Functional transformations of sensory information in the brain: the mouse olfactory system as a model

Host laboratory and collaborators

Harold Cremer / IBDM /

Jean-Claude Platel / IBDM /

Hervé Rouault / CPT /

Scientific background

In rodents, odor stimuli generate collective activities in neuronal populations of the olfactory system. In the first sensory layer, olfactory sensory neurons project into discrete units called glomeruli, within the main olfactory bulb. Combinations of excited and inhibited glomeruli encode odors by forming spatial maps of activity. This system constitutes an ideal model to study how the brain receives and process information from the outside world. This process has been extensively studied both at the experimental and theoretical levels, but important gaps remain in the understanding of its information processing. Here, our objective is to examine the function of local synaptic signaling by a population of glutamatergic interneurons (the External Tufted Cells, or ETC) within glomeruli and to establish its role in olfactory processing. The experimental data will be obtained using a new genetic model developed in the Cremer lab coupled with in vivo two-photon calcium imaging. Analysis and computational modeling will be performed in the Rouault lab. The PhD student will ideally perform in vivo recording, analyze and implement a new computational model of the olfactory processing to take into account the new preponderant role of the glutamatergic interneuron population.

PhD Objectives

The principal objective of this project will be to understand how the glomerular network processes sensory information before its transfer in the olfactory cortex. Based on neuronal activity measurements using in vivo calcium imaging, in particular ETCs,  the PhD student will develop a firing-rate model of the olfactory bulb signal processing to generate new hypotheses/predictions on the role of these populations. This will be initially based on currently used models for the balanced state of excitation and inhibition as proposed in the neocortex.

Proposed approach 

Biology:  New mouse models (ND6-Cre and ND6 CreERT2) will enable us to label both local interneurons and projection neurons (Angelova et al submitted 2018) in the olfactory bulb glomeruli. Using in vivo two-photon calcium imaging we will reconstruct individual neuron morphology and in parallel record odor-evoked activity of the same cells.

Statistical analysis: Activity traces of the recorded neurons will be fitted by linear-nonlinear neuronal models. Inhibitory/excitatory properties of the neurons will be assessed and confronted with their type, ie projection and/or interneuron. Establishing the diversity of cell responses to odor presentation will serve the next goal of establishing a neural network model for the odor processing and encoding.

Computational: Current models of the olfactory system rely on strong local inhibition to generate sparse activity on the glomeruli output. However, such models based on winner-take-all dynamics do not explain the observed lack of correlation of the output upon presentation of combinations of odors. We propose that local excitation, as generated by the ETC, could provide a source of mixing of the input, by amplifying small differences, a mechanism at play in networks displaying a balanced state of inhibition and excitation. We will explore these ideas by comparing firing rate simulations of the dynamics with input-output relationships measured experimentally in the olfactory bulb.

PhD student’s expected profile

We expect the candidate to have a biology background with a keen interest in computational neuroscience.