Error using Nipype FSL example

1

Hi all,

I’m writing a script for first- and second-level modeling using this example from the Nipype documentation. Here’s the relevant bit of code from that example:

modelfit = pe.Workflow(name='modelfit')

modelspec = pe.Node(interface=model.SpecifyModel(), name="modelspec")

level1design = pe.Node(interface=fsl.Level1Design(), name="level1design")

modelgen = pe.MapNode(interface=fsl.FEATModel(), name='modelgen',
                      iterfield=['fsf_file', 'ev_files'])

modelestimate = pe.MapNode(interface=fsl.FILMGLS(smooth_autocorr=True,
                                                 mask_size=5,
                                                 threshold=1000),
                           name='modelestimate',
                           iterfield=['design_file', 'in_file'])

conestimate = pe.MapNode(interface=fsl.ContrastMgr(), name='conestimate',
                         iterfield=['tcon_file', 'param_estimates',
                                    'sigmasquareds', 'corrections',
                                    'dof_file'])

modelfit.connect([
    (modelspec, level1design, [('session_info', 'session_info')]),
    (level1design, modelgen, [('fsf_files', 'fsf_file'),
                              ('ev_files', 'ev_files')]),
    (modelgen, modelestimate, [('design_file', 'design_file')]),
    (modelgen, conestimate, [('con_file', 'tcon_file')]),
    (modelestimate, conestimate, [('param_estimates', 'param_estimates'),
                                  ('sigmasquareds', 'sigmasquareds'),
                                  ('corrections', 'corrections'),
                                  ('dof_file', 'dof_file')]),
])

When I tried running the code above, I couldn’t connect the nodes together. Instead, I got this error:

Exception: Some connections were not found
Module modelestimate has no output called corrections

And, indeed, there isn’t any field called “corrections” in modelestimate.outputs:

Bunch(copes=<undefined>, dof_file=<undefined>, fstats=<undefined>, logfile=<undefined>,
param_estimates=<undefined>, residual4d=<undefined>, results_dir=<undefined>, 
sigmasquareds=<undefined>, thresholdac=<undefined>, tstats=<undefined>, 
varcopes=<undefined>, zfstats=<undefined>, zstats=<undefined>)

How else should I feed the corrections to the contrast manager? For reference, I’m using nipype version 0.13.1.

Thanks!

2

corrections was removed from FSL in recent releases. you can remove that pair.

1 Like
3

Thanks for the quick response! It seems to be working now.

4

Hi here, Apologies for replying to such an old thread. However, I’m running into this issue and the offered solution does not work. I can not find any help/documentation online to solve this problem.
When I run the workflow with corrections added as iterfield for the ContrastMgr() it requires and input. When I then add the connection from the modelestimate module, it raises an error because modelestimate does not generate corrections output (we know this). But then when I remove the corrections from the iterfield and in the connections, it raises an error that ContrastMgr() requires this input, as if it was not deprecated. I have not been able to find a fix. Suggestions?

5

Could you please let us know which versions of FSL and Nipype you are using?

6

Thanks for your fast reply! FSL version 6.0 and Nipype version 1.1.8.

7

those two should work together. however contrast manger is deprecated after FSL 5.0.7, so your workflow would have to use a form like this: