NeuroPSI – Université Paris-Saclay

Optimizing cortical feedback for a Brain-machine interface

Summary of my work
Pioneering studies aim to improve the everyday life of motor-impaired patients by providing motor rehabilitation devices controlled directly by brain activity. In order to use these neuroprostheses efficiently, patients need online sensory feedback to guide and correct ongoing movements. It is known that precise somatosensory information from the body parts, and not only visual infor- mation, is vital for dexterous control. Thus, brain-machine interfaces (BMIs) should both read neural activity from the brain and feed back sensory infor- mation about the prostheses current state. Recent efforts in closed-loop BMI systems are addressing this challenge promisingly. However, an understanding of the neuronal mechanisms of sensorimotor integration will be necessary to optimize sensory feedback delivery. In this thesis, we are using a low-latency, closed loop brain-machine interface for head-fixed mice, which combines elec- trophysiological recordings in M1 and optogenetic stimulation in the primary somatosensory cortex (S1). We aim to reveal general rules about how the brain uses spatio-temporal patterns of cortical activity in order to generate feedback- corrected motor commands, and further understand the mechanisms behind the computational rules for sensory-guided behavior. Firstly we showed that taking into account the topographical organization of the whisker barrel cortex, which highly reflects the organization of the whisker pad, favors the learning of a sensory guided task. Secondly, we implemented an ultra-fast incremental control algorithm to study the impact of latency in BMI learning. We expect that mimicking the physiological intrinsic latency of the sensorimotor system should promote learning.

Composition du Jury

• • Sylvain Crochet, Rapporteur (EPFL, Lausanne)
  • Thomas Brochier, Rapporteur (Institut de Neurosciences de la Timone, Marseille)
  • Gisella Vetere, Examinatrice (ESPCI, Paris)
  • Brice Bathellier, Examinateur (Institut de l’Audition, Paris)
  • Sami El-Boustani, Examinateur (University of Geneva, Genève)

et

• Valérie Ego-Stengel, Luc Estebanez & Daniel Shulz (Institut NeuroPSI, Saclay)