[FieldTrip] Memory-efficient t-testing against zero (or any scalar)

[Stephen Whitmarsh]

Dear Fieldtrippers,

I’ve been hitting a pretty high memory ceiling, so I’m trying to reduce the use of RAM as much as possible.

I am at the point of using (source)statistics, in which I need to (t-)test against zero. Normally, I would just create dummy FT data-structures, in which I replace the parameter that is tested with 0. However, in my case this would result in an asphyxiating (more than) doubling of data in memory.

Has anyone already created a ft_statfun that simply tests against a scalar (for all data points)? Or, if I would need to write my own, which existing ft_statfun would be most convenient do you think? Or perhaps there is another way around it?

Thanks!
Stephen


Stephen Whitmarsh, PhD
PostDoctoral Researcher | NatMEG

Swedish National Facility for
Magnetoencephalography
Office: +46 (0)8 524 833 33
MEG lab: +46 (0)8 524 833 09
stephen at NatMEG.se | NatMEG.se

[NatMEG_POS_resized]

Karolinska Institutet
Department of Clinical Neuroscience
Nobels väg 9, D2:240 | 171 77 Stockholm
Stephen.Whitmarsh at ki.se | ki.se


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20150907/ea244d3b/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image003.png
Type: image/png
Size: 10116 bytes
Desc: image003.png
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20150907/ea244d3b/attachment-0001.png>