Parsing pathways of visual perception

点击翻译按钮获取中文摘要

Visually-guided behaviours evoke widespread changes in spiking activity throughout sensory and motor circuits. Determining how these complex patterns of neuronal spiking are read out (i.e., decoded) to generate perception and guide behavior is of central importance for understanding how we interact with the world, yet there is very little causal evidence for how the brain accomplishes this. Output neurons from visual cortex (V1) exhibit distinct dynamics depending on whether they project to cortical, subcortical, or hindbrain targets and recent work shows that simple visual behaviours may depend more strongly on some V1 outputs than others. This suggests that V1 may represent multiple, distinct streams of information. However, the degree to which activity arising in distinct V1 outputs differentially impacts visually-guided behaviours warrants further exploration. Here, we seek to address two fundamental questions regarding how spiking in V1 gives rise to perception and action: 1) Is activity arising in some V1 projections more closely linked to perception than others? and 2) Does the same V1 output differentially affect behaviour depending on contingencies between sensory input and motor output? My trainees and I will focus on V1 neurons that project either the superior colliculus (SC) or the Striatum (Str) as both structures receive direct V1 input and perform sensorimotor functions that are critical for perceptual decision-making. Mice will be trained to perform challenging perceptual tasks and work reliably at psychophysical threshold. We will optogenetically modulate activity in distinct subpopulations of V1 neurons that project either to the SC or Str. In the first experiment (Objective 1), mice will report direct optogenetic activation of V1 projection neurons. Here, behavioral reports provide unequivocal evidence that the experimenter-controlled patterns of activity can be decoded to guide behavior, which has immense relevance for neuroprosthetic devices. A second set of mice (Objective 2) will be trained to report changes in visual motion in two different visual tasks (detect versus discriminate visual motion). We will selectively inhibit V1 projections to the SC or Str at different times to assess the timing of their contributions to behaviour and whether those contributions are task-dependent. Across experiments, in vivo electrophysiological recordings during behaviour will helps us link sensory input and task performance with population spiking in V1, SC, and the Str (Objective 3). This program of research will yield much needed insight into how neuronal signals in arising in different V1 output channels give rise to perception and behaviour. Furthermore, the HQP conducting this work will develop a strong skillset that will prepare them for careers in academia or private industries such as the emerging neurotechnology sector.