Abstract
Functional neuroimaging studies in humans and neuropsychological case studies provide evidence for a variety of extrastriate cortical areas involved in visual motion perception. Multiple mechanisms underlying processing of different motion types have been proposed, however, support for cortical specialization has remained controversial so far. We therefore studied motion perception in 23 patients with focal lesions to various cortical areas due to ischemic stroke or hemorrhage. We considered luminance- (first order) and contrast-defined (second-order) motion in gratings, as well as translational motion, heading from radial flow, and biological motion in random dot kinematograms. Stimuli were presented with different signal-to-noise ratios. We measured detection thresholds for the contra- and ipsilesional visual hemifields and compared patients' performance with data from an age-matched healthy control sample (N=124). Elevated thresholds and significant threshold asymmetries between both visual hemifields were defined as deficits. We found specific deficits in 10 patients, whereas 7 patients showed impaired perception of multiple motion types. Lesions for all patients were mapped onto a template brain scan. Functional areas were explored by lesion density plots and subtraction analysis to compare lesions of patients with and without specific deficit. Results emphasized a dissociation between basic motion processing and processing of complex motion. Anatomical analysis confirmed critical occipito-temporo-parietal areas for perception of translational motion and moreover specific functional areas for first- and second-order motion perception in parieto-frontal and occipto-temporal regions, respectively. In contrast, heading from radial flow perception proved to be remarkably robust to most lesions. We exclusively identified the frontal eye fields as a critical structure. Biological motion perception relied on distinct pathways involving temporal, parietal, and frontal areas. Although precise functional roles of identified areas cannot be determined conclusively, results clearly indicate regional specialization for motion types of different complexity. We propose a network for motion processing involving widely distributed cortical areas. This research is supported by the German Research Foundation, Graduate program ‘NeuroAct: Neuronal representation and action control’. We thank the Neurological Clinic Braunfels for support during data collection.