Date of Award
Doctor of Philosophy (PhD)
Navigating through unpredictable environments requires an efficiently organized sensory system capable of adapting to behavioral demands. In the mammalian visual system, two adaptations which have arisen to meet these demands are parallel processing and top-down feedback of internally generated expectations and contextual information for adjusting responses to match the needs of the current behavioral task. Mouse visual cortex exhibits the latter of these organizing principles, like other mammals, in the form of a molecular layer (Layer 1) which receives feedback to contextually adapt sensory responses from the outside world. The first of these organizing principles, parallel processing streams, generally takes the form in the visual cortex of interdigitating modules with distinct connectivity and physiology. Modular organization clusters similarly tuned cells, improving the signal-to-noise ratio, while minimizing total axonal wire length. Early studies on mouse visual cortex were inconclusive, however, in finding anatomical or physiological clustering representing parallel processing modules. Subsequent research revealed that mouse visual cortex is in fact modularly organized: Layer 1 contains regularly spaced patches of high and low muscarinic acetylcholine receptor 2 expression (M2+ patches and M2- interpatches). Cells aligned with M2+ patches are tuned for shape while those alignedwith M2- interpatches are tuned for features of stimulus motion. In this thesis I further explored M2 modularity by investigating whether M2+ patches and M2- interpatches show specializations for visual guidance of particular behavioral tasks. In the first study, I address this question anatomically by demonstrating that mouse higher visual cortex is modularly organized, comprising interspersed, parallel input and output streams to brain systems responsible for visual processing, landmark identification, and fear regulation. I use pathway tracing and retinotopic mapping to show that the areas exhibiting this modularity reside within the visual ventral stream and establish that the strongest outputs to the amygdala emerge from an anatomically distinct region, the postrhinal area. In the second study, I investigate the relationship between M2 modules and locomotion, which increases the response gain of visual cortex neurons. By using calcium recordings in awake mice during locomotion and visual stimulation, I find that M2- interpatches are more responsive to locomotion and that locomotion increases long-range interpatch-interpatch correlations, allowing for integration across the visual field. I show anatomical organizations which may underlie this physiology, in increased inputs from somatostatin interneurons to M2- interpatches. Finally, I show that M2- interpatches have specific connectivity with the primary and secondary motor cortex. Together, these studies demonstrate that visual cortex is organized into modular subnetworks with distinct specializations for top-down visual guidance of behavior.
Chair and Committee
Timothy Holy, Camillo Padoa-Schioppa, Larry Snyder, David van Essen,
Meier, Andrew M., "Modularity of Feedback for State-Dependent Guidance of Navigation in Mouse Visual Cortex" (2021). Arts & Sciences Electronic Theses and Dissertations. 2373.