Purchase this article with an account.
David Osher, Sean Tobyne, Keith Congden, Samantha Michalka, David Somers; Structural and functional connectivity of visual and auditory attentional networks: insights from the Human Connectome Project. Journal of Vision 2015;15(12):223. doi: 10.1167/15.12.223.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Recent work in our laboratory has suggested that human caudal lateral frontal cortex contains four interleaved regions in each hemisphere that exhibit strong sensory-specific biases in attention tasks (Michalka et al, 2014). Two visually-biased attention regions, superior and inferior pre-central sulcus (sPCS, iPCS), anatomically alternate with two auditory-biased attention regions, caudal inferior frontal sulcus (cIFS) and the transverse gyrus intersection the precentral sulcus (tgPCS). These small regions were identified in fMRI studies in a small number of individual subjects. Here, we have investigated these regions and their putative networks by mining the WashU-Minn Human Connectome Project (HCP) dataset. We used data from the 482 HCP participants with both diffusion-weighted imaging and resting-state fMRI. We defined seed regions from our individual subject data in a task that contrasted auditory and visual spatial attention. Probabilistic activation maps were constructed and thresholded to generate ROIs. These ROIs served as seed regions for resting state and tractography analyses of the HCP dataset. Stronger functional connectivity was observed for the sPCS and iPCS than for tgPCS and cIFS with superior parietal lobule visual attention regions, and conversely stronger connectivity was observed for the tgPCS and cIFS than for sPCS and iPCS with superior temporal lobe auditory attention regions. A similar pattern was observed with tractography for all ROIs, except for tgPCS. We next analyzed the whole-brain connectivity patterns of these ROIs using a multivariate approach; we found that the modality of sensory-bias can be predicted well above chance in both hemispheres at a voxelwise scale (L:71%, R:80%), using only the connectivity pattern of an individual voxel. A long-term goal of this analysis is to develop reliable methods for identifying fine-scale brain networks in large population datasets, which could have important clinical applications. Our preliminary results reveal both successes and challenges of these efforts.
Meeting abstract presented at VSS 2015
This PDF is available to Subscribers Only