[Mne_analysis] Strange patterns of temporal decoding results

Ning Mei nm2241 at nyu.edu
Fri Jan 25 07:36:34 EST 2019
Search archives:

        External Email - Use Caution        

Dear all,

I want to implement the temporal decoding in the sensor space of a MEG data
(
https://mne-tools.github.io/0.16/auto_examples/decoding/plot_decoding_time_generalization_conditions.html#sphx-glr-auto-examples-decoding-plot-decoding-time-generalization-conditions-py),
but the results of either the individual subjects or the average of all
subjects have long horizontal lines, suggesting at some certain time
points, the trained linear SVM classifier can generalize the learned
patterns to all the time points, and at the same time there are neighboring
and horizontal stripes showing no decoding at all, both seem to oscillate.

Here is an example of how I did:
https://colab.research.google.com/drive/1C1ZKn8Z0jxpaaOhw0v6P8SEzorYET9fR.

Please let me know if I did anything wrong in some of the steps.

P.S.
For some reason, in our lab, we are still using MNE-python 0.16.2
instead of the latest one, and hopefully, this won't be an issue.

Thanks!

Ning Mei
Predoctoral Researcher BCBL
www.bcbl.eu

Legal disclaimer/Aviso legal/Lege-oharra: www.bcbl.eu/legal-disclaimer
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20190125/c1ec43f9/attachment-0001.html 


More information about the Mne_analysis mailing list