Discovering the neural correlates of the muscular activity involved in a Reach-and-Grasp task in Healthy and Deficient Mammals

This is a project to understand how the brain controls muscular contractions during the reach and grasp task. Like many other movements, muscular activity during this activity is highly correlated. This parallel processing allows the movement to be more efficient. How do these correlations arise? To answer this question, we will analyze two types of neuronal recordings to see if the patterns of neural synchronization over the course of this task are able to explain what is seen at the muscular level. This will greatly improve our understanding of the neuronal control of this movement for grasping objects which is essential to our activities of daily living. We will also use this understanding of neuronal cluster formation to make predictions concerning the reach and grasp task (this computation is essential to Brain Computer Intarfaces). This part of the project will be done with Machine Learning. We will use two types of neuronal recordings to investigate the reach and grasp task 1. Multunit recordings from over 200 electrodes in the monkey brain (recordings from Dr Numa Dancause of the Université de Montréal) 2. Human EEG recordings (to be performed at INSERMU1093, UB). The ability to make predictions concerning the reach and grasp task with the human EEG data would be useful for the control of prosthetic limbs. We propose an improvement over previous efforts to do this, by testing the most recent advance in Machine Learning predictions for time series – the transformer networks which have produced the impressive capabilities of ChatGPT. This method is to be compared with older methods such as convolutional neural networks. Finally we will analyze alterations in neural clustering using a reversible mammalian models of neuronal command inactivation from the laboratory of Dr Dancause. The multicellular recordings from monkeys in this Canadian laboratory provide us with the rare opportunity for a fine grained investigation of brain plasticity and neuronal clustering following brain injury.

The CAPS laboratory brings together experts in Neuroscience, Biomechanics, Physiology and Sports training as well as researchers and clinicians from six university hospital departments (Geriatrics, Motor Rehabilitation, Rheumatology, Orthopaedics, Neuroradiology and Psychiatry). The lab conducts animal and human research. The latter includes healthy volunteers, high-level athletes, aging subjects and patients. Our activities in applied research are organized around 3 platforms: The PIT (The Technology Investigation Platform of the Dijon CHU), the CEP (Centre for Performance Expertise) and the Marey Centre.

The PhD contract for these studies will begin in October 2024. The ideal candidate for this project will have taken classes in Machine Learning as well as Neuroscience. The candidate must know how to program. Some notions concerning the neuronal control of movement would also be appreciated. The study is to be carried out at the French INSERMU1093 laboratory in Dijon, France.