Côte d'Azur University | Open PhD position in Consciousness Modeling
How does the brain’s wiring and function shape our state of consciousness—from deep sleep to psychedelics? We're recruiting a PhD student to explore structure-function dynamics using whole-brain modeling, fMRI, and HPC at INRIA. Join us to study the Modulation of Dynamical Structure–Function Correlation Across States of Consciousness.
Profile and skills required
The candidate must hold a Master or equivalent degree when starting the PhD.
Project description
This doctoral project investigates the dynamic interplay between structural and functional connectivity across different states of consciousness—such as deep sleep, general anesthesia, coma, and psychedelic experiences. The central hypothesis is that the correlation between structural connectivity (SC), reflecting anatomical brain wiring, and functional connectivity (FC), describing statistical dependencies in neural activity, is modulated by changes in consciousness. Combining empirical neuroimaging with whole-brain modeling, the project aims to uncover mechanistic insights into the neural basis of consciousness.
Building on recent advances in connectomics and whole-brain modeling, this work addresses a key gap: the lack of a unified framework to explain how structural constraints give rise to diverse functional patterns across consciousness states. This is timely, as the field shifts toward integrative modeling and cross-modal data fusion—offering a unique chance to contribute to theoretical neuroscience.
The thesis is structured in two phases:
Phase I – Empirical Characterization: Using multimodal neuroimaging (MRI, fMRI), we will quantify SC–FC coupling across varying consciousness states. Analytical tools include stochastic modeling of whole-brain dynamics [1], connectome harmonic decomposition [2], dynamic functional connectivity [3,4], and higher-order statistical interdependencies [5]. This phase aims to assess the generalizability and sensitivity of these tools to state-dependent changes.
Phase II – Computational Modeling: We will develop whole-brain simulations to reproduce and interpret SC–FC dependencies across consciousness states [6,7].
Two biophysical models will be explored:
The Adaptive Exponential Integrate-and-Fire (ADEX) model [8], which describes single-neuron dynamics with adaptive currents.
The Dynamic Mean Field (DMF) model [9], capturing population-level dynamics.
A key innovation is the use of GPU-based high-performance computing to simulate networks with up to ~10^6 ADEX neurons per region—bypassing mean-field approximations to generate firing rates and infer fMRI BOLD signals [7]. The mean field-free approach [10] allows us to track cell-scale parameters and evaluate their whole-brain effects without mean-field assumptions [11]. Another innovation involves Bayesian optimization to fit ADEX model parameters to empirical data. For DMF, we will explore how parameter changes reproduce distinct consciousness states within a single model framework.
References:
[1] Gilson et al., Phys. Rev. E, 107(2):024121, 2023.
[2] Luppi et al., Nat. Commun., 15(1):2171, 2024.
[3] Demertzi et al., Sci. Adv., 5(2):eaat7603, 2019.
[4] Castro et al., Commun. Biol., 7(1):1224, 2024.
[5] Gatica et al., Brain Connect., 11(9):734–744, 2021.
[6] Cofre et al., Brain Sci., 10(9):626, 2020.
[7] Destexhe et al., Nat. Comput. Sci. (Accepted), 2025.
[8] Brette & Gerstner, J. Neurophysiol., 94(5):3637–3642, 2005.
[9] Herzog et al., Netw. Neurosci., 8(4):1590–1612, 2024.
[10] Theodoropoulos et al., PNAS, 97(18):9840–9843, 2000.
[11] Cormier et al., Stoch. Process. Appl., 130(5):2553–2595, 2020.
Source and more details: https://adum.fr/as/ed/voirproposition.pl?langue=&matricule_prop=64039&site=sticuca
