Advanced installation and setup

MNE is written in pure Python making it easy to setup on any machine with Python >=2.6, NumPy >= 1.6, SciPy >= 0.7.2 and matplotlib >= 1.1.0.

Some isolated functions (e.g. filtering with firwin2) require SciPy >= 0.9.

To run all documentation examples the following additional packages are required:

  • PySurfer (for visualization of source estimates on cortical surfaces)
  • scikit-learn (for supervised and unsupervised machine learning functionality)
  • pandas >= 0.8 (for export to tabular data structures like excel files)
  • h5py (for reading and writing HDF5-formatted files)

Note. For optimal performance we recommend installing recent versions of NumPy (> 1.7), SciPy (> 0.10) and scikit-learn (>= 0.14).

Development Environment

Note that we explicitly support the following Python setups since they reflect our development environments and functionality is best tested for them:

  • Anaconda (Mac, Linux, Windows)
  • Debian / Ubuntu standard system Python + Scipy stack
  • EPD 7.3 (Mac, Linux)
  • Canopy >= 1.0 (Mac, Linux)

CUDA Optimization

If you want to use NVIDIA CUDA for filtering (can yield 3-4x speedups), you’ll need to install the NVIDIA toolkit on your system, and then both pycuda and scikits.cuda, see:

https://developer.nvidia.com/cuda-downloads

http://mathema.tician.de/software/pycuda

http://wiki.tiker.net/PyCuda/Installation/

https://github.com/lebedov/scikits.cuda

To initialize mne-python cuda support, after installing these dependencies and running their associated unit tests (to ensure your installation is correct) you can run:

>>> mne.cuda.init_cuda() 

If you have everything installed correctly, you should see an INFO-level log message telling you your CUDA hardware’s available memory. To have CUDA initialized on startup, you can do:

>>> mne.utils.set_config('MNE_USE_CUDA', 'true') 

You can test if MNE CUDA support is working by running the associated test:

$ nosetests mne/tests/test_filter.py

If all tests pass with none skipped, then mne-python CUDA support works.

Multi-threading

For optimal performance we recommend using numpy / scipy with the multi-threaded ATLAS, gotoblas2, or intel MKL. For example, the Enthought Canopy and the Anaconda distributions ship with tested MKL-compiled numpy / scipy versions. Depending on the use case and your system this may speed up operations by a factor greater than 10.

matplotlib

For the setups listed above we would strongly recommend to use the Qt matplotlib backend for fast and correct rendering:

$ ipython --matplotlib=qt

On Linux, for example, QT is the only matplotlib backend for which 3D rendering will work correctly. On Mac OS X for other backends certain matplotlib functions might not work as expected.

IPython notebooks

To take full advantage of mne-python’s visualization capacities in combination with IPython notebooks and inline displaying, please explicitly add the following magic method invocation to your notebook or configure your notebook runtime accordingly.

%matplotlib inline

If you use another Python setup and you encounter some difficulties please report them on the MNE mailing list or on github to get assistance.

Installing Mayavi

Mayavi is only available for Python2.7. If you have Anaconda installed (recommended), the easiest way to install mayavi is to do:

$ conda install mayavi

On Ubuntu, it is also possible to install using:

$ easy_install "Mayavi[app]"

If you use this method, be sure to install the dependencies first: python-vtk and python-configobj:

$ sudo apt-get install python-vtk python-configobj

Make sure the TraitsBackendQt has been installed as well. For other methods of installation, please consult the Mayavi documentation.

Configuring PySurfer

Some users may need to configure PySurfer before they can make full use of our visualization capabilities. Please refer to the PySurfer installation page for up to date information.

Inside the Martinos Center

For people within the MGH/MIT/HMS Martinos Center mne is available on the network.

In a terminal do:

$ setenv PATH /usr/pubsw/packages/python/anaconda/bin:${PATH}

If you use Bash replace the previous instruction with:

$ export PATH=/usr/pubsw/packages/python/anaconda/bin:${PATH}

Then start the python interpreter with:

$ ipython

Then type:

>>> import mne

If you get a new prompt with no error messages, you should be good to go.

We encourage all Martinos center Python users to subscribe to the Martinos Python mailing list:

https://mail.nmr.mgh.harvard.edu/mailman/listinfo/martinos-python