Combining beta images for the decoding toolbox (TDT)

tdt
#1

Hi. I’d like to use TDT to analyse data from a factorial fMRI study. I’m interested in comparing two of the conditions (A and B), but because of the factorial design (including conditions X/Y and R/S) I have four beta images for each condition (AXR, AXS, AYR, AYS, etc.). I could reanalyse the data pretending that the other conditions don’t exist, but that feels like bad practice. Is there a way of combining beta images across irrelevant factors for input into TDT?

Many thanks,
Kate

Factorial design with TDT
#2

Hey Kate,

there are a couple of simple options, the simplest one is to simply use the wildcard functionality (*) when calling decoding_describe data. E.g.

decoding_describe_data(cfg,{‘A*’, ‘B*’},[-1, 1],regressor_names,beta_loc,xclass)

will map all conditions starting with A to label -1 and all with B to 1. Then simply proceed as usual.

For more complex condition name patterns, check the regular expression option largely following SPM. E.g.

labelnames{1} = ‘regexp:^cond1 bin[(1)(2)]$’
will find all regressors matching ‘^cond1 bin[(1)(2)]$’

More info in from e.g. “help describe_data”

Best
Kai

2 Likes
#3

Hi Kai

Thanks for the quick response. That sounds like a really simple solution - I was expecting something far more complicated.

Best wishes,
Kate