Characterizing multiple independent behavioral correlates of cell firing

Burgess N, Cacucci F, Lever C, O'Keefe J (2005) Characterizing multiple independent behavioral correlates of cell firing in freely moving animals. Hippocampus.

The Matlab program pxd.m illustrates the use of a factorial Maximum Likelihood model in separating the relative influences of location and orientation on cell firing. The sample data files listed below contain spike and dwell-time data from a place cell (files prefixed "pc"), a head-direction cell (hd) and a theta-modulated place by direction cell (tpd), binned by direction and location. The examples relate to the first cell in each figure in the paper, but the way the data are binned and smoothed here is slightly different.

How to use

Download zip archive (below) and extract all files into a directory on your Matlab path. Start Matlab and type:

> help load_pxd_data

for instructions on loading the example data, and

> help pxd

for instructions on using the pxd algorithm.

Download Zipped Files

(94KB)

Contents

Matlab files

pxd.m
loglikelihood.m
p_estimate.m
d_estimate.m
plot_field.m
plot_dirs.m
find_2dratemap.m
find_circ_ratemap.m
load_pxd_data.m

Data files

pc_c1_pxds.txt
pc_pxdt.txt
hd_c1_pxds.txt
hd_pxdt.txt
tpd_c1_pxds.txt
tpd_pxdt.txt

Author

Neil Burgess, n.burgess@ucl.ac.uk.

A.

B.

C. 

Sample output (from Matlab 6.1, The Mathworks Inc.). A Place Cell (pc); B Head Direction (hd) Cell; C Theta-modulated place by direction (tpd) cell.

In each window, the upper pair of axes show the result of the pxd algorithm. The lower pair show the result of the traditional histogram of spikes over dwell times. Note that the apparent directional selectivity of the place cell, and the apparent locational selectivity of the head direction cell are reduced when pxd algorithm is used. By contrast both types of selectivity remain in the tpd cell.