Workshop on Local Field Potentials

Manchester 4th April, 2008

Organisers: Rasmus Petersen, Stefano Panzeri, Marcelo Montemurro

This one-day meeting was the first “hot topic workshop” run under the auspices of the UK Mathematical Neuroscience Network. It is becoming increasingly clear that local field potentials (LFPs) can tell us a great deal about neuronal computation. However, proper analysis and modelling of LFPs presents mathematical and computational challenges, which makes a cross-disciplinary approach potentially highly fruitful. The aim of the meeting was to bring together a mix of researchers, representing mathematical, computational and experimental approachers and to present the latest developments in this rapidly developing field. The meeting was fully booked and attended by 80 participants.

Stefano Panzeri (University of Manchester and IIT Genova) opened the meeting and reported his joint work with the group of Nikos Logothetis on the role of LFPs in sensory information coding. He reported that, in visual cortex, the power of low frequency LFP oscillations conveys information about rich natural stimuli which is indenpedent from the information carried by spikes recorded from the same electrode, whereas LFPs in the gamma frequency range carry information which is largely redundant to that carried by spikes. Jozsef Csicsvari (University of Oxford) reported his work about LFP recordings in behaving animals, which revealed the impotance of network oscillations in organising hippocampal cell assemblies and in memory and spatial navigation.

After lunch, Alain Destexhe (CNRS. Gif-sur-Yvette) talked about his work on the power spectrum of cortical LFPs and its relation to cortical activity, presentign evidence that the power law frequency scaling of LFPs reflects important information about the state and operation of the cortical network. Nicolas Brunel and Alberto Mazzoni (University of Paris V and ISI Turin) reported their modeling work, done in collaboration with Stefano Panzeri (Manchester) on the encoding of information about dynamics inputs by sparsely connected networks network of excitatory and inhibitory spiking neurons. They found that slow input frequencies are encodeded by this network as slow frequency LFP fluctuations that are locked to the slow input dynamics, whereas gamma-range LFP oscillations encode information about the input average spike rate.

In the third and final session, Stuart Baker (University of Newcastle) reported his studies on beta band oscillations in primary motor cortex. Using directed cohererence analysis, he showed that these oscillations are mediated by bidirectional interactions between cortex and the periphery, and may have a role in sensorimotor integration. Rosalyn Moran (UCL) gave the final talk of the meeting, in which she presented a study of the use of nerual mass models to analyze LFPs recorded in different structures and to gain quantitative insights into synaptic activity in vivo.

Report prepared by Rasmus Petersen June 2nd, 2008.