search

poldracklab / ds003-post-fMRIPrep-analysis

An exemplary task analysis workflow to run on OpenNeuro's ds000003 data, after preprocessing with fMRIPrep
Apache License 2.0
29 stars 8 forks source link

[ENH] switch to two-tailed test #2

Closed rciric closed 5 years ago

rciric commented 5 years ago

In accordance with recent discussions, this PR implements two-tailed statistical testing for multiple comparisons correction.

The resulting map for FWE ends up being quite conservative, so we might want to consider plotting the cluster map. The expected DM deactivations are present.

Cluster map views:

Screen Shot 2019-05-21 at 2 59 38 AM Screen Shot 2019-05-21 at 3 00 17 AM

FWE map views:

Screen Shot 2019-05-21 at 2 59 50 AM Screen Shot 2019-05-21 at 3 00 23 AM
poldrack commented 5 years ago

this is a small sample so that's not surprising...

On Tue, May 21, 2019 at 3:01 AM Rastko Ćirić notifications@github.com wrote:

In accordance with recent discussions, this PR implements two-tailed statistical testing for multiple comparisons correction.

The resulting map for FWE ends up being quite conservative, so we might want to consider plotting the cluster map. The expected DM deactivations are present.

Cluster map views: [image: Screen Shot 2019-05-21 at 2 59 38 AM] https://user-images.githubusercontent.com/13735271/58087074-af5c5f00-7b74-11e9-90d4-559e21f348e5.png [image: Screen Shot 2019-05-21 at 3 00 17 AM] https://user-images.githubusercontent.com/13735271/58087075-aff4f580-7b74-11e9-9afc-a5bee8161ee8.png

FWE map views: [image: Screen Shot 2019-05-21 at 2 59 50 AM] https://user-images.githubusercontent.com/13735271/58087083-b8e5c700-7b74-11e9-8b61-f547b3600d8c.png [image: Screen Shot 2019-05-21 at 3 00 23 AM] https://user-images.githubusercontent.com/13735271/58087085-b8e5c700-7b74-11e9-8321-7b81b6d45f17.png

You can view, comment on, or merge this pull request online at:

https://github.com/poldracklab/ds003-post-fMRIPrep-analysis/pull/2 Commit Summary

  • (enh) switch to two-tailed test

File Changes

Patch Links:

- https://github.com/poldracklab/ds003-post-fMRIPrep-analysis/pull/2.patch

https://github.com/poldracklab/ds003-post-fMRIPrep-analysis/pull/2.diff

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/poldracklab/ds003-post-fMRIPrep-analysis/pull/2?email_source=notifications&email_token=AAGUVEDUQTBBH7H43FTA7WLPWPCBFA5CNFSM4HOJLXDKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4GU5PSGA, or mute the thread https://github.com/notifications/unsubscribe-auth/AAGUVEGQ5A7NSNEPDBUUUODPWPCBFANCNFSM4HOJLXDA .

-- Russell A. Poldrack Albert Ray Lang Professor of Psychology Professor (by courtesy) of Computer Science Bldg. 420, Jordan Hall Stanford University Stanford, CA 94305

poldrack@stanford.edu http://www.poldracklab.org/

satra commented 5 years ago

@rciric - you could also run a permutation test with PALM on this. would require a function node at this point (no direct interface in nipype yet).

it's really nice to see a lot of "putative" word production regions coming online in this task.

oesteban commented 5 years ago

this is a small sample so that's not surprising...

Indeed, just 13.

it's really nice to see a lot of "putative" word production regions coming online in this task.

@satra Should those regions also activate for the pseudo-word trials? Please note the contrast here is words+nonwords vs. baseline.

poldrack commented 5 years ago

yes, you would expect the pattern to be fairly similar for a rhyme task, probably with more activity for pseudowords

On Wed, May 22, 2019 at 9:22 AM Oscar Esteban notifications@github.com wrote:

this is a small sample so that's not surprising...

Indeed, just 13.

it's really nice to see a lot of "putative" word production regions coming online in this task.

@satra https://github.com/satra Should those regions also activate for the pseudo-word trials? Please note the contrast here is words+nonwords vs. baseline.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/poldracklab/ds003-post-fMRIPrep-analysis/pull/2?email_source=notifications&email_token=AAGUVEDIJO2JQ5EBQQAT4DDPWVXN5A5CNFSM4HOJLXDKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODV7TA2Q#issuecomment-494874730, or mute the thread https://github.com/notifications/unsubscribe-auth/AAGUVEFHIQBPCRZ6XKTPR7LPWVXN5ANCNFSM4HOJLXDA .

-- Russell A. Poldrack Albert Ray Lang Professor of Psychology Professor (by courtesy) of Computer Science Bldg. 420, Jordan Hall Stanford University Stanford, CA 94305

poldrack@stanford.edu http://www.poldracklab.org/