[FieldTrip] cluster-based statistics with one covariate

[Stephen Politzer-Ahles]

Dear Caroline,

If your conditions are within-subject, then I think the best way to do this
is to construct a dataset representing the conditionA - conditionB
difference for each subject, and then using ft_statfun_indepsamplesregrT to
regress the subject-wise difference on age. See e.g.
https://mailman.science.ru.nl/pipermail/fieldtrip/2016-May/010528.html for
a thread discussing a similar approach.

Best,
Steve


---
Stephen Politzer-Ahles
The Hong Kong Polytechnic University
Department of Chinese and Bilingual Studies
http://www.mypolyuweb.hk/~sjpolit/
<http://www.nyu.edu/projects/politzer-ahles/>



> Message: 3
> Date: Mon, 10 Apr 2017 14:50:13 +0200 (CEST)
> From: Caroline Beese <beese at cbs.mpg.de>
> To: fieldtrip at science.ru.nl
> Subject: [FieldTrip] cluster-based statistics with one covariate
> Message-ID:
>         <1329277450.115066.1491828613491.JavaMail.zimbra at cbs.mpg.de>
> Content-Type: text/plain; charset=utf-8
>
> Dear all,
>
> I'm trying to take age as a covariate into my cluster-based statistics
> comparing two dependent measures.
>
> I have read about two options:
>
> A) to use ft_regressconfound before I do the stats or
> B) to create a design with a third row using the .cvar field in the
> 'depsamplesregrT' statfun.
>
> I understand that for correcting head motions ft_regressconfound is a
> great option but for taking age into account as a covariate, it doesn't
> seem intuitive to me to do this within-subject on a trial-by-trial basis
> having only one number per person. Or can you use this function differently?
>
> However, for option B) the  subfunction 'unitselvec' is missing and this
> seems to be an issue for several years. So I was wondering whether anyone
> has ever tried to fix this problem themselves and wouldn't mind sharing?
>
> More generally, if anyone has successfully implemented a cluster-based
> statistics with one or more covariates, please share your experience, it
> would be greatly appreciated.
>
> Best Regards,
> Caroline
>
> --
> Caroline Beese
> PhD student
>
> Max-Planck Institute for Human Cognitive and Brain Sciences
> Department of Neuropsychology
> Stephanstr. 1a
> 04103 Leipzig
>
> office phone: +49 (0) 341 9940 129
>
> www.cbs.mpg.de/staff/beese-11586
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20170412/af43a0d6/attachment-0002.html>