ICA output charts

[stjep]

Trying to understand the output that is generated by the --ica flag, but I'm having a hard time interpreting the plots next to each component.

The brainmaps look similar enough to what MELODIC generates that I'm (reasonably) okay with those.

It's unclear to me what the chart next to each component is indicating as there is no key or legend. I checked the docs and there is no mention of ICA other than highlighting that the flag exists. I also had a quick look in the code, which usually points out relevant papers, but couldn't find anything after a cursory look.

[chrisgorgo]

Indeed the documentations should be improved. Indeed we are running melodic to generate thos components. The graphic on the right is the unthresholded spatial map of the component (with a black overlay of the brain edge for reference). On the right in blue is the timecourse (it should be in seconds, but relies on correct RepetitionTime value in the sidecar JSON file (see BIDS spec) which seems to be off in your case - maybe you specified RT in msec instead of sec?). The plot in orange is the frequency decomposition (should be in Hz but also seems off in your case).

[chrisgorgo]

One more point the spatial maps are maximum intensity projection (in other words "glass brains").

[stjep]

Thanks @chrisfilo. My TR was set to 2.0 (2000 ms), but I think my slice acquisition times are off, so I'm going to re-run it with that removed to see if the axis notation is fixed.

[chrisgorgo]

Does your dataset pass the validator checks ( https://github.com/INCF/bids-validator) with no warnings?

On Mon, Jul 31, 2017 at 2:42 PM, Dan Stjepanović [email protected] wrote:

Thanks @chrisfilo https://github.com/chrisfilo. My TR was set to 2.0 (2000 ms), but I think my slice acquisition times are off, so I'm going to re-run it with that removed to see if the axis notation is fixed.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/poldracklab/mriqc/issues/613#issuecomment-319205009, or mute the thread https://github.com/notifications/unsubscribe-auth/AAOkp5qBl93ytflNADBh2cQTWd9G5mBcks5sTkoygaJpZM4Oo8hC .

[stjep]

I'm just testing this out on 1-2 functional files, but they do pass the BIDS validator. Anything else I could try to pin this down?

[chrisgorgo]

Could you share the content of the relevant JSON file?

On Mon, Jul 31, 2017 at 3:40 PM, Dan Stjepanović [email protected] wrote:

I'm just testing this out on 1-2 functional files, but they do pass the BIDS validator. Anything else I could try to pin this down?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/poldracklab/mriqc/issues/613#issuecomment-319216859, or mute the thread https://github.com/notifications/unsubscribe-auth/AAOkp_wshKiC9XI1qWurkJnL79r3P1Veks5sTlfogaJpZM4Oo8hC .

[chrisgorgo]

I think it actually might be a bug in niworkflows https://github.com/poldracklab/niworkflows/issues/191

[stjep]

Ah, (somewhat) glad to hear it's not just me.

For completion sake, the contents of JSON file are below. It's the absolute bare minimum to be able to push this data through (I'm testing different sequences, so the data should be clean):

{
    "RepetitionTime": 2.0,
    "TaskName": "rest",
    "Instruction": "Lie still and keep your eyes open"
}

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.