Access to parietal action representations after stroke lesions in visual cortex

点击翻译按钮获取中文摘要

PROJECT SUMMARY The ability to recognize and use objects according to their function (e.g., fork, hammer, pencil) requires integration of visual, semantic and action knowledge across occipital, temporal and parietal areas. Left parietal regions support critical aspects of object-directed action, such as grasping and object manipulation. This research activity uses a combination of fMRI and behavioral measures in patients with ischemic strokes to early and extrastriate visual areas to test the following hypotheses: Aim 1: There is a visual pathway to the parietal grasp region (aIPS) that bypasses processing in primary visual cortex. Aim 2: Left ventral extrastriate cortex is necessary to access manipulation information for visually presented objects. Aim 3A: Ballistic grasping actions to objects in the hemianopic field are influenced by volumetric properties (size, orientation) of targets. Aim 3B: Left ventral extrastriate lesions impair object function (e.g., `scissors used to cut') and disrupt access to manipulation knowledge from visual input. The research leverages strengths of fMRI (whole brain correlational measure) and neuropsychology (causal inference) to test new hypotheses about vision and action. `Tools' (i.e., small manipulable objects) are an excellent domain in which to address broader questions about the integration of sensory, motor and cognitive processing. This is because tool recognition and tool use require the integration of distinct sensory, motor and cognitive representations, and the neural substrates of tool processing are well described. The research program emphasizes fresh perspectives on longstanding ideas about the dorsal and ventral visual pathways, by a) undertaking the first systematic investigation of the types of information about objects that are extracted by visual pathways that bypass primary visual cortex, and by b) studying how some parietal areas depend on inputs from the ventral stream in order to access the correct action for a given object. The research activity innovates by testing hypotheses about how lesions at different stages in the cortical visual hierarchy affect downstream processing in parietal cortex, combining neural and behavioral measures to study brain damaged patients (generating causal evidence), and by combining univariate and multivariate measures to `read out' the information content of brain regions (parietal cortex) that are anatomically remote from a lesion. The research advances understanding of how lesions in one brain region disrupt computations in other parts of the brain that depend on the damaged region for their inputs, a phenomenon (`dynamic diaschisis') that applies to brain injury generally. Advancing understanding of these basic issues using causal data has broad implications for understanding how the brain selects the correct action for the correct object, and more generally for theories of conceptual organization and causal reasoning. Understanding how the brain accesses actions from visual input has implications for related fields, such as robotics, neuroprosthetics, and evidenced based approaches for rehabilitating function after brain injury.