Neural Engineering

Transformative Technologies

Work Package: WP5
Programme: P11
Deliverable: 11.1: Neural decoding from 20-neuron patterns with development of a Matlab toolbox

Deliverable due date: month 24

This document reports on the progress of work on Deliverable 11.1. The planned task has been completed but it has yet to be applied to
in vivo data. We studied how synaptic activity in small ensembles of neuron (~20 per ensemble) could give rise to particular activity patterns. The data provided by our Japanese collaborator were insufficient to produce a fully article so we instead presented a poster at FENS 2014 in relation to their data (http://fens2014.meetingxpert.net/FENS_427/poster_101436/program.aspx/anchor101436). We also submitted (and had accepted) an abstract to the selective and prestigious Cosyne 2015 conference (http://cosyne.org/cosyne15/Cosyne2015_program_book.pdf). Following this success, an article has been recently submitted to “Neural Computation” studying how activity in ensembles of neurons can shape the spatial distribution of active synapses. A pre-print version of this publication is available online at (http://www.biorxiv.org/content/early/2016/01/31/029330). All the code used in this work will be freely available upon publication of the manuscript on the git hub page of the main author (https://github.com/rcaze).
In the aforementioned publications we simulated activity in multiple small ensembles of neurons (~20 neurons). We generated correlations in these populations and demonstrated how this could explain multiple recent experimental observations. Notably, this activity shapes how synapses distribute on the post-synaptic neuron. This study makes testable predictions and can help to better understand data from the recordings of neuron population activity.
In such a way we accomplish Deliverable 11.1: “Neural decoding from 20-neuron patterns with development of a Matlab toolbox”. We used a free open-source software package (Python) to perform our study as Matlab needs to be purchased, thus extending the breadth and availability of our study.
The work also covers milestone 6.

Contributors: Romain Caze, Claudia Clopath, Amanda Foust, Simon Schultz, K Kitamura, M Kano (Imperial, University of Tokyo)