Freely Moving Animal (FMA) Toolbox
INTRODUCTION
The purpose of the Freely Moving Animal (FMA) Toolbox is to help analyze
electrophysiological and behavioral data recorded from freely moving animals.
These data typically include broadband brain signals (e.g. local field potentials),
spike data (timestamps as well as full waveforms), moment-to-moment position of
the animal, and behavioral events (stimulus presentation, photodetector crossing,
brain stimulation, etc.)
FMAToolbox is part of a larger data analysis framework, including Klusters (a powerful
and easy-to-use cluster cutting application), NeuroScope (an advanced viewer for
electrophysiological and behavioral data), and NDManager (a simple application to
manage recording parameters and data preprocessing).
INSTALLATION
The easiest way to install FMAToolbox is using the pathtool. Click 'Add with subfolders...'
and select the 'FMAToolbox' directory.
For efficiency reasons, the toolbox includes a few functions in C/C++. These can be
easily compiled for your platform. In the command window, type:
>> compilefma;
or simply click here.
Some functions in the toolbox depend on other Matlab toolboxes:
* xmltree to read XML parameter files
* chronux to perform time-frequency analyses
* mYm to provide database functionality
* export_fig (optional) to export figures as PNG, EPS, PDF, etc.
LICENSE
This toolbox was developed by Michaël Zugaro (CRNS-Collège de France, Paris,
France). It is free software distributed under the General Public License (GPL).
CONTENTS
The toolbox is organized in distinct functional categories:
Data - Experimental data handling
General - General-purpose data processing and statistical tests.
Analyses - Specialized analyses (PSTH, place maps, tuning curves...)
Plot - Enhanced and specialized plotting functions.
Database - Database access and storage
IO - Input/output helper functions, including batch handling
OVERVIEW
Functions provided in the Data category can be used to read experimental data files.
The file format is the same as in other programs such as Klusters, NeuroScope and
NDManager (see e.g. http://neuroscope.sourceforge.net/UserManual/data-files.html).
However, the rest of the toolbox is independent from these file formats, so you
are free to use any format you like (of course, you would then have to add your own
code to read your files).
Upon loading, the data are stored in simple matrices where the first column contains
timestamps and the remaining columns contain the corresponding values. For instance,
moment-to-moment positions of the animal are stored as
0 100.625 74.375
0.0256 99.375 75
0.0512 99.375 75
0.0768 100.625 75
0.1024 100.625 75
... ... ...
where the first column is time (in seconds), the second is the X coordinate of the
animal, and the third is its Y coordinate. Positions stored in such a matrix are
referred to as position 'samples'.
Many functions in FMAToolbox provide basic processing of samples, such as time
selection, interpolation or smoothing. These are grouped in the General category,
together with statistical measures and tests, interval handling, binning, etc.
More specialized analysis functions are grouped in the Analyses category, including
peri-event time histograms (PETH), place maps, tuning curves, spectrograms, current
source density, ripple detection, etc.
Results can be plotted using enhanced or specialized functions provided in the Plot
category, and stored in a local or network database using functions in the Database
category.
Last, the IO category contains low-level disk access functions (which should normally
not be used directly), as well as batch-processing handling functions.
GETTING STARTED
A typical data analysis session would start by loading the experimental data:
>> SetCurrentSession
A dialog pops up, where you can navigate your disk to choose the recording session to
analyze. Files are then loaded and kept in memory. You can now access the data very
easily. For instance, if you need the position samples, do:
>> p = GetPositions;
To plot the firing map of the place cell corresponding to cluster 5 on tetrode 3:
>> s = GetSpikes([3 5]);
>> map = FiringMap(p,s);
>> figure;
>> PlotColorMap(map.rate,map.time); % time is used to dimm the colors
To plot theta phase precession information (phase as a function of position, phase
distribution as a function of position, and phase as a function of firing rate) for
the same neuron:
>> lfp = GetLFP(20); % theta is measured on channel 20
>> theta = FilterLFP(lfp,'passband','theta');
>> phi = Phase(theta);
>> precession = PhasePrecession(p,s,phi);
>> figure;
>> PlotPhasePrecession(precession);
Now let us plot a spike raster synchronized on stimulation events:
>> stims = GetEvents('stimulation');
>> [sync,i] = Sync(s,stims);
>> figure;
>> PlotSync(sync,i);
and the corresponding PSTH and time-varying spike distribution:
>> [h,t] = SyncHist(sync,i);
>> figure;
>> bar(t,h); % PSTH
>> [h,t,n] = SyncHist(sync,i,'mode','dist');
>> figure;
>> PlotColorMap(h,1,'x',t,'y',n); % distribution
For more information, read the help topics listed above (CONTENTS).
NOTE
You may wish to add the following line to your startup file:
Browse('on');
This will automatically activate browsing features in every figure.