Altered Cellular Organisation in the Cerebral Cortex: A New Imaging Measure of Cortical Microstructure to Meet the Challenge of Heterogeneity in Autism

Friday, May 12, 2017: 4:10 PM
Yerba Buena 9 (Marriott Marquis Hotel)
S. Chance1, M. Torso2, R. McKavanagh3, M. Ravishankar3, K. Miller2, S. Sunaert4 and M. Jenkinson2, (1)John Radcliffe Hospital, Oxford, Oxfordshire, UNITED KINGDOM, (2)University of Oxford, Oxford, United Kingdom, (3)University of Oxford, Oxford, UNITED KINGDOM, (4)University of Leuven, Leuven, BELGIUM
Background: The key neuropathological findings include altered cortical minicolumns, reductions in cortical interneurons, cerebellar Purkinje cells and various dysplasias. Several theories attempt to integrate these findings but are challenged by the underlying heterogeneity of this condition. Until now, brain microanatomy has only been assessed in post-mortem or in vitro samples whereas macrostructure and cognition has been assessed in living people. It is desirable to identify a marker of microscopic cortical organisation in the same living populations as studied by cognitive neurospychology.

Objectives: 3 experiments were conducted to confirm the utility of a new MRI measurement of diffusivity applied to the grey matter of the cerebral cortex.

Methods: The newly developed ‘CHIPS’ (Cortical High-Intensity Profile Segmentation) software was used to calculate several unique measures by comparing diffusivity data with the minicolumn axis in the cortex, derived from DTI of cortical grey matter.

i) Post-mortem MRI was conducted to validate the DTI signal as an index of minicolumn alteration by comparison with histological measurements from the identical region. 6 ASD and 6 TD control brains were studied.The 3T MRI acquisition protocol on formalin-fixed tissue included: Modified spin-echo sequence with 3D segmented-EPI, and Structural 3D balanced steady state free precession (BSSFP) sequence. Semi-automated analysis of minicolumns on histological brain sections was conducted using light microscope digital photomicrographs for comparison with semi-automated CHIPS measurements. Five different cortical regions were examined including prefrontal cortex and primary visual cortex.

ii) Discovery cohort - in vivo 7T DTI acquisitions were conducted in Oxford, UK. Data was collected from a final sample of 10 TD individuals and 10 ASD (intellectually able and/or Asperger’s). Automated whole brain CHIPS analysis was conducted and a VBM analysis was also performed on the accompanying structural scans.

iii) Test Cohort – in vivo 3T DTI data from 24 TD controls and 20 ASD participants acquired in Leuven, Belgium. Automated whole brain CHIPS analysis was conducted and a VBM analysis was also performed on the accompanying structural scans.


i) The post-mortem DTI confirmed the predictions of the model indicating a clear correlation between new minicolumn diffusion measures and histological minicolumn width across both diagnostic groups. Also a significant difference in the minicolumn diffusion angle between TD and ASD in brain regions with wider minicolumns. Both effects were evident in all cortical regions assessed.

ii) The high resolution in vivo DTI found a difference in the components of diffusion which indicated wider minicolumns in ASD yielding clear separation of diagnostic groups for 18/20 participants (90% classification accuracy).

iii) The 3T DTI results found CHIPS cortical diffusivity values yielding clear separation of diagnostic groups for 42/44 participants (95% classification accuracy).

Conclusions: Diffusivity parameters in the cerebral cortex correspond to variation in the minicolumnar organisation of neurons spanning the cortical layers. Minicolumns are more widely spaced in several cortical regions in ASD and this can be detected using novel diffusivity measures. The detection of cortical microstructural changes in ASD using MRI opens the door to possible early assessment in young infants whose symptoms are unclear.