Role of Mouse Visual Cortex in Perception of Stimulus Features

Abstract

<p>Information about diverse visual features is encoded in distributed visual cortical areas. Which area/areas are required for perception for a given visual feature and how the encoded visual information is computed to guide visual perception are the main questions of this dissertation. Here I take advantage of the mouse visual system to tackle these questions. I first identify mouse visual cortical areas (primary visual cortex (V1) and higher visual areas (HVAs)) that are required for perception of different visual features via a combination of optogenetics and mouse psychophysics. I find that V1, LM (lateromedial) and AL (anterolateral) are required for discriminating orientation and detecting contrast. However, PM (posteromedial) is not involved in the orientation discrimination task. Instead, suppressing PM increases contrast detection threshold and consistently increases false alarm rate in both contrast and speed increment detection tasks. The effects of PM suppression on false alarm rate remain intact even when the visual stimuli are presented outside of the affected visual field, suggesting a non-visual specific role. To understand the computation that transforms sensory encoding to decision choice, I next use visual adaptation as a tool to determine a decoder adopted by the mouse to discriminate orientation (target versus. distractor orientation) via a combination of in vivo calcium imaging and modeling approaches. This decoder sums the neuronal population response suboptimally with higher positive weights biased towards target preferring neurons without negatively weighting the distractor preferring neurons. This decoder scheme could also be used for detecting contrast, serving as a potential reason why similar areas are required for both orientation discrimination and contrast detection tasks. To dissect the sensory and cognitive contribution of each area to the visual tasks, I attempt to use the sensitivity and bias measures from the Signal Detection Theory (SDT). However, both behavioral and neuronal evidence suggests that changes in bias can result from changes in sensory encoding, decision criterion or both, limiting the usage of SDT in dissociating two key components of perceptual decision-making process: sensory encoding versus decision criterion placement. Lastly, since adaptation can induce changes in sensory encoding and thus has profound perceptual consequences, I also apply in vivo extracellular single-unit recording in the mouse visual areas to characterize adaptation profiles. I observe a cascaded increase of adaptation along the geniculate visual pathway and much heterogeneity of adaptation within any recorded visual areas. In all, these serials of studies provide rich neuronal, behavioral and computational evidence to link between sensory encoding and perceptual behaviors.</p>