- m@cortex
March 2006
testing m@cortex
include your own sample data in a testing routine to ensure consistency of behavior with new releases of the package. More info here.
the @facts class
A simple object class to accumulate data, produce grouped bar charts, and run simple stats. If you are planning to perform t-tests or n-way ANOVAs on a set of data, then building a @facts object may be a good idea.
Indeed, the new functions @cortex/anovan and @spix/anovan internally create and use @facts objects.
Documentation is still largely incomplete. Please look into the above code for an example of how to use @facts.
new analysis functions
go here for a more or less complete list
key factors
It is now possible to group subsets of keys from a @key object into "levels", and group levels into "factors". Factor definitions are embedded into the object itself, and can be used to perform n-way analysis of variance through @cortex/anovan.
January 2005
the @spix class
I believe that @spix objects can make your life easier when you have a bunch of units collected in the same experiment to analyze. Please, see the tutorial for an introduction.
'nex' data file analysis
I embedded code written by NPB at the NIMH intor the @cortex constructor, which can now read ".nex" data files and create standard @cortex objects. Testing done by ELH at NYU shows that spike data retrieved using the @cortex class are consistent with those obtained with @ctxplx. No tests have been performed so far on analog data. Anyone interested in contributing?
@key objects replace @rules
The @rules class has been replaced by the @key class. Although @rules continue to work as before, their use is deprecated, and I recommend that you change your code accordingly. Upon request, I can explain why I ended up doing this...
In any case, the only real difference between @rules and @key usage is related to "complex" keys (formerly "groups of rules"), i.e., keys that are defined by multiple lines in the source key file. Look up the tutorial to see how things work now.
GIOCODE.DEFault values
The text file giucode/giucode.def contains default values used by several functions. See comments embedded in the file itself for details on each of the valid parameters.
One of the consequences of having a set of default values is that some function calls have a simpler syntax than before. For example, in order to load EOG data contained in a standard CORTEX file into a @cortex object you set some values in 'giucode.def' instead of adding a 3rd and 4th argument to the constructor call.
including "short" trials in spike matrices
When retrieving a spikematrix (either directly of through analysis functions such as meanfiringrate), whenever the epoch requested by the key exceeded the actually available time in a trial, that trial was excluded by the analysis.
Now, by editing the line
includeshorttrials = false;
in the [spikes] section of giucode.def you can force those trials to be included instead. All missing milliseconds are padded with NaNs, and analysis functions should treat those as missing values. This new functionality still needs to be tested thoroughly.
given_response to replace expected_response
There has always been some confusion about the name and meaning of a couple of CORTEX header fields. Look up this page for additional information. From now on, use of the dot-property expected_response in both the @cortex and @key class is deprecated, and replaced by given_response. Parameters in giucode.def also reflect this change.
care for a QUICKIE?
You can quickly display spike density plots for a single unit or for any unit included in a @spix object with a simple GUI. Just do one of:
>> quickie(cobj, cluster, key)
>> quickie(spix)
use it during informal lab presentations!
cortview for m@cortex
You can browse the content of a @cortex object with this simple GUI:
>> cortview(cobj) >> cortview(cobj,kobj)
overloaded ANOVA functions
ANOVA1 is the Stats Toolbox's implementation of 1-way analysis of variance. If, however, you invoke anova1 with a @cortex object as the first argument, the m-file @cortex/anova1.m will be called instead. The function allows you to quickly run mean firing rate comparisons across conditions (i.e., keys).
In the future I might also add a @cortex/anovan function to execute n-factor analyses.
Indeed, ANOVAN is already available for @spix objects.
epp and touchscreen support in the @cortex class
Although the job is not quite finished, @cortex objects now include the ability to store "epp" data, as well as "touchscreen" events. Please do let me know if you intend to use these features.
various 'giutils' available
the folder 'giutils' in the distribution package contains a bunch of small functions, some of which are general-purpose enough that you may want to check them out.