[FieldTrip] BEMCP

RICHARDS, JOHN RICHARDS at mailbox.sc.edu
Thu Mar 17 15:57:47 CET 2016

"The correlation between concentric spheres and FEM (attached) it is not so bad” .  I am not sure what you are correlating.  Is this over the time of a beamformer model, or over participants with a single-time model?  In either case, I think the correlations for the FEM-Concentric spheres are bad, many are less < 0.8, and a number are in the range from .9 to .95.

This result confirms the difference in the FEM and concentric-sphere models.  These are exacerbated even further when you use a subject-specific realistic FEM model and a concentric spheres model.

The results between the BEMCP and concentric spheres suggests that the BEMCP model is not working.  Most of the correlations should be positive.  I have not used correlations as you have, but have done ROIs in an individual, and correlated the values of the ROIs over the head for one subject, or multiple subjects.  I find higher correlations for the BEMCP and Concentric sphere models, than between either of these  and a FEM model.  However, in my case I am not using the standard model from FT, but using subject-specific BEMCP, concentric spheres, and FEM models.

If the FEM model is working for you, why not use it?


John E. Richards Carolina Distinguished Professor
Department of Psychology
University of South Carolina
Columbia, SC  29208
Dept Phone: 803 777 2079
Fax: 803 777 9558
Email: richards-john at sc.edu<mailto:richards-john at sc.edu>
HTTP: jerlab.psych.sc.edu

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20160317/9c5a956f/attachment-0002.html>

More information about the fieldtrip mailing list