From The CNBC Wiki
Revision as of 01:24, 14 April 2013 by Tarrlab (talk | contribs) (Image Similarity Toolbox)

Jump to: navigation, search

Various bits of code that might help you run or analyze an experiment... Also see our Resources Page

Science Code Manifesto


(.zip archive) is a random collection of routines written by Michael Tarr to work with PsychToolbox. The key routine is readdata which makes reading in text files somewhat less painful. Now includes an improved readdata routine (v 1.3).

Image Similarity Toolbox

A Matlab toolbox for computing image similarities using a variety of computer vision algorithms. Code written at Tarrlab by Darren Seibert and Daniel Leeds. Available for download through GitHub and as a zip file


Code and manual for: Willenbockel, V., Sadr, J., Fiset, D., Horne, G. O., Gosselin, F., & Tanaka, J. W. (2010). Controlling low-level image properties: The SHINE toolbox. Behavior Research Methods, 42(3), 671-684.


Code and manual for: Armstrong, B. C., Watson, C. E., & Plaut, D. C. (2012) SOS! An algorithm and software for the stochastic optimization of stimuli. Behavior Research Methods.


Code, manual, and norms for: Armstrong, B. C., Tokowicz, N., & Plaut, D. C. (2012). eDom: Norming software and relative meaning frequency norms for 544 homonyms. Behavior Research Methods.

|Stat ("pipestat" or unixstat)

One of the easiest to use stat packages around and it is the free. It is entirely command-line based, deceptively simple, and the best tool out there for fast data manipulation. Gary Perlman (|STAT's author) distributes both DOS and UNIX versions for free, the links below are for a recompile I did for OS X.
OS X binaries (zipped tar archive). I have recompiled the source for Mac OS X if anyone is interested. I never heard back from Gary about distributing the OS X versions, so I have made the binaries and the associated man pages available. Included in this distribution are several variations on the original version. Specifically, I have created four "new" commands: dl, ttest, mpttest, and a new, improved anova (that reports true n's and can automatically compute cell means by subjects). The man page for anova reflects this change. dl is self-documenting via "dl -h" and ttest and mpttest are simply that - they take two columns of numbers and return the appropriate statistic. I recommend putting these binaries in /usr/local/bin (which should already be in your PATH, but if not add it). NOTE: These have recently been recompiled for 64 bit Intel Macs - they seem to work fine, but they have not been extensively tested.
man pages (zipped tar archive). I recommend putting these man pages in /usr/local/share/man/man1 (which you definitely will need to add to your MANPATH).


Lots of code from Tommy Poggio's group at MIT, including implementations of HMAX and other interesting stuff.

HMAX Code from Max Riesenhuber

Useful page summarizing many different packages/tools for use in visual psychophysics.