Welcome to my personal webpage !

I am a Computational Neuroscientist at the Paris Brain Institute (ICM) where I study the principles of information processing in cortical networks by using different theoretical and computational tools (see below).

In the laboratory (teams of Nelson Rebola and Alberto Bacci), we combine this computational approach with neurophysiological recordings (optophysiology, electrophysiology) and experimental manipulations of brain activity (optogenetics, chemogenetics, genetic deletion).

Theoretical Neuroscience

I work on models of cortical activity at the cellular and network scale. I use analytical approaches together with numerical simulations to study the emergent properties of those systems. Various examples of such theoretical modelling work can be found in my publication list.

Data Science

Our research strongly relies on the use of machine learning to analyze and interpret neurophysiological data.

I design and implement multivariate models of neural activity from behavioral and sensory features (notably Generalized Linear Models). We use those models to (1) extract functional principles in the healthy brain and (2) characterize functional deficits in disease models of brain activity. We also use various machine learning tools such as Artificial Neural Networks (ANNs) and dimensionality reduction techniques in the preprocessing of our neurophysiological data.

Open Science

Being strongly convinced that open data and open source software is a major driver and a necessary condition for scientific progress, I am commited to Open Science in my research.

In the era of data-intensive research, this however comes at the cost of important efforts during the research process. Data needs to be standardised and softwares need to be shared appropriately. I do spend a lot of energy in the design and implementation of such processes (see Softwares). All of the code and data of my research is made to be directly uploaded to modern research material platforms (such as Ebrains) independently from manuscript publication in scientific journals.

Data Engineering

We do data-intensive research. Our experimental approach consists in recording neural activity in the neocortex during behavior using optical imaging (e.g. 2-photon calcium imaging) and electrophysiology (e.g. Neuropixels probe recordings). In practice, this means the processing of massive amounts of data on a daily basis (e.g. the recordings made in Van Velze et al. corresponds to the processing of >1TB per day).

Part of my work consists in the design and maintenance of the preprocessing pipeline for such data.

Software Engineering

Our original experimental approach and our data management strategy requires the development of custom softwares.

I am the lead developer of physion: a full software suite for neurophysiology in the context of visual processing in behaving mice. I am a contributor to fairgraph a high-level API for metadata management in neuroscientific research. I developed several python packages for Data Science applications in the context of neurophysiology.

My full software production is available on my Github profile: github.com/yzerlaut