devel > confusion on NIFTI_INTENT codes for overlay
Showing 1-3 of 3 posts
May 21, 2008 06:05 PM | Nick Schmansky
confusion on NIFTI_INTENT codes for overlay
Is it invalid to have, in a single GIFTI file, DataArrays
containing multiple intent codes of _SHAPE, _TIME_SERIES and _NONE?
The operative word being AND.
Or, in other words, are _SHAPE, _TIME_SERIES and _NONE mutually exclusive?
Case in point: in freesurfer's tksurfer app (the surface renderer), there is an option to load an overlay file. Lets say the overlay is a time series of functional activation. Now, the example Caret functional file, Human.colin.R.FUNCTIONAL.71723.func.gii, has two data arrays of data with intent code NIFTI_INTENT_NONE. So what I have done in our gifti file code is just search the gifti file for that intent code. But that same data could have been given the NIFTI_INTENT_TIME_SERIES code. So I need to search for that too. And additionally, NIFTI_INTENT_SHAPE is also valid overlay data for tksurfer. So I need to search for that too. So if a file contains multiple combinations of these, then how will my reader code know which is the overlay to load?
It seems a gifti file should be considered valid only if has data arrays of only one of these intent codes, otherwise you can't tell which one to load as an overlay, and you probably don't want an OR statement in the code to load all DataArrays with any of these three intent codes.
Nick
Or, in other words, are _SHAPE, _TIME_SERIES and _NONE mutually exclusive?
Case in point: in freesurfer's tksurfer app (the surface renderer), there is an option to load an overlay file. Lets say the overlay is a time series of functional activation. Now, the example Caret functional file, Human.colin.R.FUNCTIONAL.71723.func.gii, has two data arrays of data with intent code NIFTI_INTENT_NONE. So what I have done in our gifti file code is just search the gifti file for that intent code. But that same data could have been given the NIFTI_INTENT_TIME_SERIES code. So I need to search for that too. And additionally, NIFTI_INTENT_SHAPE is also valid overlay data for tksurfer. So I need to search for that too. So if a file contains multiple combinations of these, then how will my reader code know which is the overlay to load?
It seems a gifti file should be considered valid only if has data arrays of only one of these intent codes, otherwise you can't tell which one to load as an overlay, and you probably don't want an OR statement in the code to load all DataArrays with any of these three intent codes.
Nick
May 21, 2008 10:05 PM | Richard Reynolds
confusion on NIFTI_INTENT codes for overlay
Hi Nick,
I don't have a big personal opinion on whether it should be valid to
have such a combination of INTENT codes in a single dataset (though
it does seem odd to mix TIME_SERIES with any other).
But in case you are using gifticlib (are you?), it has gifti_find_DA()
and gifti_find_DA_list(), that return one or many DA pointers for a
given intent code.
- rick
I don't have a big personal opinion on whether it should be valid to
have such a combination of INTENT codes in a single dataset (though
it does seem odd to mix TIME_SERIES with any other).
But in case you are using gifticlib (are you?), it has gifti_find_DA()
and gifti_find_DA_list(), that return one or many DA pointers for a
given intent code.
- rick
May 22, 2008 07:05 PM | Nick Schmansky
RE: confusion on NIFTI_INTENT codes for over
yes, i am using gifti_find_DA to locate data arrays of particular
intent codes. my question is what to do if a file has multiple
intent codes of the types _NONE, _SHAPE and _TIME_SERIES. if it
does has multiple DataArrays with these differing intent codes,
then it is impossible to know which one to read when it is desired
to load 'overlay' data. currently freesurfer just looks for all
data arrays of that type and loads any and all of those three in a
our internal time series array. but the possibility exists of
having a mix of intent code data types in this list.
