Cortical Areas and Neural Connections Underlying Scene Processing

Massachusetts General Hospital, Boston MA

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): In human subjects, neuroimaging and other techniques have revealed that different regions of visual cortex respond specifically to distinct types of visual stimuli. For instance, faces selectively activate one specific set of brain regions, and scenes activate a different set, including the areas named 'PPA,' 'TOS' and 'RSC'. Understanding the function and connections of the regions that respond to scenes in the environment will provide fundamental insights about the overall strategies used by normal subjects to navigate and to perceive visual scenes. Moreover, brain damage to each of these cortical centers has been implicated in numerous neurological syndromes, including prosopagnosia, navigation agnosia and Williams Syndrome. Because the techniques available to study human subjects cannot reveal all of the basic neural mechanisms underlying the function of this network, animal models provide an essential means to understand the neural basis of the environmental perception. One goal of this proposal is to provide such a model. Our first goal (Aim #1) is to demonstrate the existence of three scene-responsive areas by using functional magnetic imaging (fMRI) in non-human primates, then to compare the cortical maps quantitatively, to show that the scene-responsive regions correspond across humans and monkeys. Based on our preliminary data, we anticipate successful completion of this aim. This will enable the use of classical, minimally invasive techniques (e.g. neural tracers) to clarify the specific circuits of scene processing (Aim #2). Three MRI-based techniques (including novel methods) will be used to trace the neural connections between each of these three areas in primates. The use of multiple tracing techniques will furnish integrated information about the cortical connections, and validate each of the techniques in a common system. Aim #2 will also answer specific questions about the neural connections underlying neural scene processing: do these three areas connect with each other, and/or with the dorsal (the 'what') stream, and/or multi-synaptically to the hippocampus, in which 'place-coding' neurons are well-known? In Aim #3, we will use fMRI to track sensory-driven information (in Aims #1/2) to higher brain levels in primate cortex, which are driven during scene recognition tasks. In humans, an identical recognition task produced robust activity in the cortical patches distinct from those produced in a face recognition task. Our hypothesis is that fMRI activity will be produced in homologous cortical areas, when monkeys are performing an equivalent recognition task. Successful completion of all aims will use different MRI-based methods to demonstrate a scene- processing network in alert primates, ranging from sensory-driven to task-driven, and the connections between these areas.