Relating Alterations in White Matter Circuitry to Cognitive Performance

Background: Diffusion tensor imaging (DTI) has consistently demonstrated widespread reductions in fractional anisotropy (FA) in adults with autism (e.g. Keller et  al., 2007), however DTI-based quantitative tractography studies have been less consistent (e.g., Catani et al., 2008; Contouro et al., 2008; Pugilese et al., 2009; Thomas et al., 2011). DTI-based tractography suffers from a well-known limitation in its ability to track connectivity where tracts cross. Within the last year, we have made transformative technological advances in High Definition Fiber Tracking (HDFT) that permit noninvasive tracing of individual fiber tracts at previously inaccessible levels of precision and accuracy (Verstynen et al., 2011; Yeh et al., 2010; 2011), allowing us to map the connectivity of individual micro-fiber bundles arising from the cortex, follow them as they turn and cross other tracts, and identify their cortical/subcortical destinations. The present study is one of the first comprehensive investigations of brain structure and function in autism between the ages of 40 and 65. The sample includes a significant number of older adults with autism, addressing an egregious lack of information about the aging brain and mind in this disorder.

Objectives: To relate anatomic connectivity alterations in specific tracts to cognitive performance drawing on brain areas connected by those tracts.

Methods: The methods include combining high-angular diffusion weighted imaging (HARDI) acquisition methods (128-directions, high b-values, 32-channel head coil) with advanced computational methods for reconstructing the diffusion orientation density function (generalized q-space imaging and fiber tractography) to map the detailed white matter connectivity in individual participants.

 Results: We have successfully collected preliminary HDFT data from a sample of 8 individuals with autism and 9 controls aged 40-65 years, and found important group differences, even in this small sample. The fractional anisotropy of the left superior longitudinal fasciculus is reduced in autism (F(1 ,14) = 6.02, p < .05) and it reliably predicts a lower-level language ability (word reading speed) (r = .69, p < .05) (measured with the WRAT). In addition, the number of fibers in the right uncinate fasciculus was reduced (F(1 ,14) = 9.58, p < .01). Importantly, the number of fibers in this tract reliably predicts higher level language abilities in autism, namely inference-making ( r = .71, p < .05) and oral expression (r = .77, p < .05) (measured by the Test of Language Competence).  In both cases, the tract properties show a strong correlation with the type of language performance that relies on the specific interregional connectivity provided by the tract. This extremely encouraging preliminary finding exemplifies the ability to relate cognitive performance to connectivity on an individual basis. Moreover, these tract measures do not predict performance in controls, where connectivity is apparently not a limiting factor.

Conclusions: These findings constitute some of the first evidence that white matter properties are disrupted in older adults with autism and that these measures are related to cognitive performance. These preliminary results demonstrate the feasibility of the approach, its applicability to older participants, and its potential fruitfulness.