Joseph Wekselblatt^1,3, Penny-Shuyi Chen^1,2, Declan Rowley^1,3, Andrew Steele⁵, Alex Huk^1,4; ¹Fuster Laboratory for Cognitive Neuroscience,, ²Neuroscience Interdisciplinary Program UCLA, ³Department of Ophthalmology UCLA, ⁴Department of Psychiatry & Biobehavioral sciences UCLA, ⁵Department of Biological sciences, California State Polytechnic University, Pomona

Introduction: Visual area MT has been studied extensively using electrophysiological techniques, but direct and comprehensive assessment of its functional architecture would benefit from imaging approaches. Here, we relied on the llissencephalic brain of marmosets, combined with novel tools to allow for easier, widespread deployment of calcium indicators (GCaMP) to characterize the functional architecture of MT at both population and cellular levels, assessed via single-photon widefield imaging, and 2-photon imaging, respectively. Methods: We packaged a calcium indicator (GCaMP8s) into a capsid engineered to (a) cross the blood brain barrier, (b) target neurons, and (c) spare the liver. We injected this construct (capb10-cag-GCaMP8s) via intravenous tail vein injection into 2 common marmosets. We then conducted MT functional mapping experiments in large fields of view (a 7 mm diameter window, and 3.5x3.5 mm^2 FOV for widefield imaging). Full-field moving dot stimuli (on vs off) were used to identify motion responsive areas, and then additional moving dot stimuli were used to measure direction tuning, speed tuning, and retinotopy. Widefield imaging was performed over the entirety of MT, and 2-photon cellular-resolution imaging was performed within different smaller ROIs within MT. Results: We identified MT in both areas due to strong and relatively homogenous expression of GCaMP throughout the imaging window. This demonstrates the viability of intravenous injections, supporting easier application and more homogenous signal over space. Widefield imaging revealed clustered direction tuning, with motion direction columns on the scale of 400-500 um per 180 degrees of direction, consistent with other estimates primarily from electrophysiological samples, The 2D spatial autocorrelation of direction preference dropped significantly over 500um, further confirming this estimate of the size of a direction hypercolumn. 2-photon imaging: identified 1336 (functionally) visual-motion stimulus driven cell across 13 sessions, and confirmed that the widefield measurements are consistent with the responses of individual neurons.

Acknowledgements: Grant/Funding support: Fuster endowment, UCLA