Global Versus Local: Double Dissociation Between MT+ and V3a in Motion Processing Revealed by a TMS Study

N. Chen,P. Cai,F. Fang
DOI: https://doi.org/10.1167/14.10.287
2014-01-01
Journal of Vision
Abstract:The functional properties of motion selective areas in human visual cortex, including V3A, MT+, and IPS (intraparietal sulcus) are far from being fully understood . To examine the functional specialization of these areas for global and local motion processing, we applied offline continuous theta-burst transcranial magnetic stimulation (TMS) aided by MRI-based neuronavigation to temporarily attenuate normal functioning of unilateral V3a, MT+ and IPS in different daily sessions. Vertex was also targeted as a control site. In each session, before and after TMS, subjects were asked to discriminate the global directions of two successive random dot kinematograms (RDKs), which consisted of 400 black dots within a circular area (9째 in diameter). Four conditions were used: two motion coherence levels (100% or 40%) at two stimulus locations (left or right 9째 horizontal eccentricity). We found that V3a stimulation selectively impaired discrimination of the 100% coherence motion, while MT+ stimulation selectively impaired discrimination of the 40% coherence motion. IPS stimulation impaired discrimination of both motion stimuli. All the impairments were specific to the stimuli presented contralaterally to the TMS site. Vertex stimulation did not lead to any change in motion discrimination. The double dissociation between the TMS effects on MT+ and V3a suggest distinctive roles of these two regions in motion processing. Under the 100% coherence condition, motion discrimination could be a local motion task because every dot moves in the same direction, so that local processing of motion signals is sufficient for identifying the direction of the RDK. However, under the 40% coherence condition, global processing is required to integrate different motion directions from many dots to form a coherent motion perception. Thus, our results suggest that MT+ and V3a dominate in global and local motion processing, respectively, and the outputs of these two areas may both project to IPS for further processing. Meeting abstract presented at VSS 2014
What problem does this paper attempt to address?