Timing of conditioned eyeblink responses is impaired in children with attention-deWcit/hyperactivity disorder
Markus Frings · Kristina Gaertner · Paul Buderath · Marcus Gerwig · Hanna Christiansen · Beate Schoch · Elke R. Gizewski · Johannes Hebebrand · Dagmar Timmann
Abstract Structural changes of the cerebellum have beenreported in several psychiatric diseases like schizophrenia,autism and attention-deWcit/hyperactivity disorder (ADHD).Beside behavioral deWcits children with ADHD often showslight motor abnormalities. Cerebellar malfunction maycontribute. The cerebellum is a structure essential for motorcoordination, various forms of motor learning and timing ofmotor responses. In the present study, eyeblink condition-ing was applied to investigate learning and timing of motorresponses both in children with ADHD and children withcerebellar lesions. Acquisition, timing and extinction ofconditioned eyeblink responses were investigated in chil-dren with ADHD, children with chronic surgical cerebellarlesions and controls using a standard delay paradigm withtwo diVerent interstimulus intervals. Timing of conditionedeyeblink responses was signiWcantly impaired in children
with ADHD in the long interstimulus interval condition.Also in children with cerebellar lesions conditionedresponses (CR) tended to occur earlier than in controls.Incidences of CRs were signiWcantly reduced in childrenwith cerebellar lesions and tended to be less in childrenwith ADHD than in controls. Extinction of the CRs wasimpaired in children with cerebellar lesions in both inter-stimulus interval conditions and in children with ADHD inthe long interstimulus interval condition. Cerebellar mal-function may contribute to disordered eyeblink condition-ing in ADHD. However, because CR abnormalities diVeredbetween ADHD and cerebellar subjects, dysfunction ofnon-cerebellar structures cannot be excluded.
Structural and functional changes of the cerebellum havebeen demonstrated in psychiatric disorders such asautism and schizophrenia (Andreasen et al. 1994, 1996;Courchesne et al. 1994). In attention-deWcit/hyperactivitydisorder (ADHD) a reduced volume of the cerebellum(Castellanos et al. 2001) and especially the vermis (Berquinet al. 1998; Castellanos et al. 2001; Mostofsky et al. 1998)has been described, too. ADHD is deWned by three clinicalfeatures which include inattention, motor hyperactivity andimpulsivity and aVects approximately 3–5% of elementaryschool-aged children (Szatmari et al. 1989). The diagnosticand statistical manual of mental disorders (DSM-IV-TR)identiWes three subtypes, the predominantly inattentivetype, the predominantly hyperactive-impulsive type, and
M. Frings (&) · K. Gaertner · P. Buderath · M. Gerwig · D. TimmannDepartment of Neurology, University of Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germanye-mail: [email protected]
H. Christiansen · J. HebebrandDepartment of Child and Adolescent Psychiatry and Psychotherapy, University of Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germany
B. SchochDepartment of Neurosurgery, University of Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germany
E. R. GizewskiDepartment of Diagnostic and Interventional Radiology and Neuroradiology, University of Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germany
the combined type (American Academy of Psychiatry2000).
A role of the cerebellum in psychiatric diseases has beendiscussed in the last three decades. The traditional viewabout the cerebellum is that it is involved in the coordina-tion of movements of the extremities, the trunk, the speechmusculature, and the eyes (Holmes 1939). Moreover,results of numerous studies indicate that the cerebellum iscrucially involved in implicit motor learning such as asso-ciative and non-associative learning of aversive reXexes,adaptation of voluntary movements and learning of com-plex motor skills (Bloedel and Bracha 1995; Doyon et al.2003). Aside from behavioral deWcits, many children andadolescents with ADHD also exhibit minor motor abnor-malities (“clumsiness”) (Karatekin et al. 2003; Tervo et al.2002). The question arises, what aspects of these minormotor abnormalities derive from changes of the cerebellum.In multifocal disorders like ADHD with an involvement offrontal lobe, basal ganglia and thalamus, causal associa-tions are diYcult to infer from the correlation betweenchanges in individual regions of the brain and behavior(Gowen and Miall 2007). For that reason, the present studyexamined if abnormalities of motor learning in childrenwith ADHD are comparable to those deWcits found in chil-dren with cerebellar lesions.
Cerebellar involvement in motor learning has been exam-ined in greatest detail in classical conditioning of the eye-blink response (Gerwig et al. 2007). Conditioning of theeyeblink response is a simple task used in animals andhumans. A corneal airpuV as an unconditioned stimulus (US)is delivered to the eye, eliciting a reXexive blink as an uncon-ditioned response (UR). If a tone as a conditioned stimulus(CS) repeatedly precedes the airpuV, subjects learn to blinkthe eye to the tone as a conditioned response (CR) so that theeye is already closed when the airpuV is delivered. Inhumans, lesions of the superior parts of the cerebellum, par-ticularly of lobules HVI and Crus I impair acquisition of con-ditioned eyeblink responses (Gerwig et al. 2003).
The cerebellum is also involved in temporal processingof eyeblink conditioning. If the interval between condi-tioned and US is prolonged the CR occurs later, too (McG-linchey-Berroth et al. 1999). This precise timing is crucialto protect the eyes optimally from the airpuV. In animalstudies cortical cerebellar lesions caused shortenedlatencies of CRs (Koekkoek et al. 2003; Perrett et al. 1993).A study in patients with cortical cerebellar degenerationand cerebellar infarctions showed that lesions includinglobule V were related to timing deWcits in eyeblink condi-tioning (Gerwig et al. 2005).
So far, only one study has examined eyeblink condition-ing in children with ADHD (CoYn et al. 2005). These chil-dren showed no impairment in the acquisition ofconditioned eyeblinks. Conditioned answers had an earlier
onset and peak time latency than controls, but these resultswere not signiWcant. In this study only one short CS–USinterstimulus interval was applied. Eyeblink conditioningparadigms with diVerent interstimulus intervals are helpfulto show true timing deWcits of conditioned eyeblink reX-exes (e.g. Koekkoek et al. 2003). If there is a cerebellardependent timing deWcit it should appear at various timeintervals. Accordingly, in a previous study in patients withlesions of the superior cerebellum, CRs occurred signiW-cantly earlier in patients than in controls and cerebellarpatients showed a timing deWcit in two interstimulus inter-val conditions (Gerwig et al. 2005).
The aim of the present study was to investigate if cerebel-lar dysfunction contributes to motor abnormalities in ADHD.For that reason eyeblink conditioning in both children withADHD and children with cerebellar lesions was compared tocontrols. To assess possible timing deWcits a paradigm withtwo diVerent CS–US interstimulus intervals has beenapplied. CS–US interstimulus intervals were used in accor-dance with a previous study in humans from McGlinchey-Berroth et al. (1999). Volumes of the cerebellum in childrenwith ADHD were quantiWed to determine possible structuralabnormalities. Lesion sites in cerebellar children wereassessed based on 3D magnetic resonance images (MRI).
Methods
Subjects
The study was performed in ten male subjects that matchedthe DSM-IV-TR (American Academy of Psychiatry 2000)diagnostic criteria for the combined subtype of ADHD(mean age 12.3 § 1.34 (10–15) years). All children withADHD were treated with methylphenidate at the time oftesting. Moreover, seven children with chronic surgical cer-ebellar lesions following astrocytoma resection wereincluded (4 male, 3 female; mean age 12.3 § 2.5 (7–15)years.). None of the cerebellar children had receivedchemotherapy or cranial radiation. Mean age of controlsubjects was 12.1 § 1.8 (10–15) years (9 male, 2 female).None of the patients or controls had a deWcit in visualcapacities and audiometry was normal. All subjects andlegal representatives gave informed written consent. Thestudy was approved by the local ethical committee and hastherefore been performed in accordance with the ethicalstandards laid down in the 1964 declaration of Helsinki.
Clinical motor assessment
The neurological examination according to the ataxia ratingscale of Trouillas et al. (1997) revealed mild signs ofcerebellar ataxia (total ataxia score, <10 of 100) in four
cerebellar children and moderate signs of cerebellar ataxia(total ataxia score, 10–20) in 3. Mean total ICARS-scorewas 8.4 (range 1–14) in the cerebellar group and 1.5 (range0–10) in the ADHD group. None of the control subjectsshowed total ICARS-scores >0. DiVerences in ICARSbetween cerebellar children and ADHD children(p = 0.016) and between cerebellar children and controlsubjects were signiWcant (p = 0.001; one-way ANOVA).DiVerences between ADHD children and controls did notreach signiWcance (p = 0.068).
Neuropsychological assessment
IQ testing (short version of the German WISC, information,picture arrangement, similarities and block-design) wasperformed in all subjects (Sattler 1992). Mean estimated IQwas 101.8 § 16.2 in the ADHD group, 94.6 § 9.1 in thecerebellar group and 110.3 § 9.2 in the control group(p = 0.127; oneway ANOVA). Parents and teachers com-pleted Conners’ questionnaires for the assessment ofADHD symptoms (Conners 1997). ADHD subjects showedabnormal high values in Conners’ scale for parents andteachers at the time of diagnosis (mean Conners’ parentstotal T-score 69 § 11 (abnormal >60), mean Conners’teachers total T-score 69 § 11). One cerebellar subjectshowed increased values (Conners’ parents total T-score69, Conners’ teachers total T-score 73), but mean scores ofcerebellar children were in the normal range (Conners’ par-ents total T-score 52 § 8, Conners’ teachers total T-score52 § 10). Also in all control subjects values were in thenormal range (Conners’ parents total T-score 48 § 5, Con-ners’ teachers total T-score 48 § 7).
Brain imaging
For each subject high-resolution 3D T1-weightedMPRAGE scans were obtained on a Siemens Sonata 1.5-TMR scanner (TR 2,400 ms, TE 4.38 ms, FOV 256 mm, 160slices, voxel size 1.0 £ 1.0£1.0 mm3). In the control andADHD children volumetric analysis of MR images wasperformed semi-automatically with the help of ECCETsoftware (http://www.eccet.de/). Total cerebellar volume,total cerebral volume (cerebellum excluded) and total intra-cranial volume (TICV) were assessed as described previ-ously (Dimitrova et al. 2006). Two examiners performedvolumetric assessment independently for every subject(P.B. and K.G.). Interrater reliabilities were good (intra-class correlation coeYcients, ICCs >0.8) and mean valueswere calculated. Cerebellar and cerebral volumes were bothexpressed as percentage TICV and corrected for bodyheight (volume/body height).
In the cerebellar children surgical lesions were tracedmanually in non-normalized 3D MRI data sets and saved as
regions of interest (ROIs) using MRIcro software (http://www.mricro.com). The individual ROIs and complete 3DMRI data sets were simultaneously spatially normalizedinto a standard proportional stereotaxic space using SPM2(http://www.fil.ion.ucl.ac.uk/spm/). For more detailedinformation see Konczak et al. (2005). Based on the MNIspatial coordinates of cerebellar lesions, the aVectedcerebellar lobules and nuclei were deWned with the help of3D MRI atlases of the cerebellum (Dimitrova et al. 2002;Schmahmann et al. 2000).
Eyeblink conditioning
For eyeblink conditioning the standard delay protocoldeveloped by Gormezano and Kehoe (1975) was used. AnairpuV (intensity 400 kPa at source, 110 kPa at nozzle)served as the US and was directed through a nozzle near theouter canthus of the eye at a distance of about 10 mm. TheCS consisted of a tone with 70 dB sound pressure level andwas presented ipsilaterally. To mask environmental noise,all subjects wore headsets. In addition, the CS was superim-posed on a continuous white noise of 60 dB applied bilater-ally. Participants were set comfortably in a chair with theireyes open. A silent movie (“Mr. Bean” from Rowan Atkin-son) was shown using a DVD player to maintain vigilanceand attention. Children with unilateral cerebellar lesionsreceived the US on the aVected side. In one cerebellar childwith bilateral lesions, the more aVected right side wasinvestigated. In all ADHD children and controls the rightside was tested.
A short (440 ms) and a long (840 ms) CS–US interval wasused (McGlinchey-Berroth et al. 1999). Duration of the USwas 100 ms co-terminating with the CS. Accordingly, the CSlasted 540 ms in the short and 940 ms in the long intervalcondition. The CS consisted of a 1 kHz tone in the short andof a 4 kHz tone in the long interval condition. Half of eachpatients and controls started with the long and half with theshort CS–US interval. The time between the short and longinterval condition was 60 min. During this time participantsreceived the clinical neurological examination includingassessment of the ataxia rating scale. In both conditions,pseudorandomly presented Wve CS-alone and Wve US-alonetrials were followed by Wve blocks of 10 paired CS–US trialswith Wve CS-alone trials interspersed. At the end Wve CS-alone extinction trials were presented. Interstimulus intervalsfor all conditions varied between 20 and 35 s.
Surface EMG recordings were taken from orbicularisoculi muscles ipsilaterally to the CS und US. Signals werefed to EMG ampliWers (sampling rate 1,000 Hz, band passWlter frequency between 100 Hz and 2 kHz), full wave rec-tiWed and further Wltered oZine (100 Hz).
EMG recordings of the paired CS § US trials were ana-lysed on a trial-by-trial basis using commercial software
(Axograph V4 for Macintosh, Axon Instruments, UnionCity, CA, USA). EMG activity lasting at least 50 ms ormerging into superimposed UR of at least twice the ampli-tude of mean EMG baseline activity and clear rising slopewas deWned as CR. The time window for detection of CRwas from 150 ms after CS onset to the onset of the US.Responses earlier than 150 ms after CS onset were regardedas reXexive responses to the tone (i.e., alpha responses) andnot conditioning-related (WoodruV-Pak et al. 1996). Onsetand peak time latencies of conditioned eyeblink responseswithin the CS–US interval were visually identiWed. CRonset was marked at the earliest point at which EMG activ-ity began to rise from the pre-CS EMG baseline level.Onset was expressed as time of CR onset before onset ofthe air puV (US) and time-to-peak as time of maximumamplitude observed before onset of the airpuV. The time ofthe airpuV was set as 0 ms. Frequency of spontaneousblinks was estimated by counting responses within an inter-val of 1 min before and after the experiment.
A univariate repeated measures analysis was calculatedusing number of CRs, onset or peak time of CRs as depen-dent variable, block number (5 blocks of ten successivepaired CS–US trials) and interstimulus interval (short vs.long interstimulus interval) as within-subject factors andgroup (cerebellar vs. ADHD vs. control group) as between-subject factor. In post-hoc analysis a univariate repeatedmeasures analysis was calculated separately comparingADHD and control group, cerebellar and control group andADHD and cerebellar group.
Results
Brain imaging
Absolute cerebellar and cerebral volume did not diVerbetween ADHD subjects and controls (cerebellum: ADHD:151.1 § 10.3 cm3, controls: 148.2 § 10.8 cm3; cerebrum:
ADHD: 1331.6 § 71.6 cm3, controls: 1384.6 § 92.9 cm3;P values >0.18, one-way ANOVA). ADHD subjectsrevealed a signiWcantly larger cerebellar volume correctedfor TICV compared to controls (ADHD: 9.9 § 0.7%; con-trols: 8.89 § 0.4%; F(1,17) = 15.85, P < 0.001). Cerebel-lar volumes corrected for body height, however, did notdiVer between groups (F(1,17 = 2.62, P = 0.12). Cerebralvolume did not signiWcantly diVer between groups (cor-rected for TICV: ADHD: 77.2 § 2.5%; controls:74.1 § 4.2%; F(1,17) = 3.50, P = 0.079; corrected for bodyheight: F(1,17) = 0.02, P = 0.90).
Analysis of normalized 3D-MRI data enabled us todeWne the extent of cerebellar cortical and nuclear impair-ment in the cerebellar children. Table 1 summarizes whichcerebellar lobules were aVected and whether or not the deepcerebellar nuclei were aVected. Except for one child with avermal lesion three children had a unilateral right and threechildren a unilateral left lesion. In two children the lesionincluded hemispheral lobule V which is thought to contrib-ute to the timing of the conditioned eyeblink response inhumans (Gerwig et al. 2005) and in three children lesionsinvolved hemispheral lobule VI which has been related tothe acquisition of CRs (Gerwig et al. 2003). The centre ofoverlap was located in the inferior cerebellum within par-avermal lobule VIIIB (n = 6). Figure 1 shows the lesions ofall cerebellar children superimposed on axial MRIs of thecerebellum of a healthy subject. All unilateral lesions aresuperimposed on the left cerebellar hemisphere with right-sided lesions being Xipped to the left. Lines within the cere-bellum indicate Wssures of hemispheral lobules V and VI.
Classical eyeblink conditioning—incidence of conditioned eyeblink responses
All groups showed an increase of CR-incidences across theWve blocks, i.e. learning (block eVect F(4,100) = 3.30,P = 0.014). In the long CS–US time interval the CR-inci-dence was higher than in the short CS–US time interval
Table 1 AVected lobules and nuclei in patients with cerebellar lesions
Anatomical identiWcation of aVected cerebellar lobules was performed according to the atlas of Schmahmann et al. (2000) and identiWcation ofcerebellar nuclei according to the MRI atlas of the human cerebellar nuclei of Dimitrova et al. (2006); l left; r right; b bilateral
(interstimulus interval eVect F(1,25) = 3.32, P < 0.001).Mean total percentage CR-incidences were smallest in thecerebellar subjects, followed by the ADHD subjects andhighest in the controls for both the short and long CS–UStime interval (Fig. 2). (group eVect F(2,25) = 2.51,P = 0.1). Repeated measures ANOVA revealed no signiW-cant block by group, interstimulus interval by group orinterstimulus interval by block interaction eVects (P values>0.27).
In post-hoc analysis group eVect was signiWcant compar-ing the control and cerebellar group (F(1,16) = 5.29,P = 0.035), but not comparing the ADHD and controlgroup (F(1,19) = 1.23, P = 0.28) or comparing the ADHDand cerebellar group (F(1,15) = 1.57, P = 0.23). In post-hocanalysis repeated measures ANOVA revealed no signiW-cant block by group, interstimulus interval by group or byblock interaction eVects (P values >0.21).
Timing of conditioned eyeblink responses
Comparison of the short and long CS–US interval conditionrevealed that the time between CS and response was longerin the long compared with the short CS–US interval condi-tion for all groups (interstimulus interval eVect for onsettime F(1,25) = 98.20, P < 0.001 and peak timeF(1,25) = 89.95, P < 0.001; Fig. 3). Group eVect and inter-stimulus interval by group eVect did not reach signiWcancefor both onset and peak time (P values >0.18).
However, in post-hoc analysis comparing controls andchildren with ADHD timing of CRs was impaired in chil-dren with ADHD, i.e., onset and peak time of the CR in thelong CS–US time interval occurred earlier in the childrenwith ADHD than in controls (group by interstimulus inter-val eVect for onset time F(1,19) = 6.04, P = 0.024 and peaktime F(1,19) = 6.08, P = 0.024; group eVect P values
>0.14). Onset and peak time also occurred earlier in cere-bellar children than in controls but diVerences were not sig-niWcant (group by interstimulus interval eVect for onset andpeak time P values >0.34; group eVect P values >0.30).Comparison of the cerebellar and ADHD groups showed
Fig. 1 Lesion sites of seven children with cerebellar lesions superimposed on axial MRI scans of a healthy subject. For superimposition of the individ-ual stereotaxically normalized cerebellar lesions, right-sided lesions were Xipped to the left. Lines within the cerebellum indicate Wssures of hemispheral lobules V and VI
Fig. 2 Mean § SD percentage of incidences of conditioned responsesfor each of the Wve blocks and across all blocks (total) in the short andlong CS–US interval conditions in the group of all control subjects(black triangles), ADHD children (gray circles) and cerebellar chil-dren (white squares)
123
172 Exp Brain Res (2010) 201:167–176
no group or group by interstimulus interval eVect for onsetor peak time (P values >0.33).
In two of seven children the lesion included hemispherallobule V which is thought to contribute to the timing of theconditioned eyeblink response (Gerwig et al. 2005). Directcomparison of means of onset time between these patientsand controls showed that responses occurred earlier in chil-dren with lesions including hemispheral lobule V, i.e., thesechildren showed a timing deWcit in the short interstimulusinterval (mean § SD ¡158.6 § 76.9 ms vs. ¡113.9 §46.6 ms) but not in the long interstimulus interval(mean § SD ¡270 § 122 ms vs. ¡283.9 § 69.9 ms).
Extinction
At the end of the experiment, Wve CS-alone trials were pre-sented. To determine eVects of extinction CRs in the blockof extinction trials were compared with the last block ofpaired acquisition trials (Fig. 4). Children with ADHD andcontrols showed a decrease of CRs in extinction trials com-pared with the last block of paired acquisition trials, butcerebellar children showed an increase of CRs in extinctiontrials, i.e., cerebellar children showed no extinction (blockby group eVect F(2,25) = 3.39, P = 0.05). The decrease ofCRs in controls and the increase of CRs in cerebellar chil-dren were stronger in the long than in the short CS–USinterstimulus interval condition. ADHD children showed aweaker decrease of CRs than controls in the long interstim-ulus interval condition, i.e., ADHD children showedimpaired extinction in the long interstimulus interval condi-tion (interstimulus interval by block by group eVectF(2,25) = 4.19, P = 0.027). For all groups number of CRswas higher in the long interstimulus interval than in theshort interstimulus interval condition (F(1,25) = 4.96,P = 0.035). Group, block, interstimulus interval by groupand by block eVects did not reach signiWcance (P values>0.3).
Post-hoc analysis was performed by comparing cerebel-lar and control group, ADHD and control group and cere-bellar and ADHD group. Comparison of cerebellar childrenand controls revealed a signiWcant block by group eVectconWrming a lack of extinction in cerebellar children(F(1,16) = 7.97, P = 0.012; block, group, interstimulusinterval and interaction eVects P values >0.12). Compari-son of children with ADHD and controls showed a signiW-cant interstimulus interval by block by group eVect(F(1,19) = 6.06, P = 0.024), i.e., extinction of CRs wasimpaired in children with ADHD compared with controls inthe long CS–US interstimulus interval condition. Compari-son of the cerebellar and ADHD groups showed no group,block, interstimulus interval or interaction eVects (P values>0.08).
Interspersed CS-alone trials
Figure 5 shows a reduced number of CRs in the Wveinterspersed CS-alone trials of the long compared to theshort CS–US interstimulus interval condition for theADHD children and in both conditions for the cerebellarchildren. However, univariate analysis did not revealsigniWcant group or interstimulus interval eVects (P val-ues >0.11).
Fig. 3 Timing parameters of conditioned eyeblink responses in theshort and long CS–US interstimulus interval condition. Group meanand SD values shown for conditioned response peak latenciesexpressed in time (ms) before onset of the unconditioned stimulus (airpuV) set as 0 ms
Fig. 4 Mean § SD percentage of incidences of conditioned responsesin the blocks of extinction trials compared with the last block of pairedacquisition trials in the short and long CS–US interval conditions
Incidence of alpha-responses (short-latency responses) ear-lier than 150 ms after CS onset was higher in the short com-pared with the long CS–US interstimulus interval condition(F(1,25) = 5.14, P = 0.032). However, group and group byinterstimulus interval eVects were not signiWcant (P values>0.18). The frequency of spontaneous blinks, i.e., the num-ber of responses within an interval of 1 min before and afterthe experiment was not diVerent between groups or inter-stimulus intervals (P values >0.49).
Discussion
The present study showed various impairments of eyeblinkconditioning in children with ADHD and in children withfocal cerebellar disorders. Although eyeblink conditioningwas impaired in both groups, the pattern of abnormalitieswas not the same. First, incidence of CRs was signiWcantlyimpaired n children with cerebellar lesions and tended to beless in children with ADHD than in controls. Second, CRsoccurred signiWcantly earlier in the long interstimulus inter-val condition in the ADHD group than in controls, but tim-ing abnormalities did not reach signiWcance in the cerebellargroup. Finally, extinction of CRs was impaired in cerebellarchildren in short and long interstimulus intervals and inADHD children in the long interstimulus interval.
MRI volumetry
In contrast to previous studies no reduced volume of thecerebellum could be demonstrated in the group ofchildren with ADHD compared to controls. However, pre-vious studies did not show consistent results, either. Areduced volume of the whole cerebellum in ADHD hasbeen described (Castellanos et al. 2002). Other studiesfound a reduced volume especially of the vermis (Berquinet al. 1998; Castellanos et al. 2001; Mostofsky et al.1998). In other developmental disorders like autismresults are unclear, too. DiVerent studies have found thatthe vermis of the cerebellum is either abnormally smallor abnormally large (Courchesne et al. 1994). Anotherreason for the lack of evidence for a reduced cerebellarvolume in the present study may be that the number ofpatients is too small. Previous studies demonstratingchanges of the cerebellum have been performed in largerpatient groups.
Incidence of conditioned eyeblink responses
Incidence of conditioned eyeblink responses wasimpaired in children with cerebellar lesions and was lessin children with ADHD than in controls. However, diVer-ences between the ADHD and control group did not reachsigniWcance. The only previous study that has examinedeyeblink conditioning in children with ADHD (CoYnet al. 2005) showed no diVerences of eyeblink condition-ing rates in children with ADHD and in controls. In bothstudies children with ADHD were treated with methyl-phenidate at the time of testing. Methylphenidate is themost commonly prescribed stimulant drug for ADHD andhas demonstrated positive eVects on inattention andhyperactivity, but also on motor skills like gait or hand-writing (Leitner et al. 2007; Lerer et al. 1977; Rubia et al.2003). The eVect of medication might have hidden eVectsof impaired CR acquisition in a comparatively smallgroup of children with ADHD.
Timing of conditioned eyeblink responses
Timing of conditioned eyeblink responses was impaired inchildren with ADHD in the long interstimulus interval con-dition. CRs occurred earlier in cerebellar children than incontrols, but not signiWcantly. The lack of signiWcantresults in contrast to previous studies in adults with cerebel-lar disease (Gerwig et al. 2005) and previous animal studieswith cerebellar lesions (Koekkoek et al. 2003; Perrett et al.1993) may be explained by the population of children withcerebellar lesions in the present study. In the study byGerwig et al. (2005) a voxel-by-voxel based analysis ofMRI lesion data revealed a relation of timing deWcits with
Fig. 5 Mean § SD percentage of incidences of conditioned responsesin the interspersed CS-alone trials of the short and long CS–US intervalconditions
lesions including hemispheral lobule V of the anterior lobeof the cerebellum. In the present study, only in two patientsthe lesion involved hemispheral lobule V, whereas the cen-tre of overlap was located in the inferior cerebellum withinparavermal lobule VIIIB. Direct comparison of means ofonset time between the two patients with lesions involvinglobule V and controls showed that responses occurred ear-lier in these patients in the short interstimulus interval butnot in the long interstimulus interval. Another reason forthe lack of signiWcance comparing timing in cerebellar chil-dren and controls may be that the Wndings are based on thefew remaining CRs in a small group of cerebellar childrenwith signiWcantly reduced incidences of CRs. Therefore,direct comparison between CR timing disorders in theADHD group with the present group of cerebellar childrenis limited and it appears warranted to refer to data in adultpatients with cerebellar lesions instead.
The Wndings in children with ADHD are in accordancewith a study of eyeblink conditioning in spontaneouslyhypertensive rats, an animal model of ADHD (Chess andGreen 2008), which showed shortened latencies of CRs,too. In the previous study by CoYn et al. (2005) in childrenwith ADHD conditioned answers had an earlier onset andpeak time latency than controls, but these results were notsigniWcant. Only the time from the onset to the peak of theCR was signiWcantly shorter in the ADHD group com-pared to controls. However, this is only an indirect param-eter of a timing deWcit. Studies in cerebellar patients and inanimals with cerebellar lesions showed both signiWcantlyshorter onset and peak time latencies (Gerwig et al. 2005;Koekkoek et al. 2003; Perrett et al. 1993). To detect possi-ble deWcits of patients with ADHD in changing the onset ofthe CR, in the present study a paradigm with two diVerentinterstimulus intervals was used (McGlinchey-Berrothet al. 1999). Onset and peak time of the CR in the long butnot in the short CS–US time interval occurred earlier in thepatients with ADHD than in controls. Maybe, timingdeWcits only become evident in the longer interstimulusinterval, which has not been applied in the study by CoYnet al. (2005).
Braitenberg (1967) was the Wrst to note that the role ofthe cerebellum may be judging the timing of events. Sub-jects with cerebellar lesions are impaired in the discrimina-tion of the duration of the time interval between pairs oftones (Ivry and Keele 1989) and in estimating the velocityof dots moving on a screen (Ivry and Diener 1991). TheWndings of the present study suggest that timing deWcits inADHD are to some extent eVects of cerebellar dysfunction,maybe as part of a fronto-striato-cerebellar network (Yanget al. 2007).
In the present study, however, timing deWcits wereobserved only in the long but not in the short interstimulus
interval condition in children with ADHD. In a previousstudy of our group with a larger group of cerebellarpatients a timing deWcit was found in a short and longinterstimulus interval condition (Gerwig et al. 2005). Onemay argue, that this discrepancy between patients withcerebellar lesions and ADHD suggests that dysfunction ofnon-cerebellar structures causes timing deWcits in eye-blink conditioning in ADHD. In addition to the cerebel-lum, frontal and striatal structures likely contribute to thepathophysiology of ADHD (Bush et al. 2005). Motor tim-ing studies suggest that frontal dysfunction plays a majorrole in timing deWcits of patients with ADHD, especiallyin longer time intervals (e.g., Radonovich and Mostofsky2004). However, timing of eyeblink conditioning has notbeen examined in humans with frontal lesions so far. Inpatients with Huntington’s disease, which mainly aVectsthe striatum timing of conditioned eyeblink reXexes wasimpaired, too (WoodruV-Pak and Papka 1996). Temporalprocessing in Parkinson’s disease seems to be also pre-dominantly impaired in longer time intervals (e.g. Kochet al. 2008). Therefore, it cannot be excluded that disor-ders in frontal cortex or basal ganglia may contribute tothe observed deWcits in timing of conditioned eyeblinkresponses in ADHD.
Extinction
Repeated presentation of trials with conditioned stimulialone leads to extinction of the previously learned CR. Inthe present study, extinction was impaired in children withcerebellar lesions in the long und short interstimulus inter-val condition and in children with ADHD in the long inter-stimulus interval condition. This Wnding is in accordancewith the animal study of Chess and Green (2008), whichshowed an impaired extinction in spontaneously hyperten-sive rats, the animal model of ADHD, too. However,extinction of the CRs was impaired in children with ADHDonly in the long interstimulus interval condition and in chil-dren with cerebellar lesions in both interstimulus intervalconditions. As outlined above, one may argue that non-cer-ebellar structures play a role in extinction of eyeblink con-ditioning in children with ADHD. On the other hand,disorders in ADHD children may extend to the short inter-stimulus interval condition in a larger group of unmedi-cated ADHD children.
In conclusion, children with ADHD show signiWcantabnormalities in timing and extinction of conditioned eye-blink responses. Cerebellar malfunction may contribute.However, because the pattern of eyeblink conditioningabnormalities diVered between ADHD and cerebellar sub-jects, dysfunction of non-cerebellar dysfunction cannot beexcluded.
Acknowledgments The study was supported by an institutional re-search grant from the University of Duisburg-Essen (IFORES-pro-gram D/107-20180, D/107-20170). We thank B. Brol for technicalassistance.
References
American Psychiatric Association (2000) Diagnostic and StatisticalManual of Mental Disorders text revision (DSM-IV-TR), 4th edn.American Psychiatric Press Inc, Washington, DC
Andreasen NC, Flashman L, Flaum M, Arndt S, Swayze V, O’LearyDS, Ehrhardt JC, Yuh WT (1994) Regional brain abnormalities inschizophrenia measured with magnetic resonance imaging. J AmMed Assoc 272:1763–1769
Andreasen NC, O’Leary DS, Cizadlo T, Arndt S, Rezai K, Boles PontoLL, Watkins GL, Hichwa RD (1996) Schizophrenia and cognitivedysmetria: a positron-emission tomography study of dysfunc-tional prefrontal–thalamic–cerebellar circuitry. Proc Natl AcadSci USA 93:9985–9990
Berquin PC, Giedd JN, Jacobsen LK, Hamburger SD, Krain AL,Rapoport JL, Castellanos FX (1998) Cerebellum in attention-deWcit hyperactivity disorder: a morphometric MRI study.Neurology 50:1087–1093
Bloedel JR, Bracha V (1995) On the cerebellum, cutaneomuscularreXexes, movement control and the elusive engrams of memory.Behav Brain Res 68:1–44
Braitenberg V (1967) Is the cerebellar cortex a biological clock in themillisecond range. Progr Brain Res 25:334–346
Bush G, Valera EM, Seidman LJ (2005) Functional neuroimaging ofattention-deWcit/hyperactivity disorder: a review and suggestedfuture directions. Biol Psychiatry 57:1273–1284
Castellanos FX, Giedd JN, Berquin PC, Walter JM, Sharp W, Tran T,Vaituzis AC, Blumenthal JD, Nelson J, Bastain TM, Zijdenbos A,Evans AC, Rapoport JL (2001) Quantitative brain magneticresonance imaging in girls with attention-deWcit/hyperactivitydisorder. Arch Gen Psychiatry 58:289–295
Castellanos FX, Lee PP, Sharp W, JeVries NO, Greenstein DK, ClasenLS, Blumenthal JD, James RS, Ebens CL, Walter JM, ZijdenbosA, Evans AC, Giedd JN, Rapoport JL (2002) Developmentaltrajectories of brain volume abnormalities in children andadolescents with attention-deWcit/hyperactivity disorder. JAMA288:1740–1748
Chess AC, Green JT (2008) Abnormal topography and altered acquisi-tion of conditioned eyeblink responses in a rodent model of atten-tion-deWcit/hyperactivity disorder. Behav Neurosci 122:63–74
CoYn JM, Roody S, Schneider K, O’Neill J (2005) Impaired cerebellarlearning in children with prenatal alcohol exposure: a compara-tive study of eyeblink conditioning in children with ADHD anddyslexia. Cortex 41:389–398
Conners CK (1997) Conners’ rating scales—revised CRS–R. MHS,North Tonawanda
Courchesne E, Saitoh O, Townsend JP, Yeung-Courchesne R, PressGA, Lincoln AJ, Haas RH, Schriebman L (1994) Cerebellarhypoplasia and hyperplasia in infantile-autism. Lancet 343:60–64
Dimitrova A, Weber J, Redies C, Kindsvater K, Maschke M, Kolb FP,Forsting M, Diener HC, Timmann D (2002) MRI atlas of the hu-man cerebellar nuclei. Neuroimage 17:240–255
Dimitrova A, Zeljko D, Schwarze F, Maschke M, Gerwig M, Frings M,Beck A, Aurich V, Forsting M, Timmann D (2006) Probabilistic3D MRI atlas of the human cerebellar dentate/interposed nuclei.Neuroimage 30:12–25
Doyon J, Penhune V, Ungerleider LG (2003) Distinct contribution ofthe cortico-striatal and cortico-cerebellar systems to motor skilllearning. Neuropsychologia 41:252–262
Gerwig M, Dimitrova A, Kolb FP, Maschke M, Brol B, Kunnel A,Böring D, Thilmann AF, Forsting M, Diener HC, Timmann D(2003) Comparison of eyeblink conditioning in patients withsuperior and posterior inferior cerebellar lesions. Brain126:71–94
Gerwig M, Hajjar K, Dimitrova A, Maschke M, Kolb FP, Frings M,Thilmann AF, Forsting M, Diener HC, Timmann D (2005) Tim-ing of conditioned eyeblink responses is impaired in cerebellarpatients. J Neurosci 25:3919–3931
Gerwig M, Kolb FP, Timmann D (2007) The involvement of thehuman cerebellum in eyeblink conditioning. Cerebellum 6:38–57
Gormezano I, Kehoe EJ (1975) Classical conditioning: some method-ical-conceptual issues. In: Estes WK (ed) Handbook of learningand cognitive processes, vol 2. Lawrence Erlbaum, Hillsdale(NJ), pp 143–179
Gowen E, Miall RC (2007) The cerebellum and motor dysfunction inneuropsychiatric disorders. Cerebellum 6:268–279
Holmes G (1939) The cerebellum of man (Hughlings Jackson memo-rial lecture). Brain 62:1–30
Ivry RB, Diener HC (1991) Impaired velocity perception in patientswith lesions of the cerebellum. J Cogn Neurosci 3:355–366
Ivry RB, Keele SW (1989) Timing functions of the cerebellum. J CognNeurosci 1:136–152
Karatekin C, Markiewicz SW, Siegel MA (2003) A preliminary studyof motor problems in children with attention-deWcit/hyperactivitydisorder. Percept Mot Skills 97:1267–1280
Koch G, Costa A, Brusa L, Peppe A, Gatto I, Torriero S, Gerfo EL,Salerno S, Oliveri M, Carlesimo GA, Caltagirone C (2008)Impaired reproduction of second but not millisecond time inter-vals in Parkinson’s disease. Neuropsychologia 46:1305–1313
Koekkoek SK, Hulscher HC, Dortland BR, Hensbroek RA, ElgersmaY, Ruigrok TJ, De Zeeuw CI (2003) Cerebellar LTD and learningdependent timing of conditioned eyelid responses. Science301:1736–1739
Konczak J, Schoch B, Dimitrova A, Gizewski E, Timmann D (2005)Functional recovery of children and adolescents after cerebellartumour resection. Brain 128:1428–1441
Leitner Y, Barak R, Giladi N, Peretz C, Eshel R, Gruendlinger L,HausdorV JM (2007) Gait in attention deWcit hyperactivity disor-der: eVects of methylphenidate and dual tasking. J Neurol254:1330–1338
Lerer RJ, Lerer MP, Artner J (1977) The eVects of methylphenidate onthe handwriting of children with minimal brain dysfunction.J Pediatr 91:127–132
McGlinchey-Berroth R, Brawn C, Disterhoft JF (1999) Temporal dis-crimination learning in severe amnesic patients reveals an alter-ation in the timing of eyeblink conditioned responses. BehavNeurosci 113:10–18
Radonovich KJ, Mostofsky SH (2004) Duration judgments in childrenwith ADHD suggest deWcient utilization of temporal informationrather than general impairment in timing. Child Neuropsychol10:162–172
Rubia K, Noorloos J, Sergeant JA (2003) Motor timing deWcits in com-munity and clinical boys with hyperactive behavior: the eVect ofmethylphenidate on motor timing. J Abnorm Child Psychol31:301–313
Sattler J (1992) Assessment of children: WISC III and WPPSI-Rsupplement. San Diego, CA
Schmahmann JD, Dojon J, Toga AW, Petrides M, Evans AC (2000)MRI atlas of the human cerebellum. Academic Press, San Diego
123
176 Exp Brain Res (2010) 201:167–176
Szatmari P, OVord DR, Boyle MH (1989) Ontario Child Health Study:prevalence of attention deWcit disorder with hyperactivity. J ChildPsychol Psychiatry 30:219–230
Tervo RC, Azuma S, Fogas B, Fiechtner H (2002) Children withADHD and motor dysfunction compared with children withADHD only. Dev Med Child Neurol 44:383–390
Trouillas P, Takayanagi T, Hallett M, Currier RD, Subramony SH,Wessel K, Bryer A, Diener HC, Massaquoi S, Gomez CM,Coutinho P, Ben Hamida M, Campanella G, Filla A, Schut L,Timann D, Honnorat J, Nighoghossian N, Manyam B (1997)
International Cooperative Rating Scale for pharmacologicalassessment of the cerebellar syndrome. J Neurol Sci 145:205–211
WoodruV-Pak DS, Papka M (1996) Huntington’s disease and eyeblinkclassical conditioning: normal learning but abnormal timing. J IntNeuropsychol Soc 2:323–334
WoodruV-Pak DS, Papka M, Ivry RB (1996) Cerebellar involvementin eyeblink classical conditioning in humans. Neuropsychology10:443–458
Yang B, Chan RC, Zou X, Jing J, Mai J, Li J (2007) Time perceptiondeWcit in children with ADHD. Brain Res 1170:90–96
