Get Started¶

Prerequisites¶

Using CoSMoMVPA effectively requires:

a working installation of Matlab or Octave.
the CoSMoMVPA source code, available from GitHub; see here for instructions to get you environment ready.
optionally the tutorial data, available here (to run the exercises).
optionally some external toolboxes for AFNI, BrainVoyager, and/or FieldTrip file support; see here.
an advanced beginner level of experience in Matlab programming.
an advanced beginner level of fMRI or MEEG data analysis.
a basic understanding of MVPA concepts.
familiarity with CoSMoMVPA concepts, in particular the Dataset, Targets, Chunks, Dataset operations, Classifier, Neighborhood, and Measure concepts.

Consider the demos to see how MVPA can be performed using CoSMoMVPA.

Get your environment ready¶

To get started, you need the CoSMoMVPA code and optionally the tutorial data; see here for instructions. For the impatient:

CoSMoMVPA Matlab / Octave code:
- git users:
  git clone https://github.com/CoSMoMVPA/CoSMoMVPA.git make -C CoSMoMVPA install
- also available as zip archive.
  In that case, use cosmo set path to set the path.
Tutorial data:
- Complete tutorial data: datadb-v0.3.zip; used for LABMAN 2017 course and NMSM 2019, Noesselt’s lab 3rd Modelling Symposium, University of Magdeburg.
- Subset with fMRI AK6 and MEEG obj-6 data (for PRNI 2016 exercises): tutorial data: datadb-ak6-meg_obj6-v0.3.zip.
- Minimal set (only fMRI AK6 data for exercises): tutorial data: datadb-ak6-v0.3.zip.
CoSMOMVPA documentation:
- git users:
  git clone https://github.com/CoSMoMVPA/cosmomvpa.github.io.git
also available as zip archive <https://github.com/CoSMoMVPA/cosmomvpa.github.io/archive/refs/heads/main.zip>.

Next steps¶

Once you are ready:

run the demos.
look at the runnable examples and the associated Matlab outputs.
try the exercises.
explore the CoSMoMVPA functions.

Some examples of analyses that can be run with CoSMoMVPA: