Characterising the role and properties of distinct neuronal populations involved in sensory coding using optogenetics

Abstract

The brain encodes and integrates diverse sensory inputs to produce outputs critical for survival. Uncovering the biophysical principles underlying such transformations is critical for understating how the brain works. The modular architecture of the brain allows for efficient processing of distinct modes of sensory input by specialized cell types and local circuits. However, a complete percept requires interaction of separate input streams over long distances and across hemispheres. To study brain-wide sensory processing is difficult because current experimental approaches are often limited to small regions of interest. To address these challenges, we use a combination of electrophysiology, optogenetics and computational modelling to study how neurons and circuits in mouse visual and somatosensory cortex process visual, somatosensory and auditory information. First, we focus on binocular visual cortex, a region in the primary visual cortex (V1) that plays a crucial role in processing visual information. We develop novel methods to identify binocular and monocular neurons in vitro and in vivo, and study their cellular and circuit properties. We have identified two distinct populations of layer 2/3 pyramidal neurons in binocular visual cortex. The binocular population was found to receive long-range, monosynaptic excitatory input from the contralateral visual cortex. In contrast, the monocular population does not receive input the contralateral visual cortex. Furthermore, we show that monocular, but not binocular, neurons send callosal projections to the opposite visual cortex, indicating that callosal projections primarily carry monocular visual information. Despite similar passive and morphological properties, binocular neurons are intrinsically less excitable than monocular neurons due to a higher expression of D-type potassium channels. These data suggest that binocular neurons may have different cellular integration rules from monocular neurons during processing of visual information. Using two color optogenetics, we show that binocular layer 2/3 neurons receive direct input from the lateral geniculate nucleus and the opposite V1. Furthermore, we provide evidence that distinct populations of both excitatory and inhibitory neurons are involved in processing binocular visual inputs. Second, we focus on whisker encoding in vibrissal primary somatosensory cortex (vS1). We show that superior colliculus, which plays a key role in attention, modulates sensory processing in vS1 via a di-synaptic pathway through thalamus. Optogenetic activation of superior colliculus modulated the input-output relationship of neurons in vS1 during whisker movement, enhancing responses to low amplitude whisker deflections. We found that the impact of superior colliculus on vS1 was mediated by a powerful monosynaptic pathway from superior colliculus to the posterior medial (POm) nucleus of the thalamus. Furthermore, we show that POm neurons receiving input from SC provide monosynaptic excitatory input to layer 2/3 and layer 5 neurons in vS1. Finally, we focus on multimodal sensory integration of auditory and somatosensory information. We show that auditory cortex sends direct monosynaptic projections to specific populations of neurons in the forelimb region of somatosensory cortex, enhancing somatic action potential output. This effect was observed in a population of layer 2/3, but not layer 5, pyramidal neurons, and was associated with improved performance during a sensory goal-directed task. To summarize, in this thesis we identify and describe novel pathways and specialized circuits that process and enhance the integration of sensory information across different modalities. In doing so we provide valuable insights into the role and properties of distinct neuronal populations involved in sensory coding as well as the biophysics of computation at the circuit and single cell level implemented by the cortex.