Doctoral thesis
English

The role of afferent input in neuroprosthetic learning

DirectorsHuber, Danielorcid
Imprimatur date2022-11-16
Defense date2022-11-16
Abstract

Brain-Machine-Interfaces can potentially provide powerful means to replace impaired motor functions. To improve currently available devices, it is important to gain a better understanding of the neuronal mechanisms underlying neuroprosthetic control. We have demonstrated that learning-related changes in neuronal firing can be highly specific to the conditioned neuron. These changes are likely the result of neuronal plasticity, such as long-term potentiation or more synchronous input to the conditioned neuron. Independent of the type of plasticity involved, one key question remains: what is the origin and nature of the synaptic input driving the conditioned neuron?

To identify the afferent activity driving neuroprosthetic learning, we designed a novel two-photon microscope with multiple focal planes, capable of imaging two cortical layers simultaneously while we condition a chosen neuron. By labeling the conditioned neurons (in L2/3) and afferent incoming axons (in L1) with spectrally different calcium indicators, this optical approach allows us to image the in- and output of specific synaptic interactions simultaneously.

Our first goal was to characterize the role of the motor thalamus and premotor areas in the learning process. Both regions show high correlations with the conditioned neuron, and with their signals we can easily reconstruct the activity patterns of conditioned neurons and its neighbors. However, calcium dynamics of buttons from the two regions also show differences. Thalamic activity is slowly evolving over a learning session suggesting a stable input contribution, which could be necessary for long-term memory consolidation. In contrast,we observed that input from premotor areas showed more transient changes, which might be related to early phases of learning. Taken together, these experiments suggest a differential role of the two afferent input sources in the process of neuroprosthetic learning.

Our second goal was to track the structural plasticity of the conditioned neuron after learning. The experiment revealed that plasticity could be seen at the basal and apical levels of the neuron dendrites. Moreover, we could observe a significant quantity of shrinking spines at the apical level, confirming that single cell conditioning remodels the entire circuit rather than only adding new connections to an existing circuit. Finally, we could also observe a significant number of growing spines in the dendrites around the soma area compared to control conditions. This finding highlights the influence of more local circuits in single-cell conditioning.

Keywords
  • Brain-machine-interfaces
  • Impaired motor functions
  • Neuronal mechanisms
  • Neuroprosthetic control
  • Learning-related changes
  • Neuronal firing
  • Plasticity
  • Long-term potentiation
  • Synchronous input
  • Synaptic input
  • Afferent activity
  • Two-photon microscope
  • Cortical layers
  • Conditioned neurons
  • Afferent incoming axons
  • Motor thalamus
  • Premotor areas
  • Activity patterns
  • Calcium dynamics
  • Stable input contribution
  • Long-term memory consolidation
  • Transient changes
  • Structural plasticity
  • Basal levels
  • Apical levels
  • Dendrites
  • Shrinking spines
  • Growing spines
  • Local circuits
  • Single-cell conditioning.
Citation (ISO format)
PHILIPPIDES, Antoine. The role of afferent input in neuroprosthetic learning. Doctoral Thesis, 2022. doi: 10.13097/archive-ouverte/unige:167207
Main files (1)
Thesis
accessLevelRestricted
Secondary files (1)
Identifiers
218views
0downloads

Technical informations

Creation01/13/2023 4:37:00 PM
First validation01/13/2023 4:37:00 PM
Update time03/16/2023 10:52:12 AM
Status update03/16/2023 10:52:10 AM
Last indexation11/01/2024 4:23:06 AM
All rights reserved by Archive ouverte UNIGE and the University of GenevaunigeBlack