Disrupted Thalamic-Occipital Connectivity in Six-Week-Old Infants at High and Low Risk for ASD and Its Relation to Behavioral Phenotypes

Background:

Specific patterns of impacted thalamocortical connectivity have been found in children and adolescents with ASD using both diffusion tensor imaging (DTI) and functional connectivity magnetic resonance imaging (fcMRI; Nair et al., 2013; Nair et al., 2015). In our previous work with infant siblings at high risk for ASD (Nair et al., under review), we demonstrated that at-risk infants show a pattern of functional thalamic-prefrontal underconnectivity and thalamic-occipital overconnectivity few weeks post-birth. In comparison, little is known about the anatomical connectivity between thalamus and cortical regions in at-risk infants.

Objectives:

The goal of the current study is to examine early anatomical connectivity of thalamocortical tracts using DTI in infants at high-risk for developing ASD compared to low-risk infants.

Methods:

DTI data were acquired during natural sleep on a 3T Siemens scanner for 17 high-risk infant siblings (6 weeks post-birth) of children with ASD (HR) and 15 infants at low risk (LR) for ASD. Data were preprocessed using FDT (FMRIB’s Diffusion Toolbox). Six cortical seeds were selected from motor, occipital, parietal, prefrontal, somatosensory, and temporal regions of the UNC Infant 0-1-2 neonate atlas (Shi, 2011). Probabilistic tracking was performed to derive white matter tracts originating from each cortical seed and terminated at the thalamus. The diffusion tensor was calculated at each voxel and maps of fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD), axial diffusivity (AD) were generated for tracts between each cortical seed and thalamus. These indices were then entered into a multivariate analysis of variance to compare between groups, and correlated with fcMRI connectivity z’ scores as well as behavior measures of visual processing (Infant/Toddler Sensory Profile (Dunn, 2002); “Charlie Brown” eye-tracking paradigm).

Results:

DTI results showed aberrant patterns of thalamic-occipital connectivity in the HR group consistent with our prior fcMRI findings. Specifically, the HR group demonstrated significantly higher MD (p=.02), AD (p=.01), and RD (p=.03) values for thalamic-occipital tracts. Correlations between fcMRI z’ prime scores and DTI indices revealed a normative pattern of lower fcMRI connectivity associated with low MD scores for the LR group (p=.02). However, such a pattern was not observed in the HR group. Higher thalamic-occipital MD (p=.02), AD (p=.05), and RD (p=.02) were significantly correlated with lower Sensory Profile visual processing scores in the HR group. Additionally, we found that lower thalamic-occipital FA (p<.001) and higher RD (p=.02) in the HR group was significantly correlated with greater preference for focusing on low-level perceptual features (e.g., motion, high luminance, high saturation, edges) versus character faces within clips of Charlie Brown at age 3 months.

Conclusions:

Our findings suggest that thalamic-occipital connectivity may be disrupted in at-risk infants by six weeks of age. These disruptions may also underlie early preference for sensory aspects of visual stimuli over social stimuli. Visual processing abnormalities have been found in prior studies of older children and adolescents with ASD (Keehn et al., 2013, Klin et al., 2015). Thus, our findings of thalamic-occipital connectivity anomalies and its relation to behavior phenotypes could be an early marker of ASD risk.