Computational protocol: Reduced reward‐related neural response to mimicry in individuals with autism

Similar protocols

Protocol publication

[…] Participants were scanned in a 3T Siemens TIM Trio MRI scanner with 32 channel head coil; 32 3‐mm‐thick axial slices were acquired in descending sequential order using a multi‐echo sequence, with three different echo times [TR = 2400 ms; TE (1; 2; 3) = 20; 36; 52 ms]. Multi‐echo sequences have been shown to have considerably greater signal to noise ratio for echo‐planar images (Lombardo et al., ). DICOM files were converted to NIfTI data image files using dcm2nii in MRICron. Pre‐processing and multi‐echo ICA (Kundu et al., , ) were performed in AFNI (Cox, ). The first four volumes were discarded to allow for the stabilization of the magnetization. Procedures consisted of slice‐timing correction, realignment of the functional images for motion correction, and the functional to structural co‐registration. The multi‐echo‐ICA was then performed, and the BOLD (linear TE‐dependent signal decay) and non‐BOLD components were separated. The non‐BOLD components were used as nuisance regressors to de‐noise the functional data. The de‐noised functional images were converted to SPM 3D images with dcm2nii and spatially smoothed with a Gaussian kernel of FWHM 5 mm using SPM8 (http://www.fil.ion.ucl.ac.uk/spm). [...] Statistical parametric maps were calculated with multiple regressions of the data onto a model of the hemodynamic response (Friston et al., ). The first‐level general linear model analyses contained three regressors for mimicking, anti‐mimicking and oddball conditions, and each stimulus lasted 3‐s. Regressors were convolved with the canonical hemodynamic response function. For each ROI, the mean t‐statistics of the contrast (mimicking > anti‐mimicking faces) for each participant were extracted with MarsBaR (version 0.44) and used as dependent variables for the group‐level analysis. To test both categorical and dimensional approaches, two models of ordinary least squares regression including the handedness and conditioning accuracy as covariates were created. The first model tested the effect of group (neurotypical vs. ASD), while the second model tested the effect of AQ or EQ. Mean ± 3SD was used as the criteria to filter outliers, and none were identified. Two similar analyses were conducted at the whole brain level, (i) a random effect flexible factorial analysis with two factors: Group (ASD vs. Neurotypical) × Condition (mimicry vs. anti‐mimicry), and (ii) a random effect multiple regression with either AQ or EQ as the regressor. As in the ROI analysis, handedness and conditioning accuracy were entered in both models as covariates. Main effects of Group, Condition, and the interaction effect in the factorial analysis, as well as the effect of AQ in the multiple regression analysis, were checked. We imposed an initial voxel‐level threshold of uncorrected P < 0.001, and then a cluster‐level threshold of family‐wise error (FWE)‐corrected P < 0.05 for the entire image volume. The anatomical labels reported in the results were taken from the Talairach Daemon database (Lancaster et al., , ) or the AAL atlas (Tzourio‐Mazoyer et al., ) incorporated in the WFU PickAtlas Tool (Maldjian et al., ). The Brodmann's areas (BA) were further checked with the Talairach Client using nearest grey matter search after coordinate transformation with the WFU PickAtlas Tool. […]

Pipeline specifications

Software tools MRIcron, AFNI, SPM, AAL
Applications Magnetic resonance imaging, Functional magnetic resonance imaging
Organisms Homo sapiens