Venue: Centre Broca (zoom link TBD)
IMN
Equipe : Dynamique des Réseaux de l’Apprentissage Procédural
Thèse dirigée par Arthur Leblois
Title
Modeling the pathological oscillations generation mechanisms and network dynamics of the basal ganglia
Abstract
Parkinson’s disease (PD) is characterized by excessive oscillatory synchronization in the beta band (13–30 Hz) throughout the cortico-basal ganglia-thalamic (CBGT) network. Although these oscillations are a hallmark of the disease and correlate with motor symptoms, the circuit mechanisms that generate them remain unclear. Their properties, such as frequency, power and distribution across nuclei, also vary substantially across species and recording methods, raising the question of whether they reflect a single pathophysiological process or several distinct rhythms produced by different circuit mechanisms. This thesis investigates these generation mechanisms and how they differ across species.
First, a review of beta oscillations in PD patients, MPTP-treated monkeys and 6-OHDA-lesioned rodents, supported by a reanalysis of published datasets with a standardized pipeline, reveals pronounced interspecies differences: low-beta synchronization (≤15 Hz) characterizes patients and parkinsonian primates, whereas dopamine depletion in rodents mainly enhances higher frequencies. Second, using a CBGT rate model constrained by rodent- or primate-specific parameters, we evaluated the oscillations produced by each negative-feedback loop, comparing their spectral properties and deriving the expected phase relationships between populations, which are consistent with rodent data and yield testable predictions for primates. Interactions between generators, acting as coupled oscillators, further modulate the oscillation frequency in the full network. These results suggest that abnormal beta oscillations arise from distinct mechanisms in rodents and primates. Third, because connection strengths in the network are largely unconstrained by experimental data, we applied simulation-based inference to the rodent model to estimate the connection gains compatible with the responses of striatal, subthalamic and pallidal neurons to optogenetic stimulation. The inferred distributions indicate which connections are well constrained by these responses and which can compensate for one another.
Overall, this work supports the coexistence of multiple circuit-specific beta rhythms and highlights the importance of species-specific constraints when relating models of their generation to experimental data.
Jury
- Mme. BAHUGUNA Jyotika – Université de Strasbourg – Examinatrice
- M. GIRARD Benoît – Université de la Sorbonne – Examinateur
- M. KUMAR Arvind – KTH Royal Institute of Technology – Rapporteur
- Mme. VICH Catalina – Universitat de les Illes Baleares – Rapporteure