announce.txt

Dear MEA-Users,

We are pleased to announce the first public release of MeaBench, an
open-source suite of programs for acquisition and analysis of
multi-electrode array (MEA) recordings. MeaBench was developed by
Daniel Wagenaar at Caltech, drawing on the excellent example set by
MultiChannel Systems' MC_Rack suite (www.multichannelsystems.com).

The software runs under Linux and other Unix variants, and is freely
distributable under the terms of the GNU Public License (see
http://www.gnu.org/copyleft/gpl.html). It offers the following
functionality:

  - Acquisition of raw electrode data from MultiChannel Systems' MCCard;
  - Complete removal of mains (60 Hz) interference using template
    filtering;
  - Removal of stimulation artifacts using the SALPA algorithm [1];
  - Online and offline detection of spikes;
  - Online visualization of electrode data and spikes;
  - Continuous or windowed [saving] of raw data and spikes;
  - Saving of spike waveforms, for later spike sorting and analysis;
  - Replaying of raw and spike files, at any speed;
  - Instant-replay buffer for easy analysis of recent events;
  - Online generation of raster plots;
  - Continuous monitoring of varying noise levels;
  - A variety of utilities for analysis and data format conversion,
    including:

      * Averaging of electrode recordings over trials;
      * Conversion of binary spike files to ASCII representation;
      * Filtering of spike files based on any mathematical expression
        involving shape or timing parameters;
      * Extraction of single channels from 64 channel streams;
      * Splitting of long data files into trials;
      * Splitting of long data files into channels;
      * Computing spike rates;
      * Detecting culture-wide bursts.

  - A selection of Matlab functions to import MeaBench data is available
    upon request. (Users may also be interested in Uli Egerts comprehensive 
    set of matlab code for MEA data analysis, available for download from
    http://www.brainworks.uni-freiburg.de/projects/mea/meatools/overview.htm.)
  
MeaBench is fully modular, and any user with some Unix programming
experience can extend it to fit her or his needs. Since MeaBench can
stream live data to your extension modules, it is well suited, for
example, to drive real-time feedback systems. In fact, the ability to
communicate with other software or hardware in real-time was one of
the primary motives for the conception of MeaBench. It allowed a
reliable, sub-100 ms feedback loop time in our Neurally Controlled
Animat. [2]

MeaBench was written primarily for use with the MultiChannel Systems
MEA hardware, and a driver is included for their MCCard data
acquisition card, written by Thomas B. DeMarse with advice from
MultiChannelSystems. If you use different data acquisition hardware,
you may still find MeaBench useful, because, due to its modular
nature, it is possible to write plug-in modules to read data from your
hardware. An experimental driver for one such board (manufactured by
United Electronic Industries, but not endorsed by us at its current
state of development) is included as well.

MeaBench has been in constant use by my group in the Pine lab at
Caltech for over two years, and now also by everyone in my group at
Georgia Tech (http://www.its.caltech.edu/~pinelab/PotterGroup.htm). 
Daniel has worked countless hours creating and improving this very
functional set of tools. It remains a work in progress; we welcome
suggestions for improvement (and bug reports). Please join in the
development by submitting your code (patches and improvements) for
inclusion in future releases.

Daniel wishes to acknowledge much valuable input and support from Tom
DeMarse, Jerry Pine, and myself. We are all grateful for financial
support from the NIH-NINDS and The Burroughs-Wellcome Fund, and
cooperation, technical support, and equipment from MultiChannel
Systems.

For more information and to download MeaBench, please visit:

    http://www.its.caltech.edu/~wagenaar/meabench


With much enthusiasm for the future of MEAs,

Steve M. Potter, PhD
Assistant Professor of Biomedical Engineering
Neuroengineering Laboratory
Coulter Department of Biomedical Engineering
Georgia Institute of Technology and Emory University
http://www.neuro.gatech.edu/potter.php
steve.potter@bme.gatech.edu

References:

[1] D A Wagenaar, and S M Potter: Real-time multi-channel stimulus
    artifact suppression by local curve fitting. J. Neurosci. Meth,
    vol 120 issue 2, 2002 (in press). Preprint available at:
    www.its.caltech.edu/~wagenaar/salpa-preprint.pdf

[2] T B DeMarse, D A Wagenaar, A W Blau, and S M Potter: The neurally
    controlled animat: biological brains acting with simulated bodies.
    Autonomous Robots 11, 2001, pp 305-310.