These files accompany the following publication:

Kubilius, J., Wagemans, J., & Op de Beeck, H. (2011) Emergence of perceptual Gestalts in the human visual cortex: The case of the configural superiority effect. Psychological Science

To guarantee highest possible standards, all stages of scientific conduct must be made transparent. Publishing a paper is not sufficient anymore. Raw data, experimental and analysis scripts must be freely available as well. (Here, raw data is not included due to a massive size of fMRI files.) I believe that full disclosure leads to more collaborative and more reliable science.

(See also Nick Barnes. Publish your computer code: it is good enough.)


  • Python 2.5

  • PsychoPy 1.62, which is dependent on:
    • numpy 1.5
    • scipy 0.8
    • matplotlib 1.0
    • pyglet 1.1
    • pygame 1.9
    • pyOpenGL 3.0
    • Python Imaging Library 1.1
    • wxPython 2.4
    • setuptools 0.6
    • pywin32 (Windows only)
  • PyMVPA 0.4.4 and PyNIfTI 0.2

  • Tabular 0.0.8
    • It produces a number of warnings that might be annoying. To avoid this behavior:
      • tabular/tabular/, Line 30: change DEFAULT_VERBOSITY = 5 to DEFAULT_VERBOSITY = 0
      • tabular/tabular/
        • Line 27: add import as io; DEFAULT_VERBOSITY = io.DEFAULT_VERBOSITY
        • Line 199: put the three if statements under condition if DEFAULT_VERBOSITY > 1:
    • There is also a problem with saving nan's to a text file. Fix:
      • tabular/tabular/, Line 344: change nan to 1.#QNAN (the DEFAULT_STRINGIFIER definition)
    • To include these updates, in the terminal navigate to where tabular is and type:

      $ sudo python install

Getting started

In a console, navigate to the folder where this file is located, then go to scripts and type:

$ python --n exp

This will run the main experimental task without saving any data on your computer. (Press '5' to initiate the experiment.) You can also run localizer by typing:

$ python --n --runType loc exp

or meridian mapping:

$ python --n --runType mer exp

If you collect your own data, you can also replicate analyses. Probably the most interesting is replicating figures and tables in the paper. You can replicate, for example, Figure 3 (without saving anything on your computer) by typing:

$ python --n material f3

To see all available options, type:

$ python --help


Author:Jonas Kubilius
Insitution:K.U.Leuven (Belgium)
License:Modified BSD License
Included packages and functions:
 argparse, combinations, combinations_with_replacement, OrderedDict
Included images:,