Neuromodulatory mechanisms of predictive processing in the mouse visual cortex
The predictive processing framework is considered a universal principle in the operation of the brain(1). However, how it is implemented on the level of circuits and single neurons is an open question(2). For any such computation, external sensory information must be compared with internally generated predictions. We have previously uncovered the connectivity and synaptic integration foundations for such algorithms in the mouse visual cortex (3). In this project, we will study how neuromodulators, such as serotonin (implicated in depression and psychosis) and acetylcholine (implicated in attention and dementia), govern the integration of internal and external information.
Neuromodulation, visual cortex, predictive processing, spiking neuronal network models, computational neuroscience, optogenetics, patch-clamp
The first objective is to explore pre- and postsynaptic neuromodulatory mechanisms using spiking neuronal network models (4) developed by the Perrinet team.
The second objective is testing model predictions experimentally. This will be done in the Rancz team using optogenetics, pharmacology and whole-cell recordings in brain slices.
Proposed approach (experimental / theoretical / computational)
We will build spiking neuronal network models approximating the mouse visual cortex using connectivity data from our lab (3) and the literature. On top of recurrent local connections, we will incorporate often neglected long-range excitatory input and the associated feed-forward inhibition currently being studied by the Rancz team. Notably, various models of neuromodulation, both on the pre- and postsynaptic levels, will be included. In particular, the normative models will allow designing stimulations using a recently developed optimisation method (5). The relative contribution of neuromodulatory mechanisms in model instantiations will then be tested experimentally. We will record somatic and dendritic neuronal activity during the optogenetic stimulation of different input streams. Model predictions of the neuronal coding of prediction errors will be tested by pharmacologically blocking or activating specific neuromodulator receptors.
We will combine computational and experimental neuroscience in the proposed project. Theoretical and computational neuroscience (TCN) is essential to propose computations and their algorithmic instantiations. Experimental neuroscience (EN) can uncover the biological mechanisms underlying these model-predicted computations. TCN can tackle experimentally intractable problems by narrowing the parameter space. EN, in return, can feed back architectural insights to help build better models, e.g. for artificial intelligence. Ultimately, understanding the biological underpinnings of the mind will lead us to alleviate the suffering caused by psychiatric diseases and the human condition in general.
We seek a highly motivated and versatile candidate with a background in experimental biology who is willing to learn computer modelling. The candidate should be interested in neurophysiology and have proven programming skills (e.g. Python). In addition, knowledge of synaptic physiology, patch-clamp recordings or numerical modelling would be a plus.
Is this project the continuation of an existing project or an entirely new one? In the case of an existing project, please explain the links between the two projects
2 to 5 references related to the project
(1) Bar, M. (2009) 10.1098/rstb.2008.0321
(2) Spratling, M. W. (2017) 10.1016/j.bandc.2015.11.003
(3) Galloni, A. R., Ye, Z. & Rancz, E. (2022) doi:10.1523/JNEUROSCI.1620-21.2022
(4) Davison, A. et al. PyNN v0.9.6. (2020) http://neuralensemble.org/PyNN/
(5) Edgar Y. Walker et al. (2019) 10.1038/s41593-019-0517-x
3 main publications from each PI over the last 5 years
- AR Galloni, Z Ye, E Rancz; 2022. Dendritic domain-specific sampling of long-range axons shapes feedforward and feedback connectivity of L5 neurons. Journal of Neuroscience 42 (16), 3394-3405
- A Tran-Van-Minh, Z Ye, E Rancz; 2022. Quantitative analysis of rabies virus-based synaptic connectivity tracing. bioRxiv, accepted in PlosOne
- AR Galloni, A Laffere, E Rancz; 2020. Apical length governs computational diversity of layer 5 pyramidal neurons. Elife 9, e55761
- Victor Boutin, Angelo Franciosini, Frédéric Chavane, Laurent U Perrinet (2022).Pooling in a predictive model of V1 explains functional and structural diversity across species. PLoS Computational Biology.
- Frederic Chavane, Laurent U Perrinet, James Rankin (2022). Revisiting Horizontal Connectivity Rules in V1: From like-to-like towards like-to-All. Brain Structure and Function.
- Hugo Ladret, Nelson Cortes, Lamyae Ikan, Frederic Chavane, Christian Casanova, Laurent U Perrinet (2022). Dynamical processing of orientation precision in the primary visual cortex. bioRxiv (in revision).