___ ____ __ __
/ __)( _ \( \/ )
\__ \ )___/ ) ( Statistical Parametric Mapping
(___/(__) (_/\/\_) SPM - https://www.fil.ion.ucl.ac.uk/spm/

Copyright (C) 1991,1994-2025 Wellcome Centre for Human Neuroimaging

The Python interface to SPM

Warning

This project is young and might contain bugs. If you experience any issues, please let us know!

Installation instructions

Important - Windows

Python installation made from Microsoft Store on Windows will not work (raises DeclarativeService.dll not found), install it from Python website.

Important - Python/MATLAB compatibility

Specific versions of MATLAB are compatible with specific versions of Python.

By default, spm-python uses:

• Python 3.6: R2020b
• Python 3.7: R2021b
• Python 3.8: R2023a
• Python 3.9: R2025a
• Python 3.10 - 3.13: R2025b

Option 1 - Install the MATLAB runtime on first use

1. Install SPM-Python

   pip install spm-python

2. Run spm

   spm

3. Follow the instructions

• Advantages
  • Installs the runtime that is required for your python version
  • Does not resintall anything if a compatible runtime already exists
• Drawbacks
  • May need to be run in proviledged mode (e.g., sudo)
  • May be fiddly on Windows

Option 2 - Install the MATLAB runtime manually

1. Install the appropriate MATLAB Runtime
2. Install SPM:

   pipinstallspm-python

• Advantages
  • Graphical interface for installing the runtime
• Drawbacks
  • The correct runtime must be selected for your python version

Option 3 - Install the MATLAB runtime using an installation script

1. Install SPM-Python

   pip install spm-python

2. Run the installer

   install_matlab_runtime --version R2025b --yes

• Advantages
  • Exposes installation options (install_matlab_runtime --help)
  • Allows any runtime version to be installed. One may do:

    pip install spm-python[R2023b]
    install_matlab_runtime --version R2023b

• Drawbacks
  • For advanced users

Minimal example

Here is a minimal set of examples showcasing changes to do to use existing Matlab code in Python (taken from the OPM tutorial).

1. Importing and setting up SPM

In Matlab:

 spm('defaults', 'eeg');

In Python:

fromspmimport*spm('defaults', 'eeg')

2. Constructing objects

In Matlab:

S = [];
S.data ='OPM_meg_001.cMEG';
S.positions ='OPM_HelmConfig.tsv';
D = spm_opm_create(S);

In Python, create a Struct() instead of a struct:

S=Struct()
S.data='OPM_meg_001.cMEG'S.positions='OPM_HelmConfig.tsv'D=spm_opm_create(S)

Here, D will be a meeg object which contains a virtual representation of the Matlab object. Class methods should work as expected, e.g.:

D.fullfile()
>>>'./OPM_meg_001.mat'

Note that the alternative call that exist in Matlab (i.e., fullfile(D)) will not work.

3. Creating and handling figures

In Matlab:

S = [];
S.triallength =3000;
S.plot =1;
S.D =D;
S.channels ='MEG';
spm_opm_psd(S);
ylim([1,1e5])

In Python:

S=Struct()
S.triallength=3000S.plot=1S.D=DS.channels='MEG'spm_opm_psd(S)

This opens a Matlab figure, but we do not have the possibility of manipulating it yet (e.g., calling ylim). As of now, we can view the figures, have GUI interactions, but cannot manipulate figures with Python code.

4. Variable number of output arguments

In Matlab:

S = [];
S.triallength =3000;
S.plot =1;
S.D =mD;
[~,freq] = spm_opm_psd(S);

In Python, the number of output arguments must be specified by the nargout keyword argument:

S=Struct()
S.triallength=3000S.plot=1S.D=mD
[_,freq] =spm_opm_psd(S, nargout=2)

5. Other examples

In Matlab:

S = [];
S.D =D;
S.freq = [10];
S.band ='high';
fD = spm_eeg_ffilter(S);
S = [];
S.D =fD;
S.freq = [70];
S.band ='low';
fD = spm_eeg_ffilter(S);

In Python:

S=Struct()
S.D=DS.freq=10S.band='high'fD=spm_eeg_ffilter(S)
S=Struct()
S.D=fDS.freq=70S.band='low'fD=spm_eeg_ffilter(S)