scikit-learn, Nilearn, machine learning based on brain imaging / Octave?

josheegjosheeg
edited March 2019 in Research
[original thread title(s): "arduino seems to have a octave connection option. Octave seems to have the ability to make .nii. EEG data from arduino to nii file for ni kit learn I think octave could be used but any suggestions?"] edited to try to guess subject of question.

Hi arduino seems to have a octave connection option. Octave seems to have the ability to make .nii files for something like ni kit learn it is a new thing based on sci kit learn. and has the option of doing machine learning on brain imaging. 

Does any one have experiance or ideas how this would be done or best free programs to do it? 

Comments

  • wjcroftwjcroft Mount Shasta, CA
    edited March 2019
    Hi Josh,

    I think you are referring to these packages?



    and .nii, NIFTI file format,


    ---

    I have no experience with these packages, but they seem all related to machine learning from MRI imagery. Do any of these packages work with EEG data? OpenBCI is EEG only. There are some excellent Python EEG analysis systems, open source, such as MNE.


    But note, for EEG brain imaging you need to do what is called "source localization". And that requires significant numbers of EEG channels. In the neurofeedback field a "full cap" is used with 19 channels. And channel counts go up from there: 32, 64, etc. OpenBCI 16 channels might be feasible with some work.

    It sounds like you are also looking for some type of open source alternative to Matlab, such as Octave. But all the packages above are Python based. So I'm a bit confused. OpenBCI board firmware is developed with Arduino libraries, but typically most users do not develop custom firmware. Instead they do their signal processing on the laptop.

    Regards,

    William
Sign In or Register to comment.

© Vanilla Keystone Theme 2026