Impaired induction of long-term potentiation-like plasticity in patients with high-functioning autism and Asperger syndrome



    1.  Department of Paediatrics, Technical University Munich, Munich;
    2.  Division of Neuropaediatrics and Muscular Disorders, Department of Paediatrics and Adolescent Medicine, University Hospital Freiburg, Freiburg;
    Search for more papers by this author

    1.  Division of Neuropaediatrics and Muscular Disorders, Department of Paediatrics and Adolescent Medicine, University Hospital Freiburg, Freiburg;
    Search for more papers by this author

    1.  Division of Neuropaediatrics and Muscular Disorders, Department of Paediatrics and Adolescent Medicine, University Hospital Freiburg, Freiburg;
    2.  European Neuroscience Institute Göttingen, Göttingen;
    Search for more papers by this author

    1.  Department of Neurology, University Medical Centre Hamburg-Eppendorf (UKE), Hamburg;
    Search for more papers by this author

    1.  Department of Child and Adolescent Psychiatry and Psychotherapy, University Hospital Freiburg, Freiburg, Germany
    Search for more papers by this author

    1.  Division of Neuropaediatrics and Muscular Disorders, Department of Paediatrics and Adolescent Medicine, University Hospital Freiburg, Freiburg;
    Search for more papers by this author

    1.  Department of Neurology, University Medical Centre Hamburg-Eppendorf (UKE), Hamburg;
    Search for more papers by this author

    1.  Department of Child and Adolescent Psychiatry and Psychotherapy, University Hospital Freiburg, Freiburg, Germany
    Search for more papers by this author

    1.  Department of Paediatrics, Technical University Munich, Munich;
    2.  Division of Neuropaediatrics and Muscular Disorders, Department of Paediatrics and Adolescent Medicine, University Hospital Freiburg, Freiburg;
    Search for more papers by this author

Professor Dr Volker Mall at Kinderzentrum München gemeinnützige GmbH, Heiglhofstrasse 63, 81377 München, Germany. E-mail:


Aim  We aimed to investigate the induction of long-term potentiation (LTP)-like plasticity by paired associative stimulation (PAS) in patients with high-functioning autism and Asperger syndrome (HFA/AS).

Method  PAS with an interstimulus interval between electrical and transcranial magnetic stimulation of 25 ms (PAS25) was performed in patients with HFA/AS (n=9; eight males, one female; mean age 17y 11mo, SD 4y 5mo) and in typically developing age-matched volunteers (n=9; five males, four females; mean age 22y 4mo, SD 5y 2mo). The amplitude of motor-evoked potentials was measured before PAS25, immediately after stimulation, and 30 minutes and 60 minutes later. A PAS protocol adapted to individual N20 latency (PASN20+2) was performed in six additional patients with HFA/AS. Short-interval intracortical inhibition was measured using paired-pulse stimulation.

Results  In contrast to the typically developing participants, the patients with HFA/AS did not show a significant increase in motor-evoked potentials after PAS25. This finding could also be demonstrated after adaptation for N20 latency. Short-interval intracortical inhibition of patients with HFA/AS was normal compared with the comparison group and did not correlate with PAS effect.

Interpretation  Our results show a significant impairment of LTP-like plasticity induced by PAS in individuals with HFA/AS compared with typically developing participants. This finding is in accordance with results from animal studies as well as human studies. Impaired LTP-like plasticity in patients with HFA/AS points towards reduced excitatory synaptic connectivity and deficits in sensory-motor integration in these patients.


Autism spectrum disorder


γ-aminobutyric acid


High-functioning autism


High-functioning autism and Asperger syndrome


Interstimulus interval


Long-term potentiation


Methyl CpG binding protein 2


Motor-evoked potential


Maximum stimulator output


Paired associative stimulation


Short-interval intracortical inhibition


Transcranial magnetic stimulation

What this paper adds

  •  Significant insights into the pathophysiology of ASDs.
  •  New findings of synaptic plasticity in developmental disorders.
  •  In vivo evidence for impaired induction of long-term potentiation-like plasticity through paired associative stimulation in patients with ASDs.

Autism spectrum disorders (ASDs) are severe neurodevelopmental disorders characterized by a qualitative impairment of social interaction and communication, repetitive stereotyped patterns of behaviour, and restricted interests and activities.1 These are accompanied by gross and fine motor impairment in both high-functioning autism (HFA) and Asperger syndrome.2–4 Motor impairments range from unspecific ‘clumsiness’ to Parkinsonian-like movement patterns, indicating an involvement of the central nervous structures responsible for intact motor behaviour and motor learning.5,6

Strengthening of synaptic connections by long-term potentiation (LTP) in animals and LTP-like plasticity in humans has been shown to be a basic mechanism of motor learning.7,8 LTP deficits in animal models of autism were recently identified. Mice with mutations in the methyl CpG binding protein 2 (MECP2) show impaired LTP induction and deficits in learning and memory.9,10 In humans, MECP2 mutations cause Rett syndrome, and reduced expression of MECP2 in the frontal lobe cortex has been described in patients with ASD.11 MECP2 protein regulates the transcription of the gene encoding for brain-derived neurotrophic factor, which is strongly involved in induction of LTP-like plasticity.12 A reduced excitatory synaptic connectivity in MECP2 knockout mice precedes deficits in neuronal plasticity.9 It has also been demonstrated that a dysfunction in γ-aminobutyric acid (GABA) signalling mediates autism-like stereotypies13 and seems to play a role in the pathophysiology of HFA.14 MECP2 is critical for normal function of GABA-releasing neurons, and subtle dysfunction of GABAergic neurons could contribute to numerous neuropsychiatric phenotypes such as Rett syndrome.13

Human studies have demonstrated impairments of motor behaviour and motor learning in ASD.2 However, little is known about the underlying mechanisms. The basal ganglia, thalamus, and supplementary motor area have recently been shown to play a pivotal role in motor dysfunction in autism.5 Neuroimaging data revealed diffusely decreased connectivity across motor execution networks involving the primary motor cortex (M1), thalamus, and supplementary motor area.6

Synaptic plasticity in animal models is studied using stimulation patterns in the neurons of the hippocampus, resulting in a potentiation of synaptic transmission (i.e. LTP). In the human M1, excitability can be measured with transcranial magnetic stimulation (TMS). Synaptic plasticity can be induced using stimulation paradigms such as paired associative stimulation (PAS), a protocol adopted from animal experiments inducing associative LTP.15 PAS consists of repeated peripheral electric stimulation paired with TMS applied to the contralateral motor cortex at a defined interstimulus interval (ISI) and requires an intact sensorimotor integration.15 This stimulation leads to an increase in amplitudes of motor-evoked potentials (MEPs)16 with similar properties to LTP in animal models: long-lasting, input-specific, N-methyl-d-aspartate (NMDA)-receptor dependent, rapidly evolving, fully reversible, and bidirectional. Therefore, the MEP increase is generally referred to as LTP-like plasticity because of its similar characteristics to LTP in vitro.17,18

In animal models it has been demonstrated that autism-like behaviour is caused, at least in part, by GABAergic dysfunction.13 Also, from human studies, there is growing evidence for GABAergic abnormalities in ASD.5,14,19 Post-mortem examination found significant reduction in the number of GABA-A receptors in certain brain areas of patients with autism,20 and in vivo studies found a significant reduction of short-interval intracortical inhibition (SICI) in patients with high-functioning autism and Asperger syndrome (HFA/AS).14

To test the hypothesis that in patients with ASD LTP-like plasticity is impaired because of abnormal SICI, we investigated LTP-like plasticity (using PAS) and GABA-mediated SICI (using an established paired-pulse protocol21,22) in patients with HFA/AS.



Fifteen patients with ASD aged 15 to 29 years (mean age 17y 10mo, SD 3y 6mo; one female, 14 males) and a diagnosis of HFA (n=7), Asperger syndrome (n=6), or pervasive developmental disorder not otherwise specified according to the International Classification of Diseases, 10th revision (ICD-10)23 and DSM-IV2 criteria (n=2; both male) were recruited and diagnosed by an experienced child and adolescent psychiatrist at an outpatient clinic of the University Hospital Freiburg, Germany (for detailed description of clinical data, see Table I). IQ scores of patients with ASD ranged from 83 to 135 (mean 110.9, SD 14.9) using the Culture Fair Intelligence Test (CFT 20–R) test,24 Wechsler Intelligence Scale for Children 3rd and 4th editions (WISC-III, WISC-IV), Wechsler Intelligenztest für Erwachsene (WIE), or Coloured Progressive Matrices (CPM) test. Diagnostic and behavioural assessments were performed using the Autism Diagnostic Observation Schedule25 and the revised version of the Autism Diagnostic Interview.26 These assessments were performed by a trained child and adolescent psychiatrist. Nine patients (one female, eight males) with ASD (HFA [n=6]; Asperger syndrome [n=3]) aged 15 to 29 years (mean age 17y 11mo, SD 4y 5mo;) were measured with an ISI between electrical and transcranial magnetic stimulation of 25 ms (PAS25), and six patients with ASD aged 16 to 20 years (mean 17y 8mo, SD 1y 7mo; six males) with HFA (n=1), Asperger syndrome (n=3), or pervasive developmental disorder not otherwise specified (n=2) were measured with PASN20+2, a PAS protocol adapted to the individual N20 latency of each patient. The control group comprised nine age-matched typically developing volunteers aged 16 to 29 years (mean age 22y 4mo, SD 5y 4mo; four females, five males) who were screened for mental illness, developmental disorders, or other neurological or internal disorders. Most patients (13/15) and all the comparison group were right-handed according to the Edinburgh Handedness Inventory. Because there seems to be no fundamental difference between the right and left hemisphere when inducing plasticity, we did not expect left-handedness in two patients to be a confounding factor.27 None of the participants fulfilled any exclusion criteria for the safety of TMS28 or took centrally acting medication at the time of measurements. The study was approved by the local ethics committee of the University of Freiburg, Germany and was conducted according to the latest version of the Declaration of Helsinki. After full disclosure of the purpose and risks of the study procedure, all participants gave their written informed consent. Parents gave written informed consent to participation for study volunteers aged under 18 years.

Table I. Clinical data of participants with autism spectrum disorder
IDSexAge (y)DiagnosisaIQIQ testADOS communication behaviourADOS social behaviourADI-R communication behaviourADI-R social interaction
  1. aAccording to DSM-IV and ICD-10 criteria. ID, identity; ADOS, Autism Diagnostic Observation Schedule; ADI-R, Autism Diagnostic Interview Revised; F, female; M, male; CFT 20-R, Culture Fair intelligence Test; HFA, high-functioning autism; WISC-III, WISC-IV; Wechsler Intelligence Scale for Children 3rd and 4th editions; PDD-NOS, pervasive developmental disorder not otherwise specified; CPM, Coloured Progressive Matrices; WIE, Wechsler Intelligenztest für Erwachsene.

1F18Asperger110CFT 20-R6131310
2M29HFA135CFT 20-R
3M18Asperger106CFT 20-R471325
4M17HFA83CFT 20-R44418
5M16HFA132CFT 20-R451126
6M15Asperger120CFT 20-R251419
7M18HFA122CFT 20-R35
8M15HFA104CFT 20-R581925
9M15HFA 82CFT 20-R77717
13M16Asperger115CFT 20-R48910

Electromyographic recording

Participants were seated comfortably in a chair reposing both hands suitably on a cushion or their lap to ensure complete relaxation. MEPs were recorded by electromyography (EMG) from the non-dominant abductor pollicis brevis muscle using Ag–AgCl surface electrodes (surface area 263mm2; AMBU, Ballerup, Denmark) mounted using the belly-tendon technique. Participants were asked to relax the target muscle during measurement. Data were bandpass filtered (20–2000Hz) and amplified using an Ekida DC universal amplifier (Ekida, Helmstadt, Germany) connected to a Micro 1401 mkII data acquisition unit (Cambridge Electronic Design, Cambridge, UK) with a sampling rate of 5kHz and stored on a personal computer for online visual display and later offline analysis using Signal software version 3 (Cambridge Electronic Design). MEP size was determined by measuring the two highest peaks of opposite polarity29 and then averaged over 20 trials for both single and conditioned pulses.

Transcranial magnetic stimulation

For focal TMS, the intersection of an eight-shaped stimulation coil with an outer wing diameter of 90mm was centred tangentially on the scalp over the right M1 of the abductor pollicis brevis, the handle of the stimulation coil pointing from dorsolateral to lateral approximately 45° in relation to the midline. This induced a current in a posterior–anterior direction, roughly perpendicular to the presumed line of the central sulcus for activating the corticospinal neurons transsynaptically.30,31 The coil was connected to a Magstim 200 Stimulator (Magstim Company, Whitland, UK) with a monophasic current waveform. We determined the optimal position for eliciting MEPs of maximum amplitude from the target muscle by moving the coil over M1 while administering stimuli of slightly suprathreshold intensity. Magnetic stimuli were administered to the cortex at a frequency of 0.1Hz, stimuli used to identify hotspot and motor thresholds were administered at 0.25Hz. We determined the resting motor threshold using a maximum-likelihood threshold-hunting procedure.32 We used 16 TMS stimuli starting at 45% of maximum stimulator output (MSO). A positive MEP was defined as a potential larger than 50μV in peak-to-peak amplitude. The initial stimulator output for evaluation was chosen to elicit a mean MEP amplitude of 800 to 1200μV (SI1mV) and was then kept constant throughout the investigations to assess changes in MEPs.

Optical navigation

The optimal coil position for eliciting MEPs of the targeted muscle was recorded using a stereotaxic, optically tracked navigation system, consisting of a camera (Polaris Vicra P6; NDI, Waterloo, Ontario, Canada), custom-made software (BrainView; Fraunhofer Institute [IPA], Stuttgart, Germany), and passive sphere markers as previously described.33 The position sensor of the camera emitted infrared light reflected by four passive sphere markers mounted on the head of the participant using a frontlet, and five additional passive sphere markers fixed to the coil. For co-registration of head and coil, anatomical landmarks (nasion, external corners of the eyes, both ears) were defined by the investigator at the beginning of the study. The position of the coil was related to the head of the participant using BrainView software, which displayed a three-dimensional figure of the head and coil in real time. The chosen position of the coil in relation to the head was maintained throughout the measuring procedure.

Somatosensory-evoked potentials

To obtain somatosensory-evoked potentials, we used a Viking-system (CareFusion, San Diego, CA, USA). Patients were seated comfortably in an examination chair in an electrically shielded room. Electrical stimuli (4.7Hz; duration 0.2 ms) were consecutively applied to both median nerves at the wrist. Active recording electrodes were placed at the head surface contralateral to the stimulus side (C3 or C4 according to the International 10-20 system), the reference electrode was placed cephalic (Fpz according to the International 10-20 system). Data were bandpass-filtered with cut-off frequencies of 10 to 1500Hz. Two hundred responses were averaged and two runs of somatosensory-evoked potentials were obtained in each patient to ensure reproducibility. Peak latencies of N20 and P25 and peak-to-peak amplitudes of N20–P25 were measured. Somatosensory-evoked potentials were considered abnormal if latency of N20 differed more than 2 ms or peak-to-peak amplitude was more than two-fold larger/smaller than the contralateral side.

Paired associative stimulation

PAS consisted of 200 peripheral and central stimulation pairs at a frequency of 0.25Hz of peripheral electric stimulation of the left median nerve at the wrist, followed by TMS of M1 at the optimal site to elicit MEPs in the abductor pollicis brevis muscle of the non-dominant hemisphere. To induce LTP-like plasticity, we used the PAS protocol adapted by Ziemann et al.18,34,35 Electrical stimulation was applied through a Digitimer DS7 electrical stimulator (Digitimer, Welwyn Garden City, Hertfordshire, UK) using a bipolar electrode with the cathode placed proximally. We identified the optimal stimulation site at the wrist, fixed the electrode, and then determined the threshold of sensory perception. During PAS, constant current square-wave pulses with a duration of 1000μs were applied at an intensity of three times the perceptual threshold. TMS intensity was set to produce MEP peak-to-peak amplitudes of about 1mV (SI1mV). The same intensity was used throughout the experiment for post-PAS evaluation. The ISI between electrical and TMS was 25 ms (PAS25ms) in nine patients and in the comparison participants. This interval induces a robust LTP-like increase of MEP amplitudes in typically developing adults.15,16,34 We adapted the ISI in six patients by their individual N20 latency plus 2 ms (PASN20+2), according to previously published data.36 Because attention level may influence the PAS effect’s magnitude,37 we constantly reminded all participants to focus their attention on the stimulated hand and ensured this by asking them to count the number of applied electrical stimuli.34 The PAS effect was determined on the mean of all time points after PAS compared with baseline as described elsewhere.34 As the application of many TMS stimuli might influence subsequent induction of plasticity,34 we limited the number of TMS stimuli before PAS to a maximum of 200.

Short interval intracortical inhibition

We measured SICI using the Kujirai21 paired-pulse protocol. A conditioning stimulus with an intensity of 80% of resting motor threshold preceded a test stimulus with SI1mV. ISIs were 2 ms and 3 ms. Conditioned and unconditioned stimuli were applied in a randomized order at a frequency of 0.25Hz. The SICI ratio was calculated as the quotient of conditioned and unconditioned MEPs.

Experimental procedures

We recorded MEP amplitudes and resting motor threshold before (pre) as well as at three time points after (post) PAS25 and PASN20+2 (post 1, 0min; post 2, 30min; post 3, 60min). SICI was measured at least 24 hours after PAS25 to avoid potential interactions.

Data analysis

We ensured there was a comparable status of relaxation in the recorded muscle during the TMS measurements by investigating offline the EMG waveform of every trial. Pre-facilitated trials (baseline EMG >50μV) were excluded from further analyses.

Trials beyond three times the SD of the mean MEP amplitude of one measurement were considered outliers and excluded from further analyses. As this criterion was not met, no MEP had to be excluded. We computed all statistical analyses using SPSS software, version 15.0 (SPSS Inc., Chicago, IL, USA). We used the same evaluation as in many TMS articles dealing with PAS (see, for example, Delvendahl et al.,34 Ziemann et al.,35 Stefan et al.38) assuming a normal distribution of MEPs. In advance, data were tested for normality by the Kolmogorov–Smirnov test (no violation of the assumption of normality was found). In addition, data were graphically normal distributed. However, these results are limited to the small sample size of the study. Statistical evaluation was performed using repeated-measure analysis of variance (ANOVA) with group (comparison and ASD) as the between-participant factor and time (pre, post 1, post 2, and post 3) as the within-participant factor. If necessary, we used the Greenhouse–Geisser correction to adjust for violations of sphericity, resulting in adjusted p-values based on adjusted degrees of freedom. In case of significant main effects or interactions, we used post-hoc two-tailed paired or unpaired t-tests with Bonferroni correction for multiple comparisons. This method was used for MEP and resting motor threshold data. To analyse baseline parameters of PAS and SICI, we used two-tailed paired t-tests. Correlation between SICI and PAS effect was performed using a Pearson’s correlation coefficient. Significance level was set at α=0.05. All values given are mean group values and SD, if not indicated otherwise.


All participants completed the study procedure and we observed no adverse events. There was no significant difference in baseline MEP amplitudes between groups undergoing PAS25 protocol (ASD: 0.95, SD 0.11mV; comparison: 1.00, SD 0.14mV; p=0.364). Also, resting motor threshold (ASD: 42.77, SD 6.59mV; comparison: 39.66, SD 6.76mV; p=0.338) and perceptual threshold (ASD: 0.56, SD 0.18mA; comparison: 0.66, SD 0.19mA; p=0.563) did not differ between the two groups.

Repeated-measures ANOVA on untransformed MEP data revealed significant main effects of group (F1,3.014=5.847, p=0.042) and time (F3,1.531=3.221, p=0.041) and an interaction group × time (F3,1.532=4.257, p=0.015; Fig. 1) in PAS25 measurements. One-way repeated-measures ANOVA revealed a significant effect of time (F3,8=4.501, p=0.012) in MEP in the comparison group. Post-hoc testing showed significant increase in MEP amplitudes after PAS25 in the comparison participants (MEP pre: 1.00, SD 0.14; MEP post 1: 1.26, SD 0.47; MEP post 2: 1.64, SD 0.92; MEP post 3: 1.71, SD 0.61; post hoc t-tests: pre– post 1: p=0.107; pre– post 2: p=0.063; pre– post 3: p=0.01) but not in patients (MEP pre: 0.94, SD 0.10; MEP post 1: 1.07, SD 0.28; MEP post 2: 0.99, SD 0.47; MEP post 3: 0.97, SD 0.37; post hoc t-tests: pre– post 1: p=0.150; pre– post 2: p=0.775 pre– post 3: p=0.814). Differences between both groups were significant at the maximum of the MEP increase at time point post 3 (p=0.007, unpaired t-test). Resting motor threshold did not change significantly after PAS stimulation both in patients with ASD (pre: 42.77, SD 6.59 MSO; post 1: 42.33, SD 7.65 MSO; post 2: 42.44, SD 9.70 MSO; post 3: 43.33, SD 7.73 MSO) and the participants without ASD (pre: 39.66, SD 6.76 MSO; post 1: 39.00, SD 6.40 MSO; post 2: 38.88, SD 7.11 MSO; post 3: 38.00, SD 7.19 MSO).

Figure 1.

 Induction of long-term potentiation (LTP)-like plasticity through paired associative stimulation (PAS). Repeated-measures analysis of variance on untransformed motor-evoked potential (MEP) data revealed significant main effects of group and time and interaction group × time. Time point post 3 differed significantly (p=0.007, unpaired t-test) between typically developing participants and patients with autism spectrum disorder (ASD) measured with PAS25. Means of untransformed data are displayed. Error bars, standard error of the mean. Pre, TMS baseline investigation.

We observed no significant increase in MEP amplitude in six patients with ASD for PASN20+2 measurements (Fig. 1). Repeated-measures ANOVA on untransformed MEP data revealed no significant main effect of time (F3,15=1.138, p=0.366). Resting motor threshold did not change significantly after PASN20+2 stimulation (pre: 48.5, SD 8.26 MSO; post 1: 50.8, SD 9.54 MSO; post 2: 49.3, SD 7.97 MSO; post 3: 50.5, SD 7.18 MSO). Repeated-measures ANOVA revealed no significant main effect of time (F3,15=0.959, p=0.438).

Together with PASN20+2 measurements, we observed no pathological somatosensory-evoked potential in patients with ASD. Mean N20 latency was 19.89 ms (SD 1.14; range 18.70–21.90) for the right cortex and 20.02 ms (SD 1.07, range 19.00–21.80) for the left cortex. Mean ISI for PASN20+2-stimulation was 22.00 ms (SD 1.26).

SICI ratios did not differ significantly either for the 2 ms or 3 ms intervals (unpaired t-test; p=0.327 and p=0.534 respectively) between patients with ASD and typically developing participants, as shown in Figure 2. We observed no significant correlation between PAS effect and SICI ratio in ASD (2 ms: r=−0.159; p=0.683 and 3 ms: r=−0.238; p=0.537).

Figure 2.

 Short-interval intracortical inhibition (SICI) displayed as a ratio of unconditioned and conditioned motor-evoked potential (MEP) for 2 ms and 3 ms intervals in autism spectrum disorder (ASD) and the typically developing comparison group (TC). SICI did not differ significantly from an aged-matched comparison group (unpaired t-test; p=0.327 and p=0.534 respectively).


We found an impaired induction of LTP-like plasticity through PAS in patients with ASD compared with typically developing participants. This was also the case when the ISI of PAS was adapted to the individual N20 latency. In our sample, GABAergic intracortical inhibition as a potential underlying pathomechanism was not impaired in patients with ASD. Our results support the hypothesis that patients with ASD have impaired LTP-like plasticity. This is in line with previous findings from animal and human studies.

Impairment of motor behaviour and motor learning in ASD has been reported not only in animal but also in human studies, although the exact underlying mechanisms still remain unknown. Genetic causes, synaptic dysfunction, and reduced excitatory synaptic connectivity as well as an altered GABAergic inhibition have been discussed as possible pathophysiological mechanisms.4,11,14 We aimed to investigate the induction of LTP-like plasticity by PAS and found it impaired in patients with ASD.

Because the induction of associative plasticity as induced by PAS depends critically on the timing between the pulse induced by peripheral (electrical) sensory stimulation and TMS of the motor cortex, integrity of the sensorimotor system may play a central role.15 Therefore, we analysed somatosensory-evoked potentials and found normal N20 latencies in all of our patients. Furthermore, we adapted the ISI of electrical and peripheral stimulation of PAS to the individual N20 latency and found similar results as with a fixed latency. However, these findings do not exclude dysfunction of the sensorimotor system as a reason for impaired induction through PAS in HFA/AS. Indeed, it has been shown that electroencephalography and neuroimaging data give strong evidence of a reduced function and connectivity of neuronal networks involving the basal ganglia, thalamus, supplementary motor area, and M1 in patients with ASD5,6 that may not be excluded by normal N20 latencies. Because these areas seem to play a key role in synaptic plasticity induced by PAS,39 reduced function and connectivity may be one reason for our findings. This may also explain different findings in the literature. Studying effects of theta burst on motor cortex plasticity in five patients, Oberman et al.40 found that the increase of MEPs lasts longer than in typically developing participants. The effect of direct motor cortex stimulation on MEPs is unlikely to be affected on connectivity of neuronal networks between the basal ganglia, thalamus, supplementary motor area, and M1, and therefore may explain the different findings of the two studies. Further, it may be responsible for findings on a behavioural level such as gross and fine motor impairment.2–4

In animal models of autism with underlying gene defects, reduced LTP induction has been demonstrated previously. Mutations of MECP2 have been identified as causing synaptic dysfunction, reduced excitatory synaptic connectivity, and learning and memory deficits in mice.9,10 Although autism is considered polygenetic, a significant reduction in MECP2 expression was found in 79% of patients with ASD11 and, therefore, may play a role in impaired synaptic plasticity in these patients. For our patients, this point remains speculative as our study did not include (epi)genetic analyses.

Attention may be a confounding factor in PAS studies as it has been shown to influence the magnitude of effects.38,41 However, none of the patients reported attention deficit disorders and we ensured that there were high attention levels in all participants throughout the whole experiment. Therefore, it seems unlikely that attention influenced PAS effects in our study.

SICI has been linked to the regulating function of GABAergic cortical interneurons in typically developing humans in the way that increased inhibition blocks the induction of LTP-like plasticity.22,34 In ASD, GABAergic inhibition was reduced in patients with HFA but not in those with Asperger syndrome.14 We found no differences in intracortical inhibition in participants with HFA and Asperger syndrome compared with participants without ASD. Owing to the relatively small sample size of our study, we cannot exclude with certainty an abnormal function of intracortical inhibitory neurons as an underlying cause of abnormal PAS in patients with ASD.

In conclusion, we demonstrated impaired induction of LTP-like plasticity through PAS in patients with HFA/AS. This may be due to synaptic dysfunction and reduced excitatory synaptic connectivity as well as to abnormal sensorimotor integration and decreased connectivity across motor networks, resulting in abnormal motor function and motor learning in these patients.


We thank the volunteers who participated in this study. The study was supported by a grant from Deutsche Forschungsgemeinschaft (MA 3306/4-1).