Authors:

CUSACK, R., OWEN, A.M.

Reference:

Year of publication:

2008

CBU number:

6871

Abstract:

Scanning volunteers without specifying a task for them to do (“resting state”) has previously been used to find functionally connected networks. In the current study, rather than choosing this approach, connectivity was measured across a broad range of different tasks in 1200+ imaging runs acquired from 300+ volunteers for FMRI studies performed at our institution. Even across this task variety, clear consistent networks were identified, with high correlation values for some distant regions. Seeds in the VLPFC and DLPFC both revealed strong connectivity with the multiple demands (MD) fronto-parietal network. No dissociation was found, consistent with a lack of modular distinction between these regions as concluded following meta-analysis. In contrast, although both also showed co-activation with the MD network, the pIPS and aIPS dissociated, with the former showing greater connectivity to the hippocampus, parahippocampal regions and medial frontal cortex. This dissociation is consistent with findings from other methods, including human DTI, macaque axon tracing, a recent resting state study, and FMRI of perceptual organisation and selective attention. This method avoids a problem inherent in resting state studies which is that a significant proportion of the observed results may arise from the fact that volunteers are performing a similar task, albeit not one specified by the experimenter. It has the added advantage that it does not require the collection of new FMRI data, and as a result can have great power at low cost. We discuss the challenges faced in the implementation of the method, and future extension to multivariate connectivity.