Cortical control of skilled forelimb movement

Motor control is a basic but fundamentally important aspect of human and animal behaviour. Descending motor output from cortical and brainstem motor areas shape the activity of spinal cord circuits to execute a wide range of motor behaviours from simple locomotion to complex, dexterous tasks such as reaching or grasping, and disrupting these pathways leads to sever motor deficits and loss of motor control. Over the past century significant advances have been made in understanding how activity patterns in motor cortex relate to high-level motor control. However, the cellular and circuit mechanisms by which principal neurons in primary motor cortex (M1) transform afferent synaptic input into behaviourally-relevant patterns of motor commands remains unresolved. Therefore, this research proposal will focus on two main questions:

How are movement representations organised in the output layer (layer 5B) of M1?
Can specific aspects of motor behaviour be decoded from the neural activity observed in M1?

To address these basic but fundamentally important questions we will use a multidisciplinary approach combining large-scale 2-photon population calcium imaging of neural activity, viral based manipulation strategies, 3D kinematic analysis of movement, quantitative behaviour and Bayesian decoders of population data. This project will be conducted in collaboration with Dr. Mattias Hennig (School of Informatics, University of Edinburgh) and Dr. Aldo Faisal (Dept. of Bioengineering, Imperial College London) and will address 3 main aims:

Aim 1 - Optimise a touchscreen-based visually-guided reach-to-touch forelimb behavioural task. To investigate the neural underpinnings of visually-guided reaching in mice we will implement and optimise a touchscreen-based behavioural paradigm that requires mice to visually locate a target on screen and reach-to-touch upon presentation of an auditory 'go' stimulus.

Aim 2 - Determine how population-level representations of movement are organised in the main output layer of M1. To characterise the neural representations of movement in M1 we will use wide-field 2-photon population calcium imaging of neuronal activity in the output layer of M1 while mice perform repeated trials of a visually-guided reach-to-touch forelimb behavioural task. Viral-based opto-/chemogenetic manipulation strategies will be used to investigate causality between the activity of subpopulations of projection neurons in M1 and behaviour.

Aim 3 - Characterise the mapping between M1 neuronal activity and forelimb kinematics with Bayesian decoders. Using population imaging data generated in Aims 1 & 2, we will build Brain Machine Interface (BMI) decoders to predict limb trajectories from neuronal population activity. By applying advanced Bayesian decoders of population data and computer vision algorithms we will develop an in-depth understanding of how L5B population and single-cell dynamics combine to generate accurate, skilled forelimb movements.

This project will provide the student with practical skills in experimental design, 2-photon imaging, statistics and advanced computational methods for data analysis, with the overarching aim of characterising the cellular and circuit computations performed by motor cortex during behaviour and importance of descending cortical information for performing precise motor actions.