Minischool

Minischool: Analysis methods for simultaneous electrophysiological recordings from a large number of electrodes (LFPs, spikes)

The purpose of this minischool is to give the audience an overview of data properties and management in multielectrode recordings and to introduce analysis methods or tools classical to the field. Courses will be theoretical (no hands-on sessions) but there will be a massive use of real data for illustration purpose.

09:15 – 12:00                     Recording neural activity: analyzing data is misleading when data itself is not well understood. The morning session aims at describing the origin of signals recorded on electrodes and how to obtain clean spiking data and local field potentials. There will be a focus on spike sorting methods. In particular but not exclusively, we will talk about detection of events, characterization of noise properties, choice and quality assessment of unsupervised classification methods as well as display of multidimensional data.

Speakers: Clément Léna (ENS Paris, 9h15-10h30), Christophe Pouzat (UMR8145, CNRS, Univ. Paris 5, 10h45-12h)

12:00 – 13:00                     Lunch

13:00 – 17:00                     Data analysis and management: our purpose is to introduce basic to advanced standardized indices or methods in spiking data or local field potentials analysis. We will start from simple statistics describing spiking data (PSTH, ISI, latencies, autocorrelation) and LFP (spectral content, peaks analysis) as well as some functional relationships between these two different kinds of neural activity. Bivariate and multivariate classical methods and concepts applying to multiarray recordings of spikes and LFPs will then be assessed: cross-correlation, coherence, Granger causality, principal and independent component analysis. We will also make an introduction to information theory and in particular information carried by one or several spike trains about some stimuli. We will talk briefly about the most used softwares in the field for data analysis. Finally, we will give perspectives on how handle data from neurophysiology efficiently: databasing, data sharing, best practices in computational neuroscience, modelling and fitting neuronal networks.

Speakers : Boris Gourévitch (UMR8195, CNRS, Univ. Paris-Sud, 13h-14h15), Alain de Cheveigné (ENS Paris, 14h30-15h45), Andrew Davison (UPR3293, CNRS, Gif-sur-Yvette, 16h-17h).

15 minutes pause between speakers.