Inferring the location of a rat from neuronal signals
Daisuke Endo and Shigeru Shinomoto (Kyoto University, Japan)

Motion of a rat

• Data1
• Data2

real space

place cells

30 60

activated place cells (synthetic)

Description of the analysis

We developed an analytical method of inferring the location of an animal from the neuronal signals. In particular, we employed the Hidden Markov Model (HMM) that maps a series of observation into a transition among hidden states. Here the observation is a set of spike counts of place cells. To test the ability of the HMM for estimating the location of the animal, we have generated spikes using synthetic place cells.

The animal locomotion data was obtained from the CRCNS (Collaborative Research in Computational Neuroscience) data sharing website, which is providing the brain experimental data. In particular, we have employed "ec014" and "ec016," the experimental data of a rat while it was exploring a square open field of 180cm times 180 cm. We would like to thank professors Kenji Mizuseki and Gyorgy Buzsaki, who have contributed their precious data for public interests.

The analysis was directed by Shigeru Shinomoto. Algorithms were developed by Daisuke Endo.

Version 1.2: 2018/10/15.