This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Morphometric abnormalities and hyperanxiety in genetically epileptic rats: A modelof psychiatric comorbidity?

Viviane Bouilleret a,b, R. Edward Hogan c, Dennis Velakoulis d, Michael R. Salzberg e, Lei Wang f, Gary F. Egan g,Terence J. O'Brien a,h, Nigel C. Jones a,⁎a Department of Medicine — Royal Melbourne Hospital, University of Melbourne, Royal Parade, Parkville, Victoria 3052, Australiab Department of Neurophysiology and Epilepsy, ApHp, CHU Bicetre, Paris 94275, Francec Department of Neurology, Washington University, St. Louis, MO 63130, USAd Melbourne Neuropsychiatry Centre, Royal Melbourne Hospital, University of Melbourne, Victoria 3052, Australiae Department of Psychiatry, St. Vincent's Health, The University of Melbourne, Victoria 3052, Australiaf Department of Psychiatry and Behavioral Sciences, Northwestern University Feinberg School of Medicine, Chicago, IL 60611, USAg Howard Florey Institute, Parkville, Victoria 3050, Australiah Department of Neurology (RMH), University of Melbourne, Victoria 3052, Australia

a b s t r a c ta r t i c l e i n f o

Article history:Received 17 September 2008Revised 30 November 2008Accepted 8 December 2008Available online 25 December 2008

Keywords:Absence epilepsyPsychopathologyGAERSMRIAnxietyHippocampus

Background: Imaging studies of epilepsy patients with comorbid affective disturbance demonstratemorphometric changes in limbic brain regions implicated in psychiatric disease. Genetic Absence EpilepsyRats from Strasbourg (GAERS), specifically bred for their epilepsy phenotype, also exhibit elevated anxiety-like behaviors suggesting a common causality. Here we examined whether relevant cerebral morphologicalalterations exist in this rat strain using volumetric measurements and large deformation high dimensionalmapping (HDM-LD), a tool recently validated to produce accurate three-dimensional surface representationsof the hippocampus.Methods: Volumetric MRI and the Open Field test of anxiety were performed in adult female GAERS (n=12)and Non-Epileptic Controls (NEC; n=11). The volumes of selected brain regions, including cortex,hippocampus, amygdala, thalamus, hypothalamus and lateral ventricles, were measured using Region-Of-Interest analysis from the MRI data and total volumes compared between the two strains.Results: GAERS had increased amygdala (right: p=0.003; left pb0.001), cortices (right: p=0.006; leftp=0.012) and ventricular volumes (p=0.002) when compared with NEC rats. Further, HDM-LD showedGAERS to have hippocampal volume loss in two regions: the medial hippocampal surface immediately caudalto the hippocampal commissure, and the lateral hippocampal surface over the mid-portion of theseptotemporal axis. GAERS exhibited increased anxiety in the Open Field compared with NEC rats: reduceddistance traveled (pb0.001) and reduced time in the centre area (p=0.042).Conclusions: Morphometric brain changes in GAERS could be relevant to their hyperanxious and epilepticphenotypes. This model may be useful in illuminating the pathogenesis of affective disorders generally, aswell as modeling psychiatric comorbidities of epilepsy.

© 2008 Elsevier Inc. All rights reserved.

Introduction

A high proportion of patients with epilepsy exhibit psychiatric co-morbidities which contribute greatly to impaired quality of life(Hermann et al., 2008). Although this was previously particularlylinked to temporal lobe epilepsy, more recent studies have demon-strated that patients with other focal and generalized epilepsy

syndromes suffer psychiatric co-morbidities at least as frequently(Adams et al., 2008; Christensen et al., 2007; Hermann et al., 2008;Jones et al., 2007; Ott et al., 2003; Tellez-Zenteno et al., 2007). Theincreased prevalence of these psychiatric disturbances, includingdepression, anxiety disorders, psychoses, cognitive disorders andincreased suicide ideation and attempts, may be attributed to thepsychosocial consequences of living with epilepsy, the effect of long-term drug therapy, or to a common underlying neurobiological insult.However several studies, examining both acquired and idiopathicepilepsies, have identified the presence of mood disturbance prior tothe onset of seizures (Forsgren and Nystrom, 1990; Hesdorffer et al.,2000, 2006; Jones et al., 2007), intimating that mood disorders pre-

NeuroImage 45 (2009) 267–274

⁎ Corresponding author. Fax: +61 3 9347 1863.E-mail address: ncjones@unimelb.edu.au (N.C. Jones).

1053-8119/$ – see front matter © 2008 Elsevier Inc. All rights reserved.doi:10.1016/j.neuroimage.2008.12.019

Contents lists available at ScienceDirect

NeuroImage

j ourna l homepage: www.e lsev ie r.com/ locate /yn img

Author's personal copy

date the epilepsy and suggesting that psychiatric disorders present arisk factor for epilepsy development, or that there is commoncausality (Salzberg et al., 2007).

While neuroimaging studies have been used extensively in anattempt to identify structural abnormalities in patients with affectivedisorders, such as major depression and bipolar disease, particularlyfocusing on limbic structures such as hippocampus and amygdala (forreviews, see Brambilla et al., 2002; Campbell and MacQueen, 2006),only a limited number of studies have investigated affectivedisturbance in patients with comorbid epilepsy (Briellmann et al.,2000; Maier et al., 2000; Tebartz van Elst et al., 2002). Two of thesestudies had small sample sizes (Briellmann et al., 2000; Maier et al.,2000), but in a study of 70 patients enlargement of amygdala volumewas identified in temporal lobe epilepsy (TLE) patients experiencingpsychoses but not in TLE patients without psychosis (Tebartz van Elstet al., 2002). This morphological regional change has also been ob-served in TLE patients with dysthymia (i.e.: chronic mild depression—

Tebartz van Elst et al., 1999, 2000) and TLE with anxiety disorders(Satishchandra et al., 2003). A recent study identified enlarged leftamygdala volumes in pediatric patients with complex partial seizuresand affective disturbance, compared to epilepsy patients without anypsychopathology (Daley et al., 2008). Taken together with the strongassociation of the amygdala with emotion (Rosen and Donley, 2006),these imaging findings suggest that changes in amygdala volume mayplay a role in the pathogenesis of affective disturbances in patientswith epilepsy.

Regional hippocampal volume changes have been observed inpatients with epilepsy (Hogan et al., 2003), schizophrenia (Cser-nansky et al., 1998) and depression (Posener et al., 2003). However,some of the changes found in these studies were discrete and onlydetected with volumetric shape analysis using MRI combined withlarge deformation high dimensional mapping. The strong relationshipbetween this limbic structure and both epilepsy (Hogan et al., 2003)and schizophrenia/psychosis (Csernansky et al., 1998; Nelson et al.,1998) suggests that it too may be primarily involved in thepathogenesis of these comorbid disorders.

GAERS (Genetic Absence Epilepsy Rats from Strasbourg) are awell-validated animal model of human Idiopathic GeneralisedEpilepsy (IGE), possessing a similar electrophysiological, ontogenicand pharmacological profile to the human condition (Danober et al.,1998). Historically, the GAERS colony was derived from a Wistar ratline and selectively bred for the seizure phenotype so that 100% ofprogeny spontaneously develop the epilepsy. A non-epilepticcontrol (NEC) strain was also derived from the original colony byselectively breeding for the lack of seizure expression, providing apowerful control strain, since any differences between the twostrains would have a high a priori chance of being aetiologicallyassociated with the epilepsy. In addition to developing epilepsy, theGAERS strain also exhibits a range of behaviors indicative ofaffective disturbance, including increased anxiety- and depressive-like behaviors (Jones et al., 2008), intimating that this rat strain alsomodels the well-documented mood disturbances observed in clinicalIGE populations (Caplan et al., 1998, 2005; Davies et al., 2003; Joneset al., 2007; Ott et al., 2003; Tellez-Zenteno et al., 2007). To ourknowledge, however, no investigation of brain morphology has beenundertaken comparing GAERS and NEC rats to determine whethermorphological abnormalities exist which might explain the observedpathologies in GAERS.

In-vivo imaging with MRI has great advantages in validity overtraditional postmortem studies, avoiding distortions inherent in thesemethods as a result of death and brain processing, for example fixationmay elicit inhomogeneous retractions of the tissues; whereas freshremoval of tissue may be modified by gravity when the brain removedfrom the skull. Using advanced in vivo MRI incorporating HDM-LD ofthe hippocampus, the aim of the current study was to determinewhether the brain regions implicated in seizure expression and

affective disorders differ between GAERS and NEC rats. Furthermore,we aimed to clarify if any such volumetric differences correlated withthe extent of the behavioral disturbance.

Methods and materials

Animals

Age-matched adult female GAERS (n=12) or Non-Epileptic Control(NEC) rats (n=11), ∼50th generation from parents originally obtainedfrom the Strasbourg population, were used in all experiments, andbred and group-housed (2–3 per cage) in the Ludwig Institute ofCancer Research/Department of Surgery Royal Melbourne Hospitalbiological research facility. The animal facilities were on a 12 h light/dark cycle, with lights on at 6am, and food (standard rat chow) andwater was available ad libitum in standard rat cages with sawdustbedding and shredded paper. Subjects sequentially underwentbehavioral testing and MRI at ∼14 weeks of age. At this age, allGAERS from our colony experience absence-like Spike-Wave-Dis-charges on the EEG (Jones et al., 2008), so GAERS were considered tobe epileptic at this time point. At all times, care was taken tominimizepain and discomfort of the animals, and all experimental procedureswere approved by the Melbourne Health and Howard Florey AnimalEthics Committees.

Magnetic resonance imaging

All Magnetic Resonance Images (MRI) were acquired on a 4.7 TBruker Biospec 47/30 Avance small animal spectrometer (Ettlingen,Germany) using a shielded gradient set (Bruker Biospec) appropriatefor rats at the Howard Florey Institute. T2-weighted axial structuralimages were obtained contiguously through the entire brain, using afast spin-echo sequence (TA, 298 s; TR, 3.1 s; TE, 67.5 ms; MTX,256 × 256; NA, 3; FOV, 60 mm×60 mm; voxel dimensions0.23 mm×0.23 mm×0.50 mm; rare factor=8). Animals were scannedin the supine position in a custom built plexiglass holder to ensureconsistent positioning of the animal within the coil. Anaesthesia wasinducedwith 5% Isoflurane in 1:1 air and oxygen (flow rates of 200ml/min), then maintained with 1.5–2.5% Isoflurane for the remainder ofthe experiment. RF pulse transmission and MR data acquisition wasperformed using a 116 mm inner diameter birdcage coil (BrukerBiospec) optimally tuned. The following sequences were acquiredduring the duration of the scan: a tri-pilot image for confirmation ofthe animal's position within the magnet, and then coronal, sagittaland axial T2 weighted images. The total scanning procedure was∼60 min.

Regional brain morphometric analysis

Quantification of the volume of selected brain regions wasperformed using Analyze 8.1 (Mayo Clinic, MN) in which Regions-Of-Interest (ROI) weremanually drawn on consecutive axial MRI slicesby an investigator blinded to the strain of animal. Twelve ROIs on amean of 35 sections per animal were drawn (Fig. 1). All delineations ofthe ROIs are described in an rostral–caudal orientation, and were asfollows: the cortex (Ctx) encompassed the 26 slices after the olfactorybulb; the striatal region (St, 7 slices) extended from the appearance ofthe corpus callosum to the thalamus; the thalamus (Tha, 7 slices)extended from the appearance of the hippocampus (HC, 8 slices) tothe first slice when the ventral hippocampus appeared; the amygdalaregion (Amyg) was delineated fromwhere the hippocampus appearedto the caudal point of the thalamus (consisting of 8 slices), and wasdefined as the area under a horizontal line passing below thethalamus; the hypothalamus (Hypotha, 4 slices) was delineated oneslice rostral to the thalamus and 3 slices caudally; and finally theventricles, incorporating both the lateral and fourth ventricles (Vent,

268 V. Bouilleret et al. / NeuroImage 45 (2009) 267–274

Author's personal copy

20 slices), started with the striatal region and progressed to the end ofthe cortex. Right (R) and left (L) sides weremeasured separately for allregions except the ventricles and the hypothalamus. Additionally, thewhole brain was drawn separately on the same slices as where theother ROI were drawn using the auto-trace tool of Analyze 8.1 basedon the intensity visually adjusted to the boundary of the brain.Calculation of the volume of each brain region was computed bymultiplying the number of voxels traced in each slice by their depth(i.e.: slice thickness) and by the number of slices.

The somatosensory cortex is thought to represent the site ofseizure initiation in GAERS (Pinault, 2003), and we thereforeestimated the thickness of the cortex at this region. Since cranialand caudal boundaries of the somatosensory cortex are difficult todelineate on theMRI images, wemeasured the somatosensory corticalthickness rather than the total volume. This was measured on 5contiguous coronal sections, starting 3 slices posterior to theappearance of the Forceps Minor of the corpus callosum (FMI) bythe following procedure: a line representing the shortest distancebetween the cortical surface and the point of intersection with ahorizontal line from the level of the top of the lateral ventricles andthe exterior of the corpus callosum (Fig.1). The somatosensory corticalthickness measurements of the 5 slices were then averaged for bothleft and right hemispheres.

Large deformation high dimensional mapping of the hippocampus

In both humans and animals, most investigators have used manualsegmentation of the hippocampus on MRI to determine hippocampalvolumes (Grohn and Pitkanen, 2007; Jupp et al., 2006; Watson et al.,1997). Large deformation high dimensional mapping (HDM-LD)utilizes the power of computer-assisted shape recognition to identifygeneral patterns within image data. Computational anatomic techni-ques produce three-dimensional surface representations of thehippocampus with resolution at a subvoxel level, enabling visualiza-tion of details of hippocampal surface anatomy (Gardner and Hogan,2005), and can show significant changes in hippocampal surfaceanatomy when total hippocampal volume changes are non-signifi-cantly different (Csernansky et al., 1998; Posener et al., 2003). HDM-LD hippocampal surface analysis shows specific patterns of hippo-campal surface changes in epilepsy (Hogan et al., 2004, 2008),schizophrenia (Csernansky et al., 1998) and depression (Posener et al.,2003), which illustrates the role of HDM-LD in a range of neuropsy-chiatric diseases. This method has been recently shown to be a validand reliable method for the segmentation of the rat hippocampus(Hogan et al., in Press).

We performed deformation segmentations as previously describedusing propriety linux-based software (Hogan et al., in Press). Briefly,the first step involved a blinded operator placing both global andhippocampus specific landmarks. These landmarks provided an initialcondition for the intensity-matching algorithm by roughly aligningthe rat atlas and scans. The global landmarks enabled scaling andalignment of the atlas brain to the target brain, relying principally onalignment of the boarders of the hemispheres. The hippocampalspecific landmarking first required identification of the septal andtemporal pole of each hippocampus (Witter and Amaral, 2004), whichspecified an axis for frame of reference for landmarking of eachhippocampus. Then, four landmarks (the medial, lateral, superior andinferior boarder of the hippocampus) were identified on five crosssections equally spaced along this axis. Images and landmarking datawere then integrated into another linux-based software program.Within this program, the mapping algorithm employed a coarse-to-fine procedure for generating a transformation field from an atlasreference MR to the target MR. The “coarse” aspect of the procedurerelied on the landmark information to derive a coarse manifoldtransformation (Joshi et al., 1995) from the atlas to the target images.The “fine” procedure involved two steps: the first was to solve theregistration problem using a linear elastic basis formulation and thefull volume data, as previously described (Christensen et al., 1994;Miller et al., 1993). This was fully automatic and only driven by thevolume data itself. The three dimensional whole brain mapscorresponded to the maximizer, whose variation solution corre-sponded to a solution of a non-linear PDE, consisting of between 107–108 parameters. The second and final step of the algorithm was tosolve the non-linear PDE corresponding to the Bayesian maximizerassociated with the fluid formulation at each voxel of the full volume(Christensen et al., 1993, 1994, 1995).

The deformation segmentation procedure resulted in generation ofcoordinates for transformation of a normal template hippocampalsegmentation (from a single normal rat, otherwise not included in thestudy) into the shape of hippocampi of the GAERS and NEC rats. Wetherefore generated coordinates for transformation for each rat in thecurrent study. To generate an ‘average’ hippocampus, we generated amean transformation for the deformation images in each group. Thismean transformation was then applied to the atlas itself to generatethe ‘average’ hippocampi for each group (Hogan et al., 2003).

To further quantify the difference between the NEC and GAERShippocampi, we calculated a Minimum Mean Squared Error (MMSE)estimation by coregistering the ‘average’ hippocampal surfaces of ratgroups (Hogan et al., 2003). By calculating differences in hippocampaltransformation coregistration using the MMSE, we made a direct

Fig. 1. Regions-of-Interests (ROIs) were drawn by a blinded reviewer on the T2-weighted axial MRI images. The regions depicted are the cortical structures (Ctx, right (R): cyan, left(L): white), the limbic structures of the amygdala (Amyg, R: red, L: green) and the hippocampi (HC, R: beige, L: salmon) and other structures of interest, including the striatum (St, R:yellow, L: blue), the hypothalamus (Hypotha: tan), the thalami (Tha, R: violet, L: deep pink) and the ventricles (Vent: magenta). The solid line within the left cortex represents thesomatosensory cortex thickness which was defined as the shortest distance between the cortical surface and a point where a horizontal line originating from the top of the lateralventricles intersects the exterior of the corpus callosum.

269V. Bouilleret et al. / NeuroImage 45 (2009) 267–274

Author's personal copy

comparison between the shape of the hippocampi in the ipsilateralcerebral hemisphere between the NEC and GAERS groups. Final resultswere projected on the NEC group hippocampal surfaces. To defineshape changes in the hippocampus, we assigned a colormap, whichrepresents the distance between the matched hippocampi. Themaximum and minimum values of the colormap were set using themaximum value of inward deformation. Deformation patterns werevisually assessed, comparing differences between GAERS and NEC ratsfor both right and left hippocampi separately. No intrahemisphericcomparison was made.

Open field test of anxiety-like behavior

The open field behavioral test assessing anxietywas performed in aclosed, quiet, light-controlled room in the Department of Medicine,Royal Melbourne Hospital, University of Melbourne. The operatingprotocol was modified slightly from our previous work (Jones et al.,2008). Briefly, animals were brought into the facility at least 60 minprior to testing to habituate them to the environment. They were thenindividually placed in the centre of a 1 m diameter circular arena withan inner circle (diameter 66 cm) separating the inner area from theouter, and allowed to explore the arena for 10 min. The lighting at thecentre of the arenawas ∼90 lx and each experiment was performed atthe same time of day (1000 h–1200 h). Each trial was video-taped, andquantification of the total distance traveled, and the time spent andnumber of entries made into the inner area of the maze wasobjectively assessed using Ethovision Tracking Software (Noldus).Reduced entries and time spent in the central area of the arena areindicative of a greater level of anxiety/fear (Prut and Belzung, 2003).

Statistical analysis of data

Comparisons of the raw volumes of the different brain regionswerecompared between GAERS and NEC using MANOVA with repeatedmeasures (region). Post-hoc planned comparison analysis was thenperformed for each individual MRI region. Total brain volume and thethickness of the somatosensory cortex, as well as animal weights andages were analyzed using Student's unpaired t-tests. For the open field

Fig. 2. Cortical analysis of the MRI obtained from GAERS (black bars) and NEC rats(white bars). GAERS exhibit significant bilateral increases in (A) total cortical volumeand (B) somatosensory cortical thickness (⁎pb0.05). Data represent mean+S.E.M.

Fig. 3. Volumetric analysis of the regions in the limbic system using MRI obtained fromGAERS (black bars) and NEC rats (white bars). (A) GAERS exhibit a significant bilateralincrease in amygdala volume (⁎pb0.05), but no significant alteration in hippocampalvolume (B). Data represent mean+S.E.M.

Table 1Region-Of-Interest analysis of various brain regions in epileptic GAERS and NEC rats

Region GAERS (mm3) NEC (mm3) GAERS(%brain)

NEC(%brain)

Whole brain 1277.7 (16.7)⁎⁎⁎ 1187.9 (15.3) 100% 100%Cortical structuresRCtx 282.0 (1.5)⁎⁎ 270.1 (3.7) 20.6 (0.1) 22.7 (0.2)###

LCtx 267.0 (1.9)⁎ 253.3 (4.8) 19.5 (0.1) 21.3 (0.2)###

RSSC thickness 1.94 mm (0.02)⁎⁎⁎ 1.81 mm (0.03) n/a n/aLSSC thickness 2.00 mm (0.03)⁎⁎⁎ 1.83 mm (0.02) n/a n/aLimbic structuresRAmyg 28.4 (0.9)⁎⁎ 24.3 (0.8) 2.2 (0.1) 2.1 (0.1)LAmyg 28.3 (0.7)⁎⁎⁎ 23.1 (.7) 2.2 (0.1)⁎⁎ 1.9 (0.1)RHC 31.7 (1.1) 30.0 (1.4) 2.5 (0.1) 2.5 (0.1)LHC 31.2 (0.8) 30.4 (1.1) 2.4 (0.1) 2.5 (0.1)Other ROIRTha 24.8 (0.7) 25.7 (1.5) 2.0 (0.1) 2.2 (0.1)LTha 22.7 (0.8) 23.2 (1.1) 1.8 (0.1) 2.0 (0.1)Hypotha 10.9 (1.0) 9.3 (0.7) 0.9 (0.1) 0.8 (0.1)RSt 34.6 (0.7) 33.4 (1.4) 2.7 (0.1) 2.8 (0.1)LSt 34.4 (1.1) 32.5 (1.5) 2.7 (0.1) 2.7 (0.1)Vent 9.0 (0.7)⁎⁎ 5.6 (0.7) 0.7 (0.1)⁎⁎ 0.5 (0.1)

Summary of regional volumes (±S.E.M) and volumes expressed as a % of whole brain(±S.E.M) obtained from GAERS (n=12) and NEC (n=11) rats. Abbreviations: R — right, L —

left; Ctx — cortex, SSC — somatosensory cortex, Amyg— amygdala, HC — hippocampus,Tha — thalamus, Hypotha — hypothalamus, St — striatum, Vent — ventricles, ROI —Region-Of-Interest. ⁎ indicates a region (either total volume or % volume) significantlylarger in GAERS, # indicates a region (either total volume or % volume) significantlylarger in NEC rats; ⁎pb0.05; ⁎⁎pb0.001, ⁎⁎⁎pb0.0001, ###pb0.0001.

270 V. Bouilleret et al. / NeuroImage 45 (2009) 267–274

Author's personal copy

test, Student's unpaired t-tests were performed to compare allvariables between strains separately. Correlations between anxiety-like behaviors and amygdala volumes were performed by normalizingthe region volume by the group mean, and then performing directlinear regression analyses. Data were analyzed using Statistica®

software (StatSoft, Tulsa, OK). Statistical significance in all cases wasset at pb0.05.

Results

At the time of MRI acquisition there were no differencesbetween the two strains in age (GAERS 105.1±1.0 days old; NEC105.1±1.0 days old: t(21)=0.004, p=0.99) or weight (GAERS 191.3±2.9g; NEC 187.6±4.1g: t(21)=0.730, p=0.47). Since the rats from thetwo strains are of equal size, we chose to plot our MRI volumetricdata as raw volumes. These are summarized in Table 1, with thecorresponding percentage of whole brain also given for each region.

Morphometric brain volumes

The total mean brain volume of GAERS was larger than NEC rats(t(21)=3.946, p=0.0007). When examining regions, an overall signifi-cant difference was observed between GAERS and NEC rats (F(1, 21)=

14.10, p=0.0017). The details of all regions studied and post-hocplanned comparison statistical analyses are summarised in Table 1.

Cortical structuresGAERS possess larger left (p=0.006) and right (p=0.012) cortical

volumes compared to NEC rats (Fig. 2A). The somatosensory cortex,the region implicated in seizure initiation in GAERS, was also thickerin this strain in both hemispheres (right: t(21)=3.861; p=0.0008, left:t(21)=4.712; pb0.0001; Fig. 2B).

Limbic structuresThe amygdala volume in GAERS is larger in both hemispheres

(right amygdala, p=0.003; left amygdala, pb0.0001: Fig. 3A). Usingthe Region-Of-Interest analysis, no significant differences wereobserved in the volume of the hippocampus of either hemisphere(pN0.05 for both hippocampi, Fig. 3B).

Other structuresThe lateral ventricular volumes in GAERS were larger than those of

NEC rats (p=0.002). However, no volumetric differences wereobserved in other structures measured between GAERS and NECrats, including the hypothalamus, striatum and the thalami of bothhemispheres (pN0.05 for all comparisons).

Fig. 4. Surface renderings using HDM-LD of the ‘average’ hippocampi of the NEC group, with a superimposed color scale to show regions of inward and outward deformation from thecomparison of the NEC and GAERS ‘average’ hippocampi. Maximal deformation difference for the left hippocampi was 0.21 mm, and for the right hippocampi was 0.25 mm.Therefore, these respective values were used to set the upper and lower limits of the color scale which represents the deformation differences (as shown in the colormap bar.) Thefigures include a colormap bar, depicting the color scale, with ‘warm’ colors depicting regions of outward deformation, and ‘cold’ colors depicting regions of inward deformation. Theright and left hippocampi are labeled, with the rostral view in the upper half of the figure, and caudal view in the lower half of the figure. In the rostral view, the septal pole is medial,while in the caudal view, the septal pole is lateral. The rostral perspective demonstrates the region of maximal inward deformation over the lateral hippocampal surface, near themid-point of the septotemporal axis. The caudal perspective shows the medial surface of each hippocampus. This view best depicts the region of maximal deformation over themedial surface of each hippocampus, which is immediately caudal to the hippocampal commissure. The regions of inward deformation show very similar patterns in the right and lefthippocampi.

271V. Bouilleret et al. / NeuroImage 45 (2009) 267–274

Author's personal copy

HDM-LD findings

HDM-LD showed the GAERS group to have accentuated regionalhippocampal volume loss in two regions (Fig. 4): the medialhippocampal surface immediately caudal to the hippocampal com-missure, and the lateral hippocampal surface over the mid-portion ofthe septotemporal axis. Maximal deformation difference for the lefthippocampi was 0.210 mm, and for the right hippocampi was0.254 mm. These correspond to the deformation patterns assessedvisually between GAERS and NEC rats. Fig. 4 shows surface renderingsof the “average” hippocampi of the NEC group, with a superimposedcolor scale to show regions of inward and outward deformation fromthe comparison of the NEC and GAERS “average” hippocampi. The toptwo images in Fig. 4 show the hippocampi viewed from a rostralperspective, with the region of maximal inward deformation over thelateral hippocampal surface, near the mid-point of the septotemporalaxis. The lower images in Fig. 4 are rotated to show the medial surfaceof each hippocampus. This view best depicts the region of maximaldeformation over the medial surface of each hippocampus, which isimmediately caudal to the hippocampal commissure. The regions ofinward deformation show very similar patterns in the right and lefthippocampi.

GAERS exhibit elevated anxiety-like behavior in the open field test

The time spent in the inner area of the Open Field, an index ofanxiety, was significantly less in GAERS when compared to NEC rats(t(21)=2.166, p=0.042; Fig. 5), which is consistent with our previousreport (Jones et al., 2008). The number of entries into the inner areawas also significantly reduced in GAERS (t(21)=2.315, p=0.031).Furthermore, the reduced time spent in the inner area by GAERSwas accompanied by significantly less exploratory activity of thearena, also indicative of an anxious phenotype (t(21)=4.644, pb0.001;Fig. 5). Within group correlation analyses relating the measures ofanxiety and amygdala volume did not indicate that amygdala volumewas negatively correlated to anxiety for either rat strain.

Discussion

Using MRI, this work examined morphological alterations of brainregions implicated in both seizure-related and anxiety-relatedpathogenesis in GAERS, a widely studied animal model of geneticgeneralised epilepsy. The results demonstrated that GAERS haveincreased cortical, amygdala and ventricular volumes, and increasedthickness of the somatosensory cortex, compared to age-matched ratsof their non-epileptic control (NEC) strain. Additionally, HDM-LD wasapplied to three-dimensionally assess hippocampal surface shape,revealing regional decreases in hippocampal volume despite nosignificant difference in the total volume of the structure.

Our previous research demonstrates that GAERS possess increasedanxiety- and depressive-like behaviors comparedwith NEC rats (Joneset al., 2008), and these affective disturbances are present prior to theonset of the epilepsy, intimating that they are not caused by seizureactivity. The presence of morphological brain changes in these ratsdemonstrated here may be related to the pathologies observed in thismodel: increased sensorimotor cortical thickness could be related tothe absence epilepsy phenotype, since this region is implicated in thegeneration of seizures in this model (Pinault and O'Brien, 2005) anddisorganized and expanded neuronal branching has been observed inthis brain region in the WAG/Rij rat, another well-characterizedgenetic model of absence epilepsy (Karpova et al., 2005). Further,altered limbic structure, including increased amygdala size andaltered hippocampal shape could be responsible for the affectivedisturbance observed, since these regions, particularly the amygdala,are strongly implicated in fear/anxiety (Rosen and Donley, 2006).Clinical studies detailed above demonstrate increased amygdalavolumes in patients with epilepsy comorbid with psychopathologies(Daley et al., 2008; Satishchandra et al., 2003; Tebartz van Elst et al.,1999, 2000, 2002). The neurobiological processes underlying theamygdala enlargement, in both clinical and experimental settings, areuncertain and detailed histological analyses were not performed inthis study. However, posited mechanisms include increased vascularperfusion andmetabolism, increased arbor or numbers of neurons andglia, increased intercellular fluid or increased dopaminergic neuro-transmission (Altshuler et al., 2000; Drevets et al., 2002; Vyas et al.,2002). While no assessment of regional volume abnormalities in otheranimal models of generalised epilepsy currently exists, studies havereported structural alterations in animal models of acquired limbicepilepsy using MRI, predominantly detailing progressive atrophy oflimbic regions following chemical induction of status epilepticus(Nairismagi et al., 2004, 2006; Wolf et al., 2002). No attempts weremade in these reports to document relationships between imagingchanges and the affective disturbance commonly reported in thesemodels.

Using HDM-LD, patients with epilepsy (Hogan et al., 2003),schizophrenia (Csernansky et al., 1998) and depression (Posener etal., 2003) have been shown to have different patterns of changes inregional hippocampal volume. Patients with schizophrenia showminimal volume loss in the lateral hippocampal head and subiculum

Fig. 5. Assessment of open field behavior of GAERS (black bars) and NEC rats (whitebars). (A) Representative traces tracking the path taken by an NEC rat (left) and a GAERS(right) in the Open Field, depictedwith the inner area of the arena. Traces exported fromEthovision™ software, black squares represent starting position. GAERS displaysignificantly greater anxiety-like behavior as evidenced by (B) significantly reducedtime spent exploring the inner exposed area of the area, and (C) significantly reducedtotal distance travelled (⁎pb0.05). Data represent mean+S.E.M.

272 V. Bouilleret et al. / NeuroImage 45 (2009) 267–274

Author's personal copy

(Csernansky et al., 1998) whereas patients with depression show ashape deformation of the subiculum without any measurable totalvolume loss (Posener et al., 2003). Additionally, studies in schizo-phrenia and depression demonstrate that HDM-LD may show regionsof subregional hippocampal surface change without significant totalhippocampal volume change. In this study HDM-LD analysis in GAERSshowed a region of maximal inward deformation over the lateralhippocampal surface, near the mid-point of the septotemporal axis,and over the medial hippocampal surface immediately caudal to thehippocampal commissure. The patterns of maximal inward deforma-tion are symmetrical over both hippocampi, which supports abihemispheric underlying pathophysiology for hippocampal changes.HDM-LD findings provide avenues for further studies of hippocampalchanges in the GAERS model. Different regions of the hippocampusalong the septotemporal axis are composed of distinctly differentsubfields (Witter and Amaral, 2004). For example, near the septal pole,only the dentate gyrus and the CA1–3 subdivisions of the hippocampusare present.Moving from the septal pole approximately 15% of thewayback toward the temporal pole, the subiculum appears. Therefore,different subregions of the hippocampus are represented uniquelyalong segments of the hippocampal surface. By creating mathemati-cally defined, high resolution surfaces, HDM-LD segmentation offerspossible avenues for studying relationships of surface anatomicalchanges correlating to histopathological findings. Changes in hippo-campal surfaces can be compared between groups, or longitudinallyover time within subjects or groups of subjects.

Some limitations to the current study should be discussed. Firstly,only female rats were used and therefore future work should examinewhether these findings are gender-specific, although we havepreviously documented alterations in anxiety levels in both maleand female GAERS (Jones et al., 2008). Furthermore, the rats weregroup-housed according to strain, as is customary for most animalfacilities. This may highlight some behavioral differences in socialinteractions which may impact on cortical thickness between thestrains. With respect to the HDM-LD mapping, the current report onlyexamines for shape differences in the hippocampus, because this isthe structure that has been analysed with this technique in humanswith epilepsy and with psychosis (Csernansky et al., 1998; Hogan etal., 2003). However, future work applying this methodology to otherrelevant brain structures, such as the amygdala, will be of greatinterest.

In conclusion the results of this study demonstrate a number ofmorphometric alterations in brain regions implicated in anxiety andseizure initiation in GAERS, an animal model of genetic generalisedepilepsy possessing a comorbid hyperanxious phenotype. Despite anenlarged total brain volume in GAERS, the increases are not global butregionally selective, affecting the amygdala, cortex and ventricles butnot other structures, including thalamus and striatum. HDM-LDanalysis showed subregional hippocampal volume changes with apattern resembling hippocampal HDM-LD findings in human studiesof depression (Posener et al., 2003). The significant volumetricdifferences observed provide a strong foundation for a longitudinalexamination of the development of these changes with respect to theinitiation of seizures and the appearance of the increased anxietyobserved in this model. These findings may also be useful inilluminating the pathogenesis of affective disorders in general, aswell as modeling psychiatric comorbidities of epilepsy.

Acknowledgments

This research was supported by an NHMRC project grant(#400088) to TJO and MS and a Special Purposes Grant from theUniversity of Melbourne to NJ.

Financial disclosureThe authors declare no conflicts, financial or otherwise, associated

with this work.

References

Adams, S.J., O'Brien, T.J., Lloyd, J., Kilpatrick, C.J., Salzberg, M.R., Velakoulis, D., 2008.Neuropsychiatric morbidity in focal epilepsy. Br. J. Psychiatry 192, 464–469.

Altshuler, L.L., Bartzokis, G., Grieder, T., Curran, J., Jimenez, T., Leight, K., Wilkins, J.,Gerner, R., Mintz, J., 2000. An MRI study of temporal lobe structures in men withbipolar disorder or schizophrenia. Biol. Psychiatry 48, 147–162.

Brambilla, P., Barale, F., Caverzasi, E., Soares, J.C., 2002. Anatomical MRI findings inmoodand anxiety disorders. Epidemiol. Psichiatr. Soc. 11, 88–99.

Briellmann, R.S., Kalnins, R.M., Hopwood, M.J., Ward, C., Berkovic, S.F., Jackson, G.D.,2000. TLE patients with postictal psychosis: mesial dysplasia and anteriorhippocampal preservation. Neurology 55, 1027–1030.

Campbell, S., MacQueen, G., 2006. An update on regional brain volume differencesassociated with mood disorders. Curr. Opin. Psychiatry 19, 25–33.

Caplan, R., Arbelle, S., Magharious, W., Guthrie, D., Komo, S., Shields, W.D.,Chayasirisobhon, S., Hansen, R., 1998. Psychopathology in pediatric complex partialand primary generalized epilepsy. Dev. Med. Child Neurol. 40, 805–811.

Caplan, R., Siddarth, P., Gurbani, S., Hanson, R., Sankar, R., Shields, W.D., 2005.Depression and anxiety disorders in pediatric epilepsy. Epilepsia 46, 720–730.

Christensen, G.E., Rabbitt, R.D., Miller, M.I., 1993. A deformable neuroanatomy textbookbased on viscous fluid mechanics. In: Prince, J. (Ed.), 27th Annual Conference onInformation Sciences and System. The Johns Hopkins University, Baltimore, pp.211–216.

Christensen, G.E., Rabbitt, R.D., Miller, M.I., 1994. 3D brain mapping using a deformableneuroanatomy. Phys. Med. Biol. 39, 609–618.

Christensen, G.E., Rabbitt, R.D., Miller, M.I., Joshi, S., Grenander, U., Coogan, T.A., 1995.Topological properties of smooth anatomic maps. Information Processing inMedical Imaging. Springer, Netherlands, pp. 101–112.

Christensen, J., Vestergaard, M., Mortensen, P.B., Sidenius, P., Agerbo, E., 2007. Epilepsyand risk of suicide: a population-based case-control study. Lancet Neurol. 6,693–698.

Csernansky, J.G., Joshi, S., Wang, L., Haller, J.W., Gado, M., Miller, J.P., Grenander, U.,Miller, M.I., 1998. Hippocampal morphometry in schizophrenia by high dimen-sional brain mapping. Proc. Natl. Acad. Sci. U. S. A. 95, 11406–11411.

Daley, M., Siddarth, P., Levitt, J., Gurbani, S., Shields, W.D., Sankar, R., Toga, A., Caplan, R.,2008. Amygdala volume and psychopathology in childhood complex partialseizures. Epilepsy Behav. 13, 212–217.

Danober, L., Deransart, C., Depaulis, A., Vergnes, M., Marescaux, C., 1998. Pathophysio-logical mechanisms of genetic absence epilepsy in the rat. Prog. Neurobiol. 55,27–57.

Davies, S., Heyman, I., Goodman, R., 2003. A population survey of mental healthproblems in children with epilepsy. Dev. Med. Child Neurol. 45, 292–295.

Drevets, W.C., Price, J.L., Bardgett, M.E., Reich, T., Todd, R.D., Raichle, M.E., 2002. Glucosemetabolism in the amygdala in depression: relationship to diagnostic subtype andplasma cortisol levels. Pharmacol. Biochem. Behav. 71, 431–447.

Forsgren, L., Nystrom, L., 1990. An incident case-referent study of epileptic seizures inadults. Epilepsy Res. 6, 66–81.

Gardner, R., Hogan, R.E., 2005. Three-dimensional deformation-based hippocampalsurface anatomy, projected on MRI images. Clin. Anat. 18, 481–487.

Grohn, O., Pitkanen, A., 2007. Magnetic resonance imaging in animal models ofepilepsy-noninvasive detection of structural alterations. Epilepsia 48 (Suppl. 4),3–10.

Hermann, B., Seidenberg, M., Jones, J., 2008. The neurobehavioural comorbidities ofepilepsy: can a natural history be developed? Lancet Neurol. 7, 151–160.

Hesdorffer, D.C., Hauser,W.A., Annegers, J.F., Cascino, G., 2000. Major depression is a riskfactor for seizures in older adults. Ann. Neurol. 47, 246–249.

Hesdorffer, D.C., Hauser, W.A., Olafsson, E., Ludvigsson, P., Kjartansson, O., 2006.Depression and suicide attempt as risk factors for incident unprovoked seizures.Ann. Neurol. 59, 35–41.

Hogan, R.E., Bucholz, R.D., Joshi, S., 2003. Hippocampal deformation-based shapeanalysis in epilepsy and unilateral mesial temporal sclerosis. Epilepsia 44,800–806.

Hogan, R.E., Wang, L., Bertrand, M.E., Willmore, L.J., Bucholz, R.D., Nassif, A.S.,Csernansky, J.G., 2004. MRI-based high-dimensional hippocampal mapping inmesial temporal lobe epilepsy. Brain 127, 1731–1740.

Hogan, R.E., Carne, R.P., Kilpatrick, C.J., Cook, M.J., Patel, A., King, L., O'Brien, T.J., 2008.Hippocampal deformation mapping in MRI negative PET positive temporal lobeepilepsy. J. Neurol. Neurosurg. Psychiatry 79, 636–640.

Hogan, R.E., Bouilleret, V., Liu, Y.R., Wang, L., Williams, J.P., Jupp, B., Myers, D. andO'Brien, T.J. in press MRI-based Large Deformation High Dimensional Mapping ofthe Hippocampus in Rats: Development and Validation of the Technique. Accepted:J MRI.

Jones, J.E., Watson, R., Sheth, R., Caplan, R., Koehn, M., Seidenberg, M., Hermann, B.,2007. Psychiatric comorbidity in children with new onset epilepsy. Dev. Med. ChildNeurol. 49, 493–497.

Jones, N.C., Salzberg, M.R., Kumar, G., Couper, A., Morris, M.J., O'Brien, T.J., 2008. Elevatedanxiety and depressive-like behavior in a rat model of genetic generalized epilepsysuggesting common causation. Exp. Neurol. 209, 254–260.

Joshi, S., Miller, M.I., Christensen, G.E., Banerjee, A., Coogan, T.A., Grenander, U., 1995.Hierarchical brain mapping via a generalized Dirichlet solution for mapping brainmanifolds. Proc. SPIE Inter. Symp. Opt. Sci. Eng. Instrum. 278–289.

Jupp, B., Williams, J.P., Tesiram, Y.A., Vosmansky, M., O'Brien, T.J., 2006. Hippocampal T2signal change during amygdala kindling epileptogenesis. Epilepsia 47, 41–46.

Karpova, A.V., Bikbaev, A.F., Coenen, A.M., van Luijtelaar, G., 2005. Morphometric Golgistudy of cortical locations in WAG/Rij rats: the cortical focus theory. Neurosci. Res.51, 119–128.

273V. Bouilleret et al. / NeuroImage 45 (2009) 267–274

Author's personal copy

Maier, M., Mellers, J., Toone, B., Trimble, M., Ron, M.A., 2000. Schizophrenia, temporallobe epilepsy and psychosis: an in vivo magnetic resonance spectroscopy andimaging study of the hippocampus/amygdala complex. Psychol. Med. 30, 571–581.

Miller, M.I., Christensen, G.E., Amit, Y., Grenander, U., 1993. Mathematical textbook ofdeformable neuroanatomies. Proc. Natl. Acad. Sci. U. S. A. 90, 11944–11948.

Nairismagi, J., Grohn, O.H., Kettunen, M.I., Nissinen, J., Kauppinen, R.A., Pitkanen, A.,2004. Progression of brain damage after status epilepticus and its association withepileptogenesis: a quantitative MRI study in a rat model of temporal lobe epilepsy.Epilepsia 45, 1024–1034.

Nairismagi, J., Pitkanen, A., Kettunen, M.I., Kauppinen, R.A., Kubova, H., 2006. Statusepilepticus in 12-day-old rats leads to temporal lobe neurodegeneration andvolume reduction: a histologic and MRI study. Epilepsia 47, 479–488.

Nelson, M.D., Saykin, A.J., Flashman, L.A., Riordan, H.J., 1998. Hippocampal volumereduction in schizophrenia as assessed by magnetic resonance imaging: a meta-analytic study. Arch. Gen. Psychiatry 55, 433–440.

Ott, D., Siddarth, P., Gurbani, S., Koh, S., Tournay, A., Shields, W.D., Caplan, R., 2003.Behavioral disorders in pediatric epilepsy: unmet psychiatric need. Epilepsia 44,591–597.

Pinault, D., 2003. Cellular interactions in the rat somatosensory thalamocortical systemduring normal and epileptic 5–9 Hz oscillations. J. Physiol. 552, 881–905.

Pinault, D., O'Brien, T.J., 2005. Cellular and network mechanisms of genetically-determined absence seizures. Thalamus and Related Systems 3, 181–203.

Posener, J.A., Wang, L., Price, J.L., Gado, M.H., Province, M.A., Miller, M.I., Babb, C.M.,Csernansky, J.G., 2003. High-dimensional mapping of the hippocampus indepression. Am. J. Psychiatry 160, 83–89.

Prut, L., Belzung, C., 2003. The open field as a paradigm to measure the effects of drugson anxiety-like behaviors: a review. Eur. J. Pharmacol. 463, 3–33.

Rosen, J.B., Donley, M.P., 2006. Animal studies of amygdala function in fear anduncertainty: relevance to human research. Biol. Psychol. 73, 49–60.

Salzberg, M., Kumar, G., Supit, L., Jones, N.C., Morris, M.J., Rees, S., O'Brien, T.J., 2007. Earlypostnatal stress confers enduring vulnerability to limbic epileptogenesis. Epilepsia48, 2079–2085.

Satishchandra, P., Krishnamoorthy, E.S., van Elst, L.T., Lemieux, L., Koepp, M., Brown, R.J.,Trimble, M.R., 2003. Mesial temporal structures and comorbid anxiety in refractorypartial epilepsy. J. Neuropsychiatry Clin. Neurosci. 15, 450–452.

Tebartz van Elst, L., Woermann, F.G., Lemieux, L., Trimble, M.R., 1999. Amygdalaenlargement in dysthymia—a volumetric study of patients with temporal lobeepilepsy. Biol. Psychiatry 46, 1614–1623.

Tebartz van Elst, L., Woermann, F., Lemieux, L., Trimble, M.R., 2000. Increased amygdalavolumes in female and depressed humans. A quantitative magnetic resonanceimaging study. Neurosci. Lett. 281, 103–106.

Tebartz van Elst, L., Baeumer, D., Lemieux, L., Woermann, F.G., Koepp, M., Krishna-moorthy, S., Thompson, P.J., Ebert, D., Trimble, M.R., 2002. Amygdala pathology inpsychosis of epilepsy: a magnetic resonance imaging study in patients withtemporal lobe epilepsy. Brain 125, 140–149.

Tellez-Zenteno, J.F., Patten, S.B., Jette, N., Williams, J., Wiebe, S., 2007. Psychiatriccomorbidity in epilepsy: a population-based analysis. Epilepsia 48, 2336–2344.

Vyas, A., Mitra, R., Shankaranarayana Rao, B.S., Chattarji, S., 2002. Chronic stress inducescontrasting patterns of dendritic remodeling in hippocampal and amygdaloidneurons. J. Neurosci. 22, 6810–6818.

Watson, C., Jack Jr., C.R., Cendes, F., 1997. Volumetric magnetic resonance imaging.Clinical applications and contributions to the understanding of temporal lobeepilepsy. Arch. Neurol. 54, 1521–1531.

Witter, M.P., Amaral, D.G., 2004. Hippocampal Formation. In: Paxinos, G. (Ed.), The RatNervous System. Elsevier Academic Press, Boston, pp. 635–704.

Wolf, O.T., Dyakin, V., Patel, A., Vadasz, C., de Leon, M.J., McEwen, B.S., Bulloch, K., 2002.Volumetric structural magnetic resonance imaging (MRI) of the rat hippocampusfollowing kainic acid (KA) treatment. Brain Res. 934, 87–96.

274 V. Bouilleret et al. / NeuroImage 45 (2009) 267–274