Children with ASD often exhibit early difficulties with action imitation, possibly due to low-level sensory or motor impairments. Impaired cortical rhythms have been demonstrated in adults with ASD during motor imitation. While those oscillations reflect an age-dependent process, they have not been fully investigated in youth with ASD. We collected magnetoencephalography data to examine patterns of oscillatory activity in the mu (8-13 Hz) and beta frequency (15-30 Hz) range in 14 adolescents with and 14 adolescents without ASD during a fine motor imitation task. Typically developing adolescents exhibited adult-like patterns of motor signals, e.g., event-related beta and mu desynchronization (ERD) before and during the movement and a postmovement beta rebound (PMBR) after the movement. In contrast, those with ASD exhibited stronger beta and mu-ERD and reduced PMBR. Behavioral performance was similar between groups despite differences in motor cortical oscillations. Finally, we observed age-related increases in PBMR and beta-ERD in the typically developing children, but this correlation was not present in the autism group. These results suggest reduced inhibitory drive in cortical rhythms in youth with autism during intact motor imitation. Furthermore, impairments in motor brain signals in autism may not be due to delayed brain development. In the context of the excitation-inhibition imbalance perspectives of autism, we offer new insights into altered organization of neurophysiological networks.
Autism Spectrum Disorder (ASD) is a complex disorder of brain development characterized, in varying degrees, by difficulties in social interaction, verbal and nonverbal communication, and repetitive behaviors [
Voluntary movements are accompanied by changes in cortical rhythms that can be detected by electroencephalography (EEG) and magnetoencephalography (MEG). Distinct oscillatory signals are associated with motor tasks but are differently modulated during movement imitation or observation.
First, movement-related changes in rhythmic activity in the mu-range (8-13 Hz) have been reported as early as infancy [
Second, rhythmic modulation in the ongoing beta (15-30 Hz) rhythm follows a pattern similar to the mu rhythm [
Third, a high-gamma band (~70-90 Hz) ERS is sometimes observed at the onset of movement [
Abnormalities in mu and beta rhythms have been described in ASD patients while performing motor imitation tasks, such as reduced mu-suppression during movement observation [
In this study, we examined mu- and beta-band oscillations in adolescents with ASD during a finger imitation task. The paradigm we chose involved simple finger-lifting imitative movement performed from the
Participants were 28 right-handed adolescents (Table
Participants’ characteristics.
| | | | |
---|---|---|---|---|
| 7 Autistic Disorder | 14 | ||
| ||||
| 14.5 ±2. 8 | 13.8 ±2. 8 | 0.66 | 0.52 |
| ||||
| 13/1 | 11/3 | n/a | 0.16 |
| ||||
| 0.8 ±0.3 | 0.8 ±0.2 | 0.46 | 0.65 |
| ||||
| 106.5 ±19.2 | 110.3 ±15.8 | 0.57 | 0.58 |
| ||||
| 104.6±21.5 | n/a | n/a | n/a |
| ||||
| 48.2±10.5 | 49.2±9.2 | 0.23 | 0.82 |
The stimuli consisted of a photorealistic animated right hand, presented in the third-person perspective (Figure
Right-hand third-person representation showing the hand at rest (left) and while performing an index lift movement (right).
MEG data were obtained in a magnetically shielded room (ETS-Lindgren, Cedar Park, TX, USA) using a Magnes 3600 WH whole-head MEG device (4D Neuroimaging, San Diego, CA, USA), comprised of 248 first-order axial-gradiometer sensors (5 cm baseline) in a helmet-shaped array. Five head position indicator coils attached to the subject’s scalp were used to determine the head position with respect to the sensor array. The locations of the coils with respect to three anatomical landmarks (nasion and preauricular points, with the intersection of the tragus and daith of the ear defining the preauriculars) and 2 extra nonfiducial points as well as the scalp surface (approximately 500 points) were determined with a 3D digitizer (Polhemus, Colchester, VT, USA). The MEG signals were acquired continuously in a 0.1-200 Hz bandwidth and sampled at 678.17 Hz and 24-bit vertical resolution.
Single axis monolithic integrated circuits Leadless Chip Carrier (LCC) accelerometers (model ADXL103; Analog Devices, Inc.) were attached to both index and pinky fingertips in order to precisely quantify movement onset. The chips are wired to approximately 3.3 m of light weight, highly flexible, miniature cable (Cooner Wire NMVF 4/30-4046) with local bypass capacitors (0.1 uf) and encapsulated in heat-shrink. Accelerometer signals were high-pass filtered at 20 Hz, rectified, and then low-pass filtered at 10 Hz in a procedure adapted from the preprocessing of electromyography data for trigger definition [
Each participant’s MEG data were coregistered with structural T1-weighted magnetic resonance imaging (MRI) data prior to source space analyses (see below MRI acquisition procedures) using common landmarks from the MEG digitization procedure and MRI scan data. Structural MRI data were aligned parallel to the anterior and posterior commissures and transformed into the Talairach coordinate system [
MEG postprocessing was performed using BESA 5.3 (MEGIS Software GmbH, Grafelfing, Germany). Artifact-free epochs (mean per condition: 51 +/- 12) in the time-domain were transformed to the time-frequency domain with a 2 Hz/25 ms sampling in BESA using complex demodulation [
Using BESA Research 5.3, cortical networks were imaged through an extension of the linearly constrained minimum variance vector beamformer [
Mean power was extracted between -500 and 500 ms for beta-ERD and 1000 ms for mu-ERD and between 750 and 2000 ms for PMBR. For statistical analyses, time-frequency results were subjected to group statistical analysis in a 2 x 2 x 2 mixed design ANOVA statistical test (group by finger by hemisphere) with finger and hemisphere treated as within-subjects measures. Separate ANOVAs were computed for the ERD and PMBR windows.
MR images and spectra were acquired using a 3.0T GE Signa HDx whole body, long bore MR scanner (GE Healthcare, Waukesha, WI, USA) at the Brain Imaging Center, University of Colorado Denver. Subjects were imaged in the supine position using a GE eight-channel phased array head coil. To comply with age- and population-related behaviors such as boredom and restlessness, subjects watched a movie during the exams using MR-compatible goggles and headphones (Resonance Technology Inc., Northridge, CA, USA) during the procedure. A T1-weighted sequence was acquired for tissue segmentation using a 3D inversion recovery fast, spoiled gradient echo (IR-SPGR) technique (matrix 256 x 256, FOV 22 cm, TR/TE/TI= 10/3/450 ms, NEX=1), resulting in 168 1.2 mm thick axial slices with an in-plane resolution of .86 mm2.
Behavioral analyses of correct responses were determined from the accelerometer data. Correct trials were defined as subject movements occurring on the correctly indicated finger within 3 s after the movement onset from the video displayed on the screen (i.e., from 1 s to 4 s after stimulus presentation). Subjects in both groups failed to respond with either finger on some trials, and these trials were excluded from the accuracy calculation. We analyzed this factor separately as level of responsiveness, defined as the number of trials responded to as a percentage of the trials presented. Accuracy, response times, and level of responsiveness were extracted for each participant and averaged across trials. Separate 2 x 2 ANOVA designs (group by finger, with finger as a repeated measure) were used to assess each behavioral variable.
Participants with ASD performed their movements around 3.23 ±.24 s after the video movement onset, averaged across both fingers, while control children imitated the finger-lifting movements after 3.13 ±.12 s. No significant main effect of group was observed,
Behavioral results.
| | |
---|---|---|
| ||
Reaction time ± SD | 3.23 ± 0.24 | 3.13 ± 0.11 |
Accuracy ± SD | 94.10% ± 6.95 | 98.80% ± 1.84 |
Responsiveness ± SD | 60.70% ± 14.40 | 72.3% ± 13.48 |
| ||
Reaction time ± SD | 3.24% ± 0.25 | 3.14% ± 0.14 |
Accuracy ± SD | 94.00% ± 7.15 | 95.90% ± 6.51 |
Responsiveness ± SD | 65.90% ± 18.16 | 74.80% ± 15.19 |
SD, standard deviation.
As expected, we found relevant motor-associated beta and mu oscillations in both hemispheres, contralateral and ispilateral to the movement, during imitation of both fingers (Figures
For beta-ERD, the main effect of group was significant,
For PMBR, the main effect of group was significant,
For mu-ERD, the main effect of group was significant,
Correlations between age and beta-ERD, beta-PMBR, and mu-ERD were examined in each group and each hemisphere using a Pearson r correlation coefficient. In the control group, there was a significant negative correlation between age and beta-ERD during imitation of the index finger,
Baseline mu and beta power were calculated for the source reconstructed waveforms and group differences were examined using a 2-sample
In the current study, children without ASD exhibited a well-established pattern of oscillatory neural activity before and after movement onsets in brain areas associated with motor processing. Beta and mu-ERD were observed prior to movement onset and during movement execution, whereas a strong PMBR response emerged following movement termination. Those responses were observed though contralateral and ispilateral sensorimotor cortices. Children with autism also exhibited each of these neural responses, although the mu and beta power changes associated with the imitations were significantly different from those of controls. While both affected and nonaffected children were able to perform the simple action of lifting a finger, their cortical activity levels were strikingly different. In the motor cortex, induced power revealed an increase in mu- and beta-ERD and a reduction in PMBR in the ASD group compared to the control group, during imitation of both finger movements. Our results provide some physiological evidence of distinct brain activity associated with imitation of hand movements in children with autism. Below, we discuss the implications of these findings for understanding the pathological cortical activity in children with autism.
Surprisingly, we found greater mu-ERD in the group of children with autism compared to their nonaffected peers. Whether this greater ERD is restricted to the motor-related signals or rather linked to the mirror neurons remains to be clarified. Following the “broken mirror” theory of autism [
We observed significantly greater ERD in the beta-band in children with ASD. Given the beta-ERD’s association with movement preparation [
Alternatively, difficulties with body part orientation [
Previous studies have shown that beta-ERD power during simple finger movements is correlated with age [
PMBR is proposed to be associated with motor deactivation or inhibition of the motor cortex by somatosensory afferents [
Current theories and experimental data strongly suggest that dysfunctional integrative mechanisms in ASD result from reduced neuronal synchronization [
Finally, it is critical to consider the mixed picture of behavioral results in the current study. We found that the participants with ASD responded to the imitative stimuli less often than controls but that when the participants with ASD responded, their responses were as accurate as those in the control group. Reduced responding could be interpreted as evidence of confusion over the imitative action requested but equally could be considered evidence of greater lapsing of attention during the task. In the current study, we cannot discern between these possibilities. Since we were focused on response-locked beta-band ERD and PMBR, we could not analyze trials on which the subjects did not respond to the stimuli, limiting our understanding of whether such trials were associated with additional differences in beta-band activity.
While beta-ERD and -PBMR are generated from the same regions, it is not clear whether they result from similar events at the neuronal or network level. Our cohort of children with ASD did not exhibit significant motor defects. We interpret the aberrant pattern of beta rhythms observed in our ASD group, especially the increased ERD, as most likely associated with the difficulty of cognitive processes involved in selecting the motor response rather than with a motor deficit itself. In contrast, the reduced PMBR may be related to reduced inhibition in the motor cortices. Indeed, PMBR is absent in young children [
Given the cross-sectional nature of the study and small sample size, we would like to warn about the highly preliminary nature of these findings. In addition, the low number of females might limit the generalization of the results. By essence, autism is a spectrum so the characteristics of children with ASD and their life circumstances are mostly heterogeneous in nature. Addressing these issues may require larger sample sizes and possibly interdisciplinary collaboration.
We have demonstrated that children and adolescents with autism may have reduced inhibitory drive in cortical rhythms as measured with MEG during motor imitation. Our results support previous theories that inhibitory dysfunction could be one of the factors underlying abnormal behaviors in autism. Further, changes in ERD suggest greater difficulty in movement planning in the autism group. Understanding these mechanisms may provide a potential target for future therapies to address motor-related symptoms, by both pharmacological and behavioral interventions. Whereas the relevance of altered brain oscillations to motor imitation problems in autism needs further clarification, monitoring pathological beta-bands features with MEG might hold promise as a biomarker for motor impairments in ASD. On this last point, although a large number of individuals with ASD have motor difficulties, they are not universally observed [
Autism spectrum disorders
Magnetoencephalography
Event-related desynchronization
Postmovement beta rebound
Electroencephalography
Gamma-aminobutyric acid
Diagnostic and statistical manual of mental disorders
Autism diagnostic observation schedule
Magnetic resonance (imaging).
Deidentified data is available on our laboratory server in.mat and.xlsx file format readily usable by any requester.
None of the authors have potential conflicts of interest to be disclosed.
The authors’ work was supported by Autism Speaks (Postdoctoral Fellowship #7592, I. Buard), the Colorado Clinical and Translational Science Institute (Co-Pilot Award, I. Buard) and the National Institute of Health (R01 MH082820, D. C. Rojas). The authors would like to thank Himaja Gaddipati for technical assistance with some of the MEG recordings and Ian Southwell for the realization of the animated hands (