Brain Network Organization Correlates with Autistic Features in Preschoolers with Autism Spectrum Disorders and Their Fathers

Background:

Autism Spectrum Disorders (ASD) are neurodevelopmental conditions characterized by an abnormal brain connectivity. Given its highly heritable origin, first-degree relatives could show a set of ‘‘sub-threshold’’ features resembling milder manifestations of ASD (broader autism phenotype –BAP). There is great interest in evaluating whether BAP characteristics extend also to neuroanatomical connectivity of first-degree relatives of ASD probands. A powerful tool for the investigation of this topic is Diffusion Tensor Imaging (DTI) technique that, for the first time, was used in parents of individuals of ASD.

Objectives:

We aimed to investigate the correlation of the structural brain network: (a) with the autistic traits in fathers and (b) with the severity of ASD in their probands. We recruited only fathers since, according to the literature, they generally have a higher probability of expressing BAP traits.

Methods:

16 ASD child-father dyads were recruited. Autism severity in children (all males, age: 1.5-5.2 y) was evaluated by ADOS calibrated severity score (ADOS-CSS). Autistic traits in fathers were assessed with the Autism-Spectrum Quotient (AQ). Structural parcellation was obtained. After pre-processing of DTI data, fiber orientation distribution was estimated using constrained spherical deconvolution. Anatomically-constrained tractography was obtained and number of streamlines weighted connectomes were generated. Global measures (global efficiency, transitivity, characteristic path length and small-world propensity) and local measures (local efficiency - EL - and clustering coefficient - CC) were computed. A general linear model was performed to evaluate correlations between network and clinical measures.

Results:

No significant correlations with global measures were found. In children, ADOS-CSS positively correlated with EL of left Paracentral Lobule (p=.019) and left Superior Temporal Gyrus (p=.037) and CC of left Cingulate Cortex (p=.02). In fathers, total AQ negatively correlated with CC of left Transverse temporal cortex (p=.04). “Attention switching” negatively correlated with EL and/or CC of bilateral Superior Frontal Gyrus (left: p=.15, p=.015; right: p=.007, p=.005), left Inferior Frontal Gyrus (p=.01; p=.01), bilateral Middle Frontal Gyrus (left: p=.038, p=.013; right: p=.009, p=.004), left Frontal Pole (CC: p=.04), left Orbitofrontal Cortex (p=.006, p=.01), left Lateral Occipital Cortex (p=.025, p=.008;), left Postcentral Gyrus (p=.032, p=.049), bilateral Cingulate Cortex (left: p=0.1, p=.015; right: p=.016, p=.016), right Supramarginal Gyrus (p=.038, p=.03), right Precuneus (p=.022, p.014), right Cuneus (p=.016, p=.026), right Superior Temporal Gyrus (CC: p=.027), right Inferior Temporal Gyrus (p=.039), and right Insula (p=.024, p=.034). “Imagination” negatively correlated with CC of Nucleus Accumbens (p=.02). “Attention to details” positively correlated with EL of left Inferior Frontal Gyrus and (p=.026) and “communication” with EL and CC of right Superior Frontal Gyrus (p=.024, p=.024).

Conclusions:

Significant correlations were found between local network measures and autism severity in ASD children and autistic traits in fathers. In fathers correlations were mostly negative indicating an association of that autistic phenotype with the decrease in local efficiency. Conversely, in ASD children all correlations were positive in line with the “early over-connectivity” theory of ASD. Interestingly, some regions of impairment overlap between the two groups suggesting an intergenarational transmission of neural substrates. Overall, these results may help elucidating the neurostrctural endophenotype of ASD.