Home Data Plots HELP

Network activity of developing cortical cultures in vitro


This is the data set from D. A. Wagenaar, J. Pine, and S. M. Potter, 2006, "An extremely rich repertoire of bursting patterns during the development of cortical cultures", BMC Neuroscience, 7:11. Downloading from this site is subject to the condition that you acknowledge our work in any publications that use our data, by citing this article.

We make no guarantees of any kind as to the usability or quality of these data, or the software and code fragments contained on this side. We have tried to copy the data used for the article faithfully into this archive, but some files may have been lost or damaged. We welcome reports on such issues, but cannot accept liability for any problems you experience as a result, directly or indirectly, of using this site. All software and code fragments are distributed under the terms of the GNU General Public License, version 2.


File formats

"Full info"-files

These contain full information about detected spikes, including 3ms of "context": raw electrode data around the time of the spike. These files are stored gzipped, so you need to "gunzip" them before use.

Use in C/C++

"Full info"-files are binary, and consist of a sequence of "Spikeinfo" structures, defined as follows:

typedef struct {
    unsigned long long time;    /* 8-byte integer, little-endian */
    short channel;/* 2-byte integer, little-endian */
    short height;
    short width;
    short context[74];
    short threshold;
    } Spikeinfo;


There are no headers in the files. Thus, one could read the information pertaining to the first 100 spikes in a file as follows:

Spikeinfo buffer[100];
int n;
FILE *fh

fh = fopen("filename.spike","rb");
n = fread(buffer,sizeof(Spikeinfo),100,fh);

(This assumes that "Spikeinfo" has been defined exactly as above.)

Use in Matlab

To load "Full info"-files into Matlab, you could use code like this:

fid = fopen('filename.spike','rb');
if (fid<0)
error('Cannot open the specified file');
raw = fread(fid,[82 n],'int16');
ti0 = raw(1,:); idx = find(ti0<0); ti0(idx) = ti0(idx)+65536;
ti1 = raw(2,:); idx = find(ti1<0); ti1(idx) = ti1(idx)+65536;
ti2 = raw(3,:); idx = find(ti2<0); ti2(idx) = ti2(idx)+65536;
ti3 = raw(4,:); idx = find(ti3<0); ti3(idx) = ti3(idx)+65536;
spikes.time = (ti0 + 65536*(ti1 + 65536*(ti2 + 65536*ti3))) / 25000;
spikes.channel = raw(5,:);
spikes.height = raw(6,:) * .3335;
spikes.width = raw(7,:) * .040;
spikes.context = raw(8:81,:) * .3335;
spikes.thresh = raw(82,:) * .3335;

This fragment loads n spikes from a file called 'filename.spike', and stores them in a structure called 'spikes'. Time is converted to seconds; height, context, and threshold to μV; and width to ms. A Matlab function to load "Full info"-files conveniently is loadspike.m.

Channel numbers

Channel numbers in "Full info"-files refer to electrode numbers. These are related to actual positions in the MEA as follows:


Conversions from electrode number to column-row format can be made in Matlab using hw2cr.m. Conversions in the other direction can be made using cr2hw.m. In C/C++, conversions can be made using the functions defined in ChannelNrs.H.

Compact ".spk.txt" files

These contain only time and channel information, no context. Files have two columns for recordings of spontaneous activity:

For recordings of evoked activity, they contain four columns:

These files are stored bzipped, so you need "bunzip2" to access them.

Compact ".tri.txt" files

These contain information about stimuli, one line per stimulus. They have two columns:

These files are stored bzipped, so you need "bunzip2" to access them.

Compact Matlab files

These contain the information from .spk.txt and .tri.txt files in a more convenient form for access from Matlab. They always contain a variable "spk", with the information from the ".spk.txt" file. For recordings from stimulus-evoked activity, they also contain a variable "tri", with the information from the ".tri.txt" file. You need Matlab version 7 or later to read these files.

Downloading tips

You can download indidvual files, or click on the "list" links to download a list of all the files in a row, in a column, or in an entire table. This list can be used with programs like wget to conveniently download multiple files. Many Linux distributions ship with "wget". Solutions for other operating systems may be available, but we cannot provide them.

If you want to download many files, clicking on each may be inconvenient. In that case, it is recommended that you use the file lists at the bottom and right edges of the tables. In Linux, those can be used as follows:

wget -I daily.spont.dense.full.0.0.3.list

You are also welcome to directly access the archive, e.g., to download all "Compact (text)"-files for daily recordings of spontaneous activity from cultures of any density:

wget -x -r http://potterlab.bme.gatech.edu/devel-data/simple-text/daily/spont

The directory structure is:



If you do choose to download entire directories in this manner, please keep in mind that the entire archive is about 45 GB in size.

(C) Daniel Wagenaar, 2005