Zhang, Zihui; (2021) Closed-loop all-optical manipulation of neural circuits during behaviour. Doctoral thesis (Ph.D), UCL (University College London).

Preview

Text zzhang_final_thesis_v12_corrections.pdf Download (12MB) | Preview

Abstract

A fundamental problem in neuroscience is to understand the causal relationship between neuronal activity in the brain and behaviour of the animal. To effectively test how neural activity drives behaviour, we must not only read out activity from identified neurons, but also make precisely targeted interventions in the activity. However, this targeting has hitherto been guided by offline analysis, and typically addressed one or a group of pre-defined neurons with the same strength and timing across trials. Since the activity of individual neurons can be highly variable, and their contribution to computation can vary from trial to trial, it is essential to be able to titrate interventions according to online activity measurements. This thesis introduces a closed-loop all-optical strategy for dynamically controlling individual neurons in cortex of awake mice during behaviour. Two-photon optogenetic stimulation was rapidly tailored and delivered to specific neurons based on the online monitoring of activity in the same population using two-photon calcium imaging. Proof-of-principle experiments demonstrate that the system can be used to clamp the level of calcium signals read out from individual neurons, to boost weak sensory-evoked neuronal responses, and to pair the activity of neuronal ensembles to a specific neuron to induce long-term changes in the network. I further refined this approach by integrating a state-of-the-art calcium imaging analysis toolbox into the system to allow automatic detection and photoactivation of arbitrary ensembles of neurons, efficient mapping of sensory tuning properties, and selective photostimulation guided by the population activity in the cortex of a mouse solving a sensory-based behavioural task. Finally, I used this method to probe the circuit mechanisms that transform sensory information into meaningful behaviour. When mice perform a texture discrimination task, neurons in layer 2/3 (L2/3) of the primary somatosensory cortex (S1) exhibit stimulus and choice selectivity. Distinct populations of neurons carry information about the stimulus irrespective of the behavioural outcome (‘stimulus neurons’) or about the choice irrespective of the presented stimulus (‘decision neurons’). Using targeted photostimulation, I found the stimulus and decision neurons exhibited different activity patterns and functional connectivity in the local circuit: there is a stronger like-to-like connectivity in neuronal ensembles encoding the same stimulus identity, and a preferential activation of decision neurons. Crucially, direct activation of the correct choice-related activity in the L2/3 circuit improves behavioural performance and therefore causally links the decision signal to behaviour. The finding that neurons encoding either the stimulus or the decision are intermingled within the same circuit challenges the idea of a strict processing hierarchy in the cortex, and the causal manipulation results suggest a direct involvement of decision neurons in S1 in the decision-making process. The method, therefore, enables experiments to directly test predictions about how particular activity patterns in defined elements of a neural circuit lead to long-term changes in the circuit and their role in driving behaviour.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as XMLCSVJSON

Downloads by country - last 12 months

Export as JSONXMLCSV

Archive Staff Only

View Item