Myelin Imaging in Autism - Preliminary Reports

Background: Autism is a pervasive developmental disorder characterized by social, communicative and behavior impairments.  A hypothesized substrate of the disorder is aberrant white matter maturation and abnormal myelination, evidence of which has been suggested by volumetric, spectroscopic and diffusion tensor magnetic resonance imaging studies. However, while these methods provide qualitative information related to white matter microstructure and fiber coherence, they do not quantitatively evaluate myelin content, which is believed to underlie the observed white matter changes. Multi-component Driven Equilibrium Single Pulsed Observation of T₁ and T₂ (mcDESPOT) is a recently proposed magnetic resonance imaging (MRI) method which allows whole-brain myelin analysis. With mcDESPOT, the MR signal is decomposed into contributions from water in the intra- and extracellular compartments, and water trapped between the lipid layers of the myelin sheath, allowing quantification of each compartment's volume fraction. These measures of myelin water content correlate strongly with histological myelin assessments.  Here we report on a pilot study comparing myelin content throughout the brain in people with Autism Spectrum Disorders (ASD) and healthy controls using mcDESPOT.

Objectives: To quantitatively compare myelin content in people with ASD and healthy controls using the mcDESPOT multi-component relaxometry MR imaging method.

Methods: To date, in-vivo data from 14 healthy male controls (aged 23-36) and 2 male autism spectrum disorder (ASD) patients (aged 25 and 26) right-handed, of normal intelligence (IQ>70) and not using medication, have been collected. Following acquisition, mcDESPOT analysis, and normalization to standard MNI space, voxel-wise maps of mean myelin content (and standard deviation) were calculated from the healthy volunteer data.  To investigate differences in the ASD patients, Z-scores were calculated voxel-wise and thresholded to identify areas of substantial difference (Z +/- 4).  Mean values were then obtained from regions of interest (ROIs) placed within identified areas.

Results: Preliminary results suggest that ASD patients have reduced myelin content bilaterally within the frontal lobe and internal capsule. This myelin reduction in the frontal lobes is consistent with prior reports of white matter volume reductions in these areas, suggesting variations in myelin content may underlie these established observations. From the ROI analysis, healthy population myelin volume fraction values obtained for the right and left frontal lobes were: 24.1 (0.002)% and 23.6 (0.0017)%, respectively, and for right and left internal capsule: 21.3 (0.009)% and 22.6 (0.0019)%, respectively.  For ASD patient #1, corresponding values were: 23.4 (0.012)%, 23.9 (0.01)%, 22.2 (0.014)% and 21.1 (0.009); and for ASD patient #2: 20.2 (0.011)%, 20.9 (0.009)%, 19.2 (0.015)% and 18.9 (0.007)%.

Conclusions: Though preliminary, these results attest to the investigative potential of the mcDESPOT approach in ASD.  While aberrant myelination has been proposed to underlie ASD symptoms, to date this hypothesis has been difficult to directly test due to the absence of a suitable myelin imaging technique. Our method, reported here for the first time, provides a new quantitative technique for measuring brain myelination in people with autism.