippocampus. 23(11):1075–1083.

Gibb SJ, Wolff M, Dalrymple-Alford JC. 2006. Odour-place paired-associate learning and limbic thalamus: comparison of anter-ior, lateral and medial thalamic lesions. Behav Brain Res. 172(1):155–168.

Goldman-Rakic PS. 1988. Topography of cognition: parallel dis-tributed networks in primate association cortex. Annu RevNeurosci. 11:137–156.

Goldman-Rakic PS, Selemon LD, Schwartz ML. 1984. Dual path-ways connecting the dorsolateral prefrontal cortex with thehippocampal formation and parahippocampal cortex in therhesus monkey. Neuroscience. 12(3):719–743.

Greve DN, Fischl B. 2009. Accurate and robust brain image align-ment using boundary-based registration. Neuroimage.48(1):63–72.

Hartley T, Bird CM, Chan D, Cipolotti L, Husain M, Vargha-Khadem F, Burgess N. 2007. The hippocampus is required forshort-term topographical memory in humans. Hippocampus.17(1):34–48.

Henson RN, Burgess N, Frith CD. 2000. Recoding, storage, re-hearsal and grouping in verbal short-term memory: an fMRIstudy. Neuropsychologia. 38(4):426–440.

Hunt PR, Aggleton JP. 1991. Medial dorsal thalamic lesions andworking memory in the rat. Behav Neural Biol. 55(2):227–246.

Igarashi KM, Lu L, Colgin LL, Moser MB, Moser EI. 2014. Coordin-ation of entorhinal-hippocampal ensemble activity during as-sociative learning. Nature. 510(7503):143–147.

Isseroff A, Rosvold HE, Galkin TW, Goldman-Rakic PS. 1982. Spatialmemory impairments following damage to the mediodorsal nu-cleusof the thalamus in rhesusmonkeys.BrainRes. 232(1):97–113.

Jeneson A, Mauldin KN, Squire LR. 2010. Intact working memoryfor relational information aftermedial temporal lobe damage.J Neurosci. 30(41):13624–13629.

JenesonA, Squire LR. 2012.Workingmemory, long-termmemory,and medial temporal lobe function. Learn Mem. 19(1):15–25.

Keller SS, O’Muircheartaigh J, Traynor C, Towgood K, Barker GJ,Richardson MP. 2014. Thalamotemporal impairment in tem-poral lobe epilepsy: a combined MRI analysis of structure, in-tegrity, and connectivity. Epilepsia. 55(2):306–315.

Kubat-Silman AK, Dagenbach D, Absher JR. 2002. Patterns of im-paired verbal, spatial, and object working memory after thal-amic lesions. Brain Cogn. 50(2):178–193.

Menke RA, Szewczyk-Krolikowski K, Jbabdi S, Jenkinson M,Talbot K, Mackay CE, Hu M. 2014. Comprehensive morphom-etry of subcortical grey matter structures in early-stage Par-kinson’s disease. Hum Brain Mapp. 35(4):1681–1690.

Milner B. 1971. Interhemispheric differences in the localization ofpsychological processes in man. Br Med Bull. 27(3):272–277.

MishkinM. 1982. Amemory system in themonkey. Philos Trans RSoc Lond B Biol Sci. 298(1089):83–95.

Mitchell AS, Chakraborty S. 2013. What does the mediodorsalthalamus do? Front Syst Neurosci. 7(37):1–19.

Nichols EA, KaoYC, VerfaellieM, Gabrieli JD. 2006.Workingmem-ory and long-term memory for faces: evidence from fMRI andglobal amnesia for involvement of themedial temporal lobes.Hippocampus. 16(7):604–616.

Nystrom LE, Braver TS, Sabb FW, Delgado MR, Noll DC, Cohen JD.2000.Workingmemory for letters, shapes, and locations: fMRIevidence against stimulus-based regional organization inhuman prefrontal cortex. Neuroimage. 11(5 Pt 1):424–446.

O’Reilly JX, Beckmann CF, Tomassini V, Ramnani N, Johansen-Berg H. 2010. Distinct and overlapping functional zones in


the cerebellum defined by resting state functional connectiv-ity. Cereb Cortex. 20(4):953–965.

Ojemann G, Fedio P. 1968. Effect of stimulation of the humanthalamus and parietal and temporal white matter on short-term memory. J Neurosurg. 29(1):51–59.

Ojemann GA, Blick KI, Ward AA Jr. 1971. Improvement and dis-turbance of short-term verbal memory with human ventro-lateral thalamic stimulation. Brain. 94(2):225–240.

Olson IR, Moore KS, Stark M, Chatterjee A. 2006. Visual workingmemory is impaired when the medial temporal lobe is da-maged. J Cogn Neurosci. 18(7):1087–1097.

Olson IR, Page K, Moore KS, Chatterjee A, Verfaellie M. 2006.Working memory for conjunctions relies on the medial tem-poral lobe. J Neurosci. 26(17):4596–4601.

Owen AM, Herrod NJ, Menon DK, Clark JC, Downey SP,Carpenter TA, Minhas PS, Turkheimer FE, Williams EJ,Robbins TW, et al. 1999. Redefining the functional organiza-tion of working memory processes within human lateral pre-frontal cortex. Eur J Neurosci. 11(2):567–574.

Oztekin I, Davachi L, McElree B. 2010. Are representations inworking memory distinct from representations in long-termmemory? Neural evidence in support of a single store.Psychol Sci. 21(8):1123–1133.

Pardoe HR, Pell GS, Abbott DF, Jackson GD. 2009. Hippocampalvolume assessment in temporal lobe epilepsy: how good isautomated segmentation? Epilepsia. 50(12):2586–2592.

Parnaudeau S, O’Neill PK, Bolkan SS, Ward RD, Abbas AI, Roth BL,Balsam PD, Gordon JA, Kellendonk C. 2013. Inhibition of med-iodorsal thalamus disrupts thalamofrontal connectivity andcognition. Neuron. 77(6):1151–1162.

Patenaude B, Smith SM, KennedyDN, JenkinsonM. 2011. A Bayes-ian model of shape and appearance for subcortical brain seg-mentation. Neuroimage. 56(3):907–922.

Pergola G, Ranft A,Mathias K, Suchan B. 2013. The role of the thal-amic nuclei in recognition memory accompanied by recallduring encoding and retrieval: an fMRI study. Neuroimage.74:195–208.

Pertzov Y, Miller TD, Gorgoraptis N, Caine D, Schott JM, Butler C,Husain M. 2013. Binding deficits in memory following medialtemporal lobe damage in patients with voltage-gated potas-sium channel complex antibody-associated limbic encephal-itis. Brain. 136(Pt 8):2474–2485.

Powell HW, RichardsonMP, SymmsMR, Boulby PA, Thompson PJ,Duncan JS, Koepp MJ. 2008. Preoperative fMRI predicts mem-ory decline following anterior temporal lobe resection.J Neurol Neurosurg Psychiatry. 79(6):686–693.

Price GR, Mazzocco MM, Ansari D. 2013. Why mental arithmeticcounts: brain activation during single digit arithmetic predictshigh school math scores. J Neurosci. 33(1):156–163.

Ranganath C, D’Esposito M. 2001. Medial temporal lobe activityassociated with active maintenance of novel information.Neuron. 31(5):865–873.

Romanski LM, GiguereM, Bates JF, Goldman-Rakic PS. 1997. Topo-graphic organization of medial pulvinar connections with theprefrontal cortex in the rhesus monkey. J Comp Neurol.379(3):313–332.

Rotshtein P, Soto D, Grecucci A, Geng JJ, Humphreys GW. 2011.The role of the pulvinar in resolving competition betweenmemory and visual selection: a functional connectivitystudy. Neuropsychologia. 49(6):1544–1552.

Saalmann YB, Pinsk MA, Wang L, Li X, Kastner S. 2012. The pulvi-nar regulates information transmission between cortical areasbased on attention demands. Science. 337(6095):753–756.

Seidenberg M, Hermann B, Pulsipher D, Morton J, Parrish J,Geary E, Guidotti L. 2008. Thalamic atrophy and cognition inunilateral temporal lobe epilepsy. J Int Neuropsychol Soc.14(3):384–393.

Shallice T, Warrington EK. 1970. Independent functioning of ver-bal memory stores: a neuropsychological study. Q J ExpPsychol. 22(2):261–273.

Sherman SM, Guillery RW. 2011. Distinct functions for direct andtransthalamic corticocortical connections. J Neurophysiol.106(3):1068–1077.

Sinjab B, Martinian L, Sisodiya SM, Thom M. 2013. Regional thal-amic neuropathology in patients with hippocampal sclerosisand epilepsy: a postmortem study. Epilepsia. 54(12):2125–2133.

Soei E, Koch B, Schwarz M, Daum I. 2008. Involvement of thehuman thalamus in relational and non-relational memory.Eur J Neurosci. 28(12):2533–2541.

Stewart CC, Griffith HR, Okonkwo OC, Martin RC, Knowlton RK,Richardson EJ, Hermann BP, Seidenberg M. 2009. Contribu-tions of volumetrics of the hippocampus and thalamus to ver-bal memory in temporal lobe epilepsy patients. Brain Cogn.69(1):65–72.

Sziklas V, Petrides M. 1999. The effects of lesions to the anteriorthalamic nuclei on object-place associations in rats. Eur JNeurosci. 11(2):559–566.

Tanibuchi I, Goldman-Rakic PS. 2003. Dissociation of spatial-, ob-ject-, and sound-coding neurons in the mediodorsal nucleusof the primate thalamus. J Neurophysiol. 89(2):1067–1077.

Todd JJ, Marois R. 2004. Capacity limit of visual short-termmem-ory in human posterior parietal cortex. Nature. 428(6984):751–754.

Van derWerf YD, Scheltens P, Lindeboom J,WitterMP, Uylings HB,Jolles J. 2003. Deficits of memory, executive functioningand attention following infarction in the thalamus; a studyof 22 cases with localised lesions. Neuropsychologia.41(10):1330–1344.

Vertes RP, Hoover WB, Szigeti-Buck K, Leranth C. 2007. Nucleusreuniens of the midline thalamus: link between the medialprefrontal cortex and the hippocampus. Brain Res Bull.71(6):601–609.

Voets NL, Zamboni G, Stokes MG, Carpenter K, Stacey R,Adcock JE. 2014. Aberrant functional connectivity in dissoci-able hippocampal networks is associated with deficits inmemory. J Neurosci. 34(14):4920–4928.

Wager TD, Smith EE. 2003. Neuroimaging studies of workingmemory: a meta-analysis. Cogn Affect Behav Neurosci.3(4):255–274.

Wickelgren WA. 1968. Sparing of short-term memory in an am-nesic patient - implications for strength theory of memory.Neuropsychologia. 6(3):235.

Winocur G. 1985. The hippocampus and thalamus: their roles inshort- and long-termmemory and the effects of interference.Behav Brain Res. 16(2–3):135–152.

Yang T, Bavley RL, Fomalont K, Blomstrom KJ, Mitz AR, Turchi J,Rudebeck PH, Murray EA. 2014. Contributions of the hippo-campus and entorhinal cortex to rapid visuomotor learningin rhesus monkeys. Hippocampus. 24(9):1102–1111.

Yonelinas AP. 2013. The hippocampus supports high-resolutionbinding in the service of perception, working memory andlong-term memory. Behav Brain Res. 254:34–44.

Zarahn E, Rakitin B, Abela D, Flynn J, Stern Y. 2005. Positive evi-dence against human hippocampal involvement in workingmemory maintenance of familiar stimuli. Cereb Cortex.15(3):303–316.


thalamo-cortical disruption contributesto short-term · 2019. 1. 20. · for this,...

Documents