NEW RESEARCH

Amygdala and Hippocampus EnlargementDuring Adolescence in Autism

Wouter Groen, M.D., Ph.D., Michelle Teluij, M.D., Jan Buitelaar, M.D., Ph.D.,Indira Tendolkar, M.D., Ph.D.

Objective: The amygdala and hippocampus are key components of the neural systemmediating emotion perception and regulation and are thought to be involved in thepathophysiology of autism. Although some studies in children with autism suggest that thereis an enlargement of amygdala and hippocampal volume, findings in adolescence aresparse. Method: We measured amygdala and hippocampus volume in a homogeneousgroup of adolescents with autism (12 through18 years; n � 23) and compared them with anage-, sex-, and IQ-matched control group (n � 29) using a validated automated segmentationprocedure in 1.5-T magnetic resonance images. All analyses were adjusted for total brainvolume. Results: Repeated-measures analysis revealed a significant group � hemisphere �brain structure interaction (p � .038), even when corrected for total brain volume. Post-hocanalysis showed that the right amygdala and left hippocampus were significantly enlarged (p� .010; p � .015) in the autism compared with the control group. There were no significantcorrelations between age and amygdala or hippocampus volume. Conclusions: The abnor-mal enlargement of the amygdala and hippocampus in adolescents with autism adds toprevious findings of enlargement of these structures in children with autism. This may reflectincreased activity of these structures and thereby altered emotion perception and regulation.Our results could therefore be interpreted in light of developmental adaptation of the autisticbrain to a continuous overflow of emotional learning experiences. J. Am. Acad. Child Adolesc.Psychiatry, 2010;49(6):552–560. Key Words: autism, emotion, amygdala, hippocampus

di

bpcatetairtisaathb

A Autism is a neurodevelopmental disor-der characterized by impaired social in-teraction, language and communication

deficits, and stereotyped repetitive behavior. De-spite the great variation in clinical presentation, acentral feature of autism is the impairment insocial and emotional behavior. Persons with au-tism typically do not exhibit adequate socio-emotional reciprocity, have difficulty under-standing social-emotional cues, and often displayblunted affect.1 Development of adequate socio-emotional behavior depends on the structuraland functional integrity of brain regions mediat-ing emotion perception and regulation.2 Volu-metric investigation of these brain regions mayprovide important clues that can help to eluci-

This article is the subject of an editorial by Dr. Bradley Peterson onpage 533.Supplemental material cited in this article is available online.

d

JOURN

552 www.jaacap.org

ate which structures contribute to the emotionalmpairment in autism.

Converging evidence from studies in humaneings and in animals suggests that emotionerception and regulation is mediated by a brainircuit in which the amygdala and hippocampusre central components.3 The amygdala enableshe instantaneous recognition and evaluation ofmotionally salient stimuli, subsequent produc-ion of affective states, and regulation of theutonomic response.4 Apart from its general rolen memory, the hippocampus is implicated in theegulation of affective behavior elicited by emo-ionally salient stimuli. It does so through inhib-tory connections with the amygdala and othertructures involved in emotion perception, en-bling the production of contextually appropri-te affective behavior.4 Thus, structural and func-ional abnormalities of the amygdala andippocampus may result in impaired affectiveehavior and emotion regulation, which has in-

eed been shown in a variety of psychiatric

AL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY

VOLUME 49 NUMBER 6 JUNE 2010

dots

tspdimailaHifiaf

AMYGDALA-HIPPOCAMPAL ENLARGEMENT IN AUTISM

disorders.23 It is therefore hardly surprising thatstructural and functional abnormalities of amyg-dala and hippocampus have been found in au-tism.24,25 Yet, it is important to take into accountthat volumetric abnormalities of the amygdalaand hippocampus in autism may have a differentimpact, as they are linked to developmentalabnormalities rather than to disease-state abnor-malities (as has been shown, for example, indepression).26 Investigating amygdala and hip-pocampus volume throughout development inchildren or adolescents with autism may there-fore reveal crucial insights in the neurobiology ofautism.

Several studies have investigated amygdalavolume in autism. As is evident from Table 1,there is substantial variance with respect to ageranges, clinical presentation, and methodologicalapproaches in the volumetric amygdala studies.Nevertheless, there seems to be convergence

TABLE 1 Overview of Studies on Amygdala and Hippoc

Study Group

AgeFirst Author, Year, Ref. A PD AS C

Dager 20075 29 16 13 3-4Nicolson 20066 21 24 6-16Rojas 20067 24 23 7-44Nacewicz 20068 23 5 26 8-25Palmen 20069a 42 42 7-25Rojas 200410 15 17 19-47Schumann 200411 19 11 11 7-12Schumann 200411 19 11 11 12-18Herbert 200312 17 15 7-11Bigler 200313 38 27 7-31Sparks 200214 29 16 26 3-4Pierce 200115 7 8 20-42Saitoh 200116 59 51 2-49Haznedar 200017 10 7 17 27b

Howard 200018a 10 10 16-40Abell 199919 15 15 28,8b

Aylward 199920 14 14 20,5b

Piven 199821 35 36 12-29Saitoh 199522 33 32 13,8b

Note: A � autistic disorder; AM � amygdala volume; AS � Asperger’s disHP � hippocampus; IQ � intelligence quotient; M � male; MF � maotherwise specified; s � shape; TBVa � total brain volume adjusted;

aAutism sample includes people with Asperger’s disorder.bMean age.cMean IQ.dCombined measurement of the amygdala and hippocampus.eDisproportional enlargement of the amygdala in the autistic disorder subfDecreased area dentata (part of the hippocampus).

across studies toward enlargement of the amyg- b

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY

VOLUME 49 NUMBER 6 JUNE 2010

ala in children with autism and disappearancef this overgrowth in adults with autism. There iso date, however, no report of a homogenousample of adolescents with autism.

Findings of abnormal amygdala volumes in au-ism fit microscopic evidence from neuropathologytudies that have demonstrated increased cell-acking density, abnormally small neurons, and aecreased number of neurons in the amygdala,

ndicative of a curtailed maturation process.27 Post-ortem studies25 and neuro-imaging studies5 have

dditionally demonstrated structural abnormalitiesn the hippocampus in autism. The neuropatho-ogic abnormalities observed in the hippocampusre consistent with those observed in the amygdala.25

ippocampal volume studies in autism show annconsistent pattern. The majority of studies did notnd significant volume differences between theutism and control group, but the authors mainlyocused on adults or included participants of a

us Volume in Autism

x

IQ MF TBVa AM

HP

F v s

13 NR Sex � � �

0 �70 Age, HC, VIQ � � �

0 60-133 Age, IQ, educ � L10 97c Age, sex � 23 �80 Age, sex, IQ � � �

11 62-140 Age, sex � L2 L10 �/�70 Age � 1 10 �/�70 Age � � 10 �80 Sex � 2d 2d

0 �62 Age, sex � � �

15 NR Age � 1e 1�

0 73-102 Age, sex � 218 41-135 NR � 2f

4 55-125 NR � � �

0 NR Age, VIQ, sex � 1 �

6 NR Age, sex, IQ � L10 106c Age, sex, IQ � 2 2

25 52-136 NR � �

7 NR Age � �

C � controls; educ � education; F � female; HC � head circumference;factors; NR � not reported; PD � pervasive developmental disorder notlume; VIQ � verbal IQ.

amp

Se

M

45454754811141412265561592302024287158

order;tchingv � vo

group.

road age range.3,17,18,21 Nonetheless, there is some

553www.jaacap.org

ase

TfaCismtvpip

tUfnIadmpIpibaawpcopk

DN

tient; N

GROEN et al.

indication of an enlargement of the hippocampusin children with autism and a normalized volumein adults with autism (Table 1). Again, evidence forvolumetric abnormalities of this structure duringadolescence is missing.

The contradictory results from the aforemen-tioned volumetric studies investigating amyg-dala and hippocampus in autism may be causedby the heterogeneity of the autism samples in-cluded in these studies. As the phenotypic vari-ation probably reflects heterogeneity on the neu-robiologic level, it is important to investigatevolumetric differences in homogeneous samplesto provide more insight into the neuropathologicunderpinnings of autism. As many studies founda volume increase of the amygdala and hip-pocampus during childhood and normalizationor decrease in adulthood, the aim of the currentstudy was to evaluate the volume of the amyg-dala and hippocampus in a homogeneous, well-defined group of adolescents with autism (be-tween 12 and 18 years of age) and a matchedcontrol group.

Most studies used manual tracing procedures,but differences in the anatomic definitions of theamygdala and hippocampus impede the compa-rability of the studies. We used an automatedsegmentation procedure, which is shown to beequal in accuracy to manual tracing methods.28

METHODParticipantsTwo groups participated in the study: 29 typicallydeveloping (TD) adolescents (five girls and 24 boys)and 23 adolescents with autism (three girls and20 boys). All participants were between 12 and 18years of age, right-handed, and Caucasian. Written

TABLE 2 Participant characteristics (mean � SD)

Autism group (n�23)

Age (years) 15.15 � 1.86Sex (m/f) 20/3Full scale IQ 99.52 � 20.12Performance IQ 99.39 � 16.64Verbal IQ 99.52 � 21.68ADI: social deficits 16.30 � 5.89ADI: communication deficits 20.00 � 8.09ADI: repetitive behavior 12.52 � 4.93

Note: ADI � Autism Diagnostic Interview–Revised; IQ � intelligence quo

informed consent was obtained from all participants m

JOURN

554 www.jaacap.org

nd their parents after complete description of thetudy. The study was approved by the local medicalthical committee.

Matching was performed rigorously; more than 200D adolescents from local high schools were assessed

or verbal IQ, performance IQ and full-scale IQ usingshort form of the Wechsler Intelligence Scale for

hildren III (WISC-III) including Vocabulary, Similar-ties, Block Design, and Picture Completion to finduitable matches. As a result, the autism group wasatched without significant differences in age, sex,

otal intelligence quotient (IQ), performance IQ, anderbal IQ (Table 2). In the autism group, sevenarticipants had previously used psychotropic med-

cation, of whom four had used risperidone and twoipamperone.

The participants with autism were recruitedhrough Karakter Child and Adolescent Psychiatryniversity Center in Nijmegen. Diagnostic assignment

ollowed DSM-IV criteria for autistic disorder. Diag-ostic characterization included the Autism Diagnosticnterview–Revised29 as assessed by a trained cliniciannd a series of clinical assessments that includedetailed developmental history, clinical observation,edical work-up, and cognitive testing. The partici-

ants with autism were tested with the full Wechslerntelligence Scale for Children III. To screen for theresence of psychiatric disorders or behavior problems

n the controls, CBCL questionnaires were completedy the parents/guardians of the control participants,nd TRF questionnaires were completed by a teachert school. None of the control participants scoredithin the clinical range on any of the CBCL or TRFroblem scales. Exclusion criteria included any medi-al condition affecting CNS function, neurological dis-rders, and substance abuse. In particular, none of thearticipants had a history of seizure disorder, as this isnown to affect the volumes of interest.30,31

ata Acquisition and Procedureeuroimaging data were acquired on a 1.5-T Sie-

Controls (n�29) p-value Statistic (t or �2)

15.65 � 1.66 0.31 1.0323/5 0.68 0.17

104.93 � 8.68 0.23 1.20105.24 � 11.56 0.14 1.49104.90 � 9.62 0.28 1.11

NA — —NA — —NA — —

A � not applicable.

ens Sonata MR scanner at the Donders Institute for

AL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY

VOLUME 49 NUMBER 6 JUNE 2010

DSgId(�ebsqabocfccaescA

RAsrvslte5bFatcv

AMYGDALA-HIPPOCAMPAL ENLARGEMENT IN AUTISM

Brain, Cognition and Behavior. For each subject, aT1-weighted, whole-brain scan was collected(MPRAGE; TI 850 ms, TR 2250 ms, TE 3.68 ms, flipangle 15°, field of view [FoV] 256 � 256 � 176 mm,voxel size 1.0 � 1.0 � 1.0 mm). The participants werefamiliarized with the set-up and normal scanningprocedures before the actual MR acquisition bymeans of a 30-minute introductory rehearsal in aseparate replica (dummy) scanner.

MRI Volumetric AnalysisVolumetric analysis of the amygdala and hippocam-pus was performed with Freesurfer (4.0.5 version), anautomated whole-brain segmentation procedure.28 Weused Freesurfer in the fully automated form. Briefly,the processing stream comprises removal of the skulland dura, automated Talaraich transformation, defin-ing the gray–white matter boundary, intensity normal-ization, and segmentation of subcortical gray matter.The Freesurfer application enables automatic labelingof subcortical structures using a probabilistic algo-rithm. Initially, each image is a rigid body registered toa probabilistic atlas based on manually labeled images.Thereafter, the image is morphed to the atlas by anonlinear transform, and a Bayesian segmentationprocedure is used. Each voxel in the MRI volume isautomatically assigned to a neuro-anatomical labelbased on probabilistic information estimated from amanually labeled training set. The labeling procedureis not biased by anatomical variability.32 The segmen-tation procedure is based on three types of probabili-ties to disambiguate labels. First, the likelihood that agiven structure occurs at a specific atlas location.Second, the likelihood of the image intensity given thattissue class. Third, the probability that a voxel belongsto a given tissue class based on likelihood of the spatialconfiguration of labels. This automated segmentationand labeling procedure has been shown to be ofaccuracy equal to that of manual tracing methods32

and relatively insensitive to changes in acquisitionparameters.28 Examples of amygdala and hippocam-pal region definitions in 1.5-T MR images can be found

TABLE 3 Summary of Repeated-Measures Analysis of Co

Effect

GroupBrain structureHemisphereGroup � brain structureGroup � hemisphereBrain structure � hemisphereBrain structure � hemisphere � group

Note: Df � degrees of freedom.

in Figures S1, S2, and S3, available online. h

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY

VOLUME 49 NUMBER 6 JUNE 2010

ata Analysistatistical analysis was performed with Statistical Pro-ram for the Social Sciences (SPSS) Edition 15.0 (SPSSnc., Chicago, IL). To investigate the presence of volumeifferences between the groups, we performed a group

autism versus controls) � hemisphere (right versus left)brain structure (amygdala and hippocampus) repeat-

d-measures analysis of covariance, with group as aetween-subjects variable and hemisphere and braintructure as within-subject variables. In these and subse-uent analyses, total brain volume was used as a covari-te to control for possible differences in brain volumeetween the autism and control group. Post-hoc analysesf covariance (ANCOVAs) with total brain volume asovariate were performed to investigate significant ef-ects. To investigate whether the developmental timeourse showed a deviant pattern in the autism group, wealculated Pearson’s correlations between age, amygdala,nd hippocampus volume in both participant groups. Toxplore whether the brain volumes were associated withocial impairment in autism, we calculated Pearson’sorrelations between the social impairment score on theDI-R and volumes of the amygdala and hippocampus.

ESULTSll volumes were distributed normally. Table 3

hows a summary of the results. As expected, theepeated-measures ANCOVA with TBV as a co-ariate revealed a significant main effect for braintructure (F(1,49) � 32.57, p � .001), reflecting thearger size of hippocampus when compared withhe amygdala. We also found a significant mainffect of group (autism versus controls; F(1,49) �.53, p � .023) and a significant main effect ofrain structure (amygdala and hippocampus;(1,49) � 32.57, p � .001). Importantly, we foundsignificant group � hemisphere � brain struc-

ure interaction (F(1,49) � 4.55, p � .038), indi-ating that the group differences of the amygdalaolume and hippocampus volume differed per

ance Results

F ratio p

5.53 .02332.57 �.0010.28 .5981.28 .2641.07 .3070.13 .7224.55 .038

vari

df

1.491.491.491.491.491.491.49

emisphere. Post-hoc ANCOVAs for each hemi-

555www.jaacap.org

dglpsdsitw

asPhtheedat(rpvpadAtaadbto

ipidaro

stwosg

GROEN et al.

sphere showed that the left hippocampus (F(1,49)� 6.31, p � .015) and the right amygdala (F(1,49)� 7.20, p � .010) were significantly enlarged inthe autism group (corrected for TBV), whereasvolume differences between groups for leftamygdala (F(1,49) � 2.69, p � .11) and righthippocampus (F(1,49) � 1.55, p � .22) did notreach significance (TBV corrected). The resultsindicate that the enlargement of the left hip-pocampus and the right amygdala is dispropor-tionate to (i.e., greater than) the enlargement ofthe total brain volume in the autism group. It isnoteworthy that the results of the group compar-isons were similar without inclusion of TBV as acovariate.

The volume measurements of the amygdala,hippocampus, and total brain volume (TBV) areshown in Figure 1 and Table 4. The volumes ofthe amygdala and hippocampus were enlargedin the autism group. Compared with the controlgroup, the autism group showed, in the right andleft amygdala, a 7.8% and a 5.8% enlargementrespectively, whereas the left hippocampus dem-onstrated a 6.25% enlargement and the righthippocampus a 3.8% enlargement. There were nosignificant differences in TBV between thegroups (t � �0.825, df � 50, p � .41).

Age correlated with neither volumes of theamygdala or hippocampus in the autism group(left amygdala: r � �0.024, right amygdala: r ��0.099, left hippocampus: r � �0.11, right hip-pocampus: r � �0.172) nor in the control group(left amygdala: r � �0.16, right amygdala: r ��0.12, left hippocampus: r � �0.07, right hip-pocampus: r � �0.11). Because of the sex speci-ficity in maturation of the amygdala and hip-pocampus,33 we calculated the age correlations inmale participants separately. However, this didnot reveal a significant correlation between ageand hippocampus or amygdala volume either.Finally, we did not find significant correlations inthe autism group between social impairmentsubscore on the ADI-R and volume of any of thestructures investigated.

DISCUSSIONIn the current study, we used an automatedsegmentation procedure to investigate amygdalaand hippocampus volumes in a well-definedsample of high-functioning adolescents with au-tism (12 through 18 years) and a carefully

matched control group. Although there were no f

JOURN

556 www.jaacap.org

ifferences in total brain volume between bothroups, we found a statistically significant en-

argement of the right amygdala and left hip-ocampus even when correcting for total brainize. Although we did not find significant evi-ence for enlargement in both hemispheres (pos-ibly because of the small sample size), visualnspection of the volume scatter plots suggestedhat the left amygdala and right hippocampus

ere also enlarged in the patient group.Our finding of an enlarged right amygdala

nd left hippocampus in autism supports earliertudies albeit of different age ranges.7,10,11,14,18,19

reviously, enlargement of the amygdala andippocampus has been found in 3- to 4-year-old

oddlers with autism.14 A study later in child-ood also provides evidence for persistence of annlarged amygdala in 8- to 12-year-olds andnlarged hippocampus in 8- to 18-year-old chil-ren and adolescents with autism, although theuthors did not find a significant enlargement ofhe amygdala in older children and adolescents12 to 18 years).11 The negative finding in this ageange may be caused by insufficient statisticalower, considering the sample size (n � 11), asisual inspection of the figures provided in thatublication suggest that participants between 12nd 18 years seem to have a larger right amyg-ala in both the low-functioning as well as thesperger’s group compared with controls. Fur-

hermore, some authors have observed enlargedmygdala18,19 and hippocampus volumes7,10 indults with autism and mixed samples of chil-ren and adults. However, these results have toe interpreted with caution, because no correc-ion for total brain volume was performed in twof these studies.18,19

Contrary to our findings, a number of stud-es observed decreased amygdala and hip-ocampus volumes or reported negative find-

ngs. Several factors may account for thisiscrepancy. First, most studies focused ondults with autism or included a broader ageange, which hampered the detection of devel-pmental changes in volume.8-10,15-17,20,21

Second, several studies investigated smallamples and samples with a heterogenic popula-ions, including mentally retarded individuals

ith autism or participants with Asperger’s dis-rder and PDD-NOS.5,8,9,15,17,18,22 Third, severaltudies did not tightly match the controlroups.5,13,15,17 Fourth, the anatomical borders

or manual segmentation of the amygdala and

AL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY

VOLUME 49 NUMBER 6 JUNE 2010

hvT

Fd

AMYGDALA-HIPPOCAMPAL ENLARGEMENT IN AUTISM

hippocampus showed a considerable variationacross studies, impeding the comparability of theresults. Finally, at least one paper16 included

IGURE 1 Volumes of right and left amygdala and hipisorder; TD � typically developing participants. *Signific

patients with autism in which a major proportion a

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY

VOLUME 49 NUMBER 6 JUNE 2010

ad epilepsy. Epilepsy is known to reduce theolume of the amygdala31 and hippocampus.30

his may explain the smaller hippocampi in the

mpus. Note: ASD � subjects with autism spectrumdifferent from controls at p � .05.

pocaantly

utism group reported by Saitoh et al.16

557www.jaacap.org

eaaceto

bvpaivapbhaitAtihpdthhdcaideiawae

GROEN et al.

In line with another study,9 we did not find acorrelation in the autism or the control groupbetween age and amygdala or hippocampus vol-ume. This is probably due to the considerableinterindividual variation in brain morphologythat impedes the identification of developmentalchanges, and to the narrow age range of theparticipants.

Taken together, the data suggest that duringchildhood there is a developmental abnormalityin amygdala and hippocampus volume in autismthat persists into adolescence. Thus, the questionarises how this enlargement can be interpreted.One might speculate that increased use may leadto an increase in volume, as has been shown forvarious brain structures including amygdala inanimal studies34 and the hippocampus35 in hu-man studies. The enlargement of the amygdalaand hippocampus may then reflect an adaptiveeffect in response to increased activity through-out childhood and adolescence in autism. In-deed, a significant correlation has been foundbetween enlarged amygdala volumes and in-creased anxiety as well as increased social impair-ments in autism, suggesting that amygdala en-largement may be a predictor of the degree of socialand emotional deficits in autism.36

As the hippocampus has an important regula-tory effect on amygdala activity via a densenetwork of reciprocal connections, it is plausiblethat the hippocampus enlarges in response toheightened amygdala activity. Although it re-mains to be elucidated in autism, affective andcognitive symptoms of stress-related disorderscan, in principle, be attributed to such an inter-active effect of both the amygdala and hippocam-pus.37,38 Given the interaction between amygdalaand hippocampus during emotional memoryprocesses, our data suggest that the enlargementof these structures during adolescence in autismmay lead to a broader unselected encoding of

TABLE 4 Mean Volumetric Data (mm3 � SD)

Autism patie

Total brain volume 1683259 � 2Right amygdala 1756.22Left amygdala 1685.35Right hippocampus 4724.91Left hippocampus 4875.26

*Significantly different from controls at p � .05.

various emotional aspects, thereby changing o

JOURN

558 www.jaacap.org

motional learning experiences. Note that a par-llel enlargement of both hippocampus andmygdala throughout childhood and adoles-ence in autism is more long lasting than, forxample, the state-related volumetric abnormali-ies during a first depressive episode, in whichnly amygdala enlargement has been found.26

Our findings in individuals with autism, com-ined with previous work, suggest that there is aolume increase of the amygdala and hippocam-us in childhood that persists into adolescence. Yet,s individuals with autism pass from adolescencento adulthood, their amygdala and hippocampusolumes seem to normalize, suggesting that there isrelative loss of volume. We hypothesize that a

rocess of chronic stress in autism, possibly causedy a hyperexcitory amygdala, may lead to initialypertrophy of the amygdala and hippocampusnd to damaging effects in later life. Daily life, withts social demands and constantly changing situa-ions, causes severe stress in children with autism.

number of studies reported significant correla-ions between enlarged amygdala volumes andncreased anxiety in autism.36,39 Animal modelsave demonstrated that stress initially evokes hy-ertrophy of the amygdala neuron.34 Yet, chronicysregulation of the stress response (also referred

o as allostatic overload) may lead to amygdala andippocampal volume decreases in later life. Thisas been supported by studies on recurrent majorepression, a disorder that is associated withhronic stress, in which volume decreases of themygdala and hippocampus have been shown laten the disease process.40,41 Animal studies haveemonstrated that chronic repeated stress evokesxcitotoxic changes in the hippocampus, resultingn degeneration.42 To conclude, a model of hyper-ctivity-induced structural changes is compatibleith the majority of previous findings on amygdala

nd hippocampus volume in autism, as initialnlargement in childhood and subsequent decline

� 23) Controls (n � 29)

73.4 1637902 � 185139.99.78* 1629.10 � 143.862.49 1593.45 � 157.638.00 4553.31 � 406.638.16* 4588.38 � 366.58

nts (n

108� 19� 20� 40� 44

f amygdala and hippocampal volume may result

AL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY

VOLUME 49 NUMBER 6 JUNE 2010

dlptoiwdh

tdafiimdesapa

AMYGDALA-HIPPOCAMPAL ENLARGEMENT IN AUTISM

in either normalized or decreased volumes inadulthood.

Future combined functional and structuralneuroimaging studies should investigate to whatextent the parallel enlargement of the amygdalaand hippocampus reflects the developmental ad-aptation of the autistic brain to a continuousoverflow of emotional learning experiences. Thiscould explain hallmark deficits in autism such asavoidance of eye contact and social withdrawal,social fear, high anxiety levels, agitated behaviorin response to sudden changes, and overarousalin response to particular sensory stimuli. Like-wise, further research is needed to establishwhether these abnormalities of the amygdala andhippocampus are caused by a primary disregu-lation within these limbic structures or by defi-cient top-down regulation from the prefrontalareas of the brain.

There are a few limitations to consider wheninterpreting the results of the current study. First,because our autism sample was restricted tohigh-functioning adolescents with autism, ourfindings are not representative of the whole au-tism population. Second, compared with the re-sults from manual tracing methods, the volumeestimates in our study yielded greater hippocam-pal volumes, as has recently been addressed byMorey et al.43 This seems to be the case for otherautomated segmentation procedures as well. Ithas also been shown, however, that manualsegmentation and automated segmentation inFreesurfer are equally sensitive for detectinggroup volume differences. Thus, as the greatervolume estimate is present in the autism group aswell as in the control group, the finding of groupdifferences are valid. Third, the cross sectionaldesign limits the inferences that we can makeregarding an abnormal developmental pattern ofthe amygdala and hippocampus in adolescentswith autism. As interindividual variation in brain

morphology is probably many times larger than

hippocampus in young children with autism spectrum disorder.AJNR Am J Neuroradiol. 2007;28:672-677.

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY

VOLUME 49 NUMBER 6 JUNE 2010

evelopmental changes within an individual,44

ongitudinal studies are necessary to confirm theroposed abnormal developmental trajectory of

he amygdala and hippocampus. Establishmentf brain–behavior relationships through concom-

tant functional and structural resonance imagingill provide promising opportunities to eluci-ate the neurobiologic underpinnings of thiseterogenic disorder.5-17,20-22

In conclusion, we report an enlargement ofhe amygdala and hippocampus in a well-efined sample of high-functioning adolescentsged 12 through 18 years with autism. Ourndings support a chronically increased activ-

ty of these brain structures in early develop-ent, evoking initial overgrowth of the amyg-

ala and hippocampus in childhood thatxtends into adolescence. Future studieshould focus on the relationship between thectivity and volume of the amygdala and hip-ocampus over time in subjects with autismnd matched controls. &

Accepted January 6, 2010.

Drs. Groen, Teluij, Buitelaar, and Tendolkar are with the DondersInstitute for Brain Cognition and Behavior, Centre for CognitiveNeuroimaging, and with Radboud University Nijmegen MedicalCentre, Nijmegen. Drs. Groen and Buitelaar are with Karakter, Childand Adolescent Psychiatry University Center.

The authors would like to thank Dr. Mark Rijpkema for his invaluableadvice.

Disclosure: Dr. Buitelaar has served as a consultant and on theadvisory board for Shire, Janssen Cilag, Eli Lilly, Pfizer, Organon,UCB, Servier, and Otsuka. He has served on the speakers’ bureau forJanssen Cilag and Eli Lilly. He has received research support fromShire. Drs. Groen, Teluij, and Tendolkar report no biomedical financialinterests or potential conflicts of interest.

Correspondence to Dr. Wouter B. Groen: Department of Psychiatry,Radboud University Nijmegen Medical Centre, Reinier Postlaan 10,PO-Box 9101, 6500 HB Nijmegen, The Netherlands; e-mail: w.groen@psy.umcn.nl

0890-8567/$36.00/©2010 American Academy of Child andAdolescent Psychiatry

DOI: 10.1016/j.jaac.2009.12.023

REFERENCES1. Lord C, Cook EH, Leventhal BL, Amaral DG. Autism spectrum

disorders. Neuron. 2000;28:355-363.2. Amaral DG. The primate amygdala and the neurobiology of

social behavior: implications for understanding social anxiety.Biol Psychiatry. 2002;51:11-17.

3. Lopez JF, Akil H, Watson SJ. Neural circuits mediating stress. BiolPsychiatry. 1999;46:1461-1471.

4. Phillips ML, Drevets WC, Rauch SL, Lane R. Neurobiology ofemotion perception I: the neural basis of normal emotion percep-tion. Biol Psychiatry. 2003;54:504-514.

5. Dager SR, Wang L, Friedman SD, et al. Shape mapping of the

6. Nicolson R, DeVito TJ, Vidal CN, et al. Detection and mapping ofhippocampal abnormalities in autism. Psychiatry Res. 2006;148:11-21.

7. Rojas DC, Peterson E, Winterrowd E, et al. Regional gray mattervolumetric changes in autism associated with social and repeti-tive behavior symptoms. BMC Psychiatry. 2006;6:56.

8. Nacewicz BM, Dalton KM, Johnstone T, et al. Amygdala volumeand nonverbal social impairment in adolescent and adult maleswith autism. Arch Gen Psychiatry. 2006;63:1417-1428.

9. Palmen SJ, Durston S, Nederveen H, van Engeland H. Noevidence for preferential involvement of medial temporal lobe

structures in high-functioning autism. Psychol Med. 2006;36:827-834.

559www.jaacap.org

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

4

4

GROEN et al.

10. Rojas DC, Smith JA, Benkers TL, et al. Hippocampus and amyg-dala volumes in parents of children with autistic disorder. Am JPsychiatry. 2004;161:2038-2044.

11. Schumann CM, Hamstra J, Goodlin-Jones BL, et al. The amygdalais enlarged in children but not adolescents with autism; thehippocampus is enlarged at all ages. J Neurosci. 2004;24:6392-6401.

12. Herbert MR, Ziegler DA, Deutsch CK, et al. Dissociations ofcerebral cortex, subcortical and cerebral white matter volumes inautistic boys. Brain. 2003;126;1182-1192.

13. Bigler ED, Tate DF, Neeley ES, et al . Temporal lobe, autism, andmacrocephaly. AJNR Am J Neuroradiol. 2003;24:2066-2076.

14. Sparks BF, Friedman SD, Shaw DW, et al. Brain structuralabnormalities in young children with autism spectrum disorder.Neurology. 2002;59:184-192.

15. Pierce K, Muller RA, Ambrose J, et al. Face processing occursoutside the fusiform ‘face area’ in autism: evidence from func-tional MRI. Brain. 2001;124:2059-2073.

16. Saitoh O, Karns CM, Courchesne E. Development of the hip-pocampal formation from 2 to 42 years: MRI evidence of smallerarea dentata in autism. Brain. 2001;124:1317-1324.

17. Haznedar MM, Buchsbaum MS, Wei TC, Hof PR, et al. Limbiccircuitry in patients with autism spectrum disorders studied withpositron emission tomography and magnetic resonance imaging.Am J Psychiatry. 2000;157:1994-2001.

18. Howard MA, Cowell PE, Boucher J, et al. Convergent neuroana-tomical and behavioural evidence of an amygdala hypothesis ofautism. Neuroreport. 2000;11:2931-2935.

19. Abell F, Krams M, Ashburner J, et al. The neuroanatomy ofautism: a voxel-based whole brain analysis of structural scans.Neuroreport. 1999;10:1647-1651.

20. Aylward EH, Minshew NJ, Goldstein G, et al. MRI volumes ofamygdala and hippocampus in non-mentally retarded autisticadolescents and adults. Neurology. 1999;53:2145-2150.

21. Piven J, Bailey J, Ranson BJ, Arndt S. No difference in hippocam-pus volume detected on magnetic resonance imaging in autisticindividuals. J Autism Dev Disord. 1998;28:105-110.

22. Saitoh O, Courchesne E, Egaas B, Lincoln AJ, Schreibman L.Cross-sectional area of the posterior hippocampus in autisticpatients with cerebellar and corpus callosum abnormalities. Neu-rology. 1995;45:317-324.

23. Phillips ML, Drevets WC, Rauch SL, Lane R. Neurobiology ofemotion perception II: implications for major psychiatric disor-ders. Biol Psychiatry. 2003;54:515-528.

24. Baron-Cohen S, Ring HA, Bullmore ET, et al. The amygdalatheory of autism. Neurosci Biobehav Rev. 2000;24:355-364.

25. Kemper TL, Bauman ML. The contribution of neuropathologicstudies to the understanding of autism. Neurol Clin. 1993;11:175-187.

26. Van Eijndhoven P, van Wingen G, van Oijen K, et al. Amygdala

volume marks the acute state in the early course of depression.Biol Psychiatry. 2009; 65:812-819.

JOURN

560 www.jaacap.org

7. Schumann CM, Amaral DG. Stereological analysis of amygdalaneuron number in autism. J Neurosci. 2006;26:7674-7679.

8. Fischl B, Van der Kouwe A, Destrieux C, et al. Automaticallyparcellating the human cerebral cortex. Cereb Cortex. 2004;14:11-22.

9. Lord C, Rutter M, Le Couteur A. Autism Diagnostic Interview–Revised: a revised version of a diagnostic interview for caregiversof individuals with possible pervasive developmental disorders. JAutism Dev Disord. 1994;24:659-685.

0. Fernández G, Effenberger O, Vinz B, et al. Hippocampal malfor-mation as a cause of familial febrile convulsions and subsequenthippocampal sclerosis. Neurology. 1998;50:909-917.

1. Pitkanen A, Tuunanen J, Kalviainen R, et al. Amygdala damage inexperimental and human temporal lobe epilepsy. Epilepsy Res.1998;32:233-253.

2. Fischl B, Salat DH, Busa E, et al. Whole brain segmentation:automated labeling of neuroanatomical structures in the humanbrain. Neuron. 2002;33:341-355.

3. Giedd JN, Vaituzis AC, Hamburger SD, et al. Quantitative MRI ofthe temporal lobe, amygdala, and hippocampus in normal humandevelopment: ages 4-18 years. J Comp Neurol. 1996;366:223-230.

4. Chattarji S, Vyas A, Mitra R, Rao BS. Effects of chronic unpredict-able and immobilization stress on neuronal plasticity in the ratamygdala and hippocampus. Soc Neurosci Abs. 2000;26:1533.

5. Maguire EA, Gadian DG, Johnsrude IS, et al. Navigation-relatedstructural change in the hippocampi of taxi drivers. Proc NatlAcad Sci U S A. 2000;97:4398-4403.

6. Juranek J, Filipek PA, Berenji GR, et al. Association betweenamygdala volume and anxiety level: magnetic resonance imaging(MRI) study in autistic children. J Child Neurol. 2006;21:1051-1058.

7. Hull AM. Neuroimaging findings in post-traumatic stress disor-der—systematic review. Br J Psychiatry. 2002;181:102-110.

8. Adami P, Konig P, Vetter Z, Hausmann A, Conca A. Post-traumatic stress disorder and amygdala-hippocampectomy. ActaPsychiatr Scand. 2006;113:360-363.

9. Munson J, Dawson G, Abbott R, et al. Amygdalar volume andbehavioral development in autism. Arch Gen Psychiatry. 2006;63:686-693.

0. Sheline YI, Wang PW, Gado MH, et al. Hippocampal atrophy inrecurrent major depression. Proc Natl Acad Sci U S A. 1996;93:3908-3913.

1. Sheline YI, Gado MH, Price JL. Amygdala core nuclei volumes aredecreased in recurrent major depression. Neuroreport. 1998;9:2023-2028.

2. McEwen BS. Stress and hippocampal plasticity. Annu Rev Neu-rosci. 1999;22:105-122.

3. Morey RA, Petty CM, Xu Y, et al. A comparison of automatedsegmentation and manual tracing for quantifying hippocampaland amygdala volumes. Neuroimage. 2009;45:855-866.

4. Caviness VS, Kennedy DN, Richelme C, Rademacher J, Filipek

PA. The human brain age 7-11 years: a volumetric analysis basedon magnetic resonance images. Cereb Cortex. 1996;6:726-736.

AL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY

VOLUME 49 NUMBER 6 JUNE 2010