MRI compatible dome display to study the neural correlates of dynamic immersive naturalistic vision.
Poster Presentation: Tuesday, May 20, 2025, 2:45 – 6:45 pm, Pavilion
Session: Scene Perception: Neural mechanisms
Schedule of Events | Search Abstracts | Symposia | Talk Sessions | Poster Sessions
HARISH KATTI1 (), AIDAN P. MURPHY2, DAVID LEOPOLD1,2; 1Section on Cognitive Neurophysiology and Imaging, National Institutes of Health, 2Neurophysiology Imaging Facility, National Institutes of Health
Introduction: Visual experiences in the real world are dynamic, immersive and often encountered while we move. This is contrary to how vision is traditionally studied using foveal presentation of static stimuli covering a small fraction of the field of view. To investigate the neural correlates of immersive real-world naturalistic vision, we captured full field of view video and displayed it in a novel, MRI-compatible, immersive, hemispheric “dome” display system. We then asked how brain-wide activity is engaged by different aspects of naturalistic visual stimulation. Methods: We used a fish-eye lens with 180o field-of-view, to create immersive, first-person point of view videos that capture the visual experience of ego-motion in natural real-world scenes. In addition to the original video, we created control condition videos that isolated static shape information (by temporal down sampling) and motion information (in the form of short-lived dot trajectories that sampled frame-wise optical flow). We then acquired whole-brain functional MRI responses as a rhesus macaque (macaca mulatta) free-viewed these videos projected onto the dome display. Our video capture, display and viewing procedure ensures that elements of the dynamic scene are reproduced veridically from the animal’s perspective. Results: We first established that repeated presentations of full-field dynamic video stimuli elicited consistent visually-driven, time-varying response patterns in visual cortical regions (as previously described for traditional video free-viewing fMRI) and negligible activity in control ROIs. We find that the intact video as well as dot motion control condition elicits activity in the superior temporal sulcus (STS), which suggests a broader role for STS beyond biological motion. We also find that large camera movements such as panning (yaw) left or right and tilting (roll), elicit robust responses in visually responsive cortex.