On Thu, Aug 22, 2019 at 12:36 PM Alexandre PEREZ ***@***.***> wrote: Issue The get_3d_peaks function from nipy.labs.statistical_mapping throws an error when using it with a mask. Minimal reproductible example from nilearn import datasets, masking, imagefrom nipy.labs.statistical_mapping import get_3d_peaks template = datasets.load_mni152_template() mask = masking.compute_gray_matter_mask(template) imgs = datasets.fetch_neurovault_motor_task() img = image.resample_to_img(imgs.images[0], template) get_3d_peaks(img, mask=mask) Error Traceback (most recent call last): File "minimal.py", line 10, in <module> get_3d_peaks(img, mask=mask) File ".../venv/lib/python3.7/site-packages/nipy/labs/statistical_mapping.py", line 199, in get_3d_peaks ff = field_from_graph_and_data(wgraph_from_3d_grid(xyz, k=18), data) File ".../venv/lib/python3.7/site-packages/nipy/algorithms/graph/graph.py", line 527, in wgraph_from_3d_grid raise ValueError('xyz should have shape n * 3') ValueError: xyz should have shape n * 3 Versions nipy==0.4.2 nilearn==0.5.2 Fix xyz must be of shape n*3 to be used by wgraph_from_3d_grid, hence the use of the 3D mask array instead of the 1D ravelled mask array in np.where. ------------------------------ You can view, comment on, or merge this pull request online at:#455 Commit Summary - Fix get_3d_peaks when using mask File Changes - *M* nipy/labs/statistical_mapping.py <https://github.com/nipy/nipy/pull/455/files#diff-0> (2) Patch Links: -https://github.com/nipy/nipy/pull/455.patch -https://github.com/nipy/nipy/pull/455.diff — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#455?email_source=notifications&email_token=AACDE2HUZR4HW2YBFRIGYQDQF26A3A5CNFSM4IOXAHX2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HG2YWJA>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AACDE2EK2BIFP4GC7CWOK3LQF26A3ANCNFSM4IOXAHXQ> .