Describe the issue:
Hi,
I am seeing two related issues that seem to come from the ROI merge logic in Suite2p.
- ROI merge in the GUI throws an index error
When merging ROIs in the GUI I occasionally get:
IndexError: index XXXX is out of bounds for axis 0 with size YYYY
This happens inside merge_activity_masks, where the merged ROI tries to reference traces using an index that no longer exists.
- stat becomes inconsistent after running Suite2p
In the resulting Fall.mat, most ROIs have the full set of fields (including xext, yext, neuropil_mask),
but several ROIs created or modified around merge operations are missing these fields.
This results in a stat list where some entries have fewer fields than others, which causes failures in downstream analysis.
Observation
Both issues appeared in the same datasets and seem to point to a shared root cause:
ROI merge operations do not fully update all ROI metadata and indices.
Question
Is this a known issue with the ROI merge/update pipeline?
And is there a recommended workaround to ensure that merged ROIs have consistent indices and complete metadata?
Happy to provide a small example or Fall.mat snippet if needed.
Thanks!
Reproduce the code example:
IndexError: index XXXX is out of bounds for axis 0 with size YYYY
Example input dataset used to reproduce the issue:
No response
Error message:
Version information:
0.14.4
Context for the issue:
No response
Describe the issue:
Hi,
I am seeing two related issues that seem to come from the ROI merge logic in Suite2p.
When merging ROIs in the GUI I occasionally get:
IndexError: index XXXX is out of bounds for axis 0 with size YYYY
This happens inside merge_activity_masks, where the merged ROI tries to reference traces using an index that no longer exists.
In the resulting Fall.mat, most ROIs have the full set of fields (including xext, yext, neuropil_mask),
but several ROIs created or modified around merge operations are missing these fields.
This results in a stat list where some entries have fewer fields than others, which causes failures in downstream analysis.
Observation
Both issues appeared in the same datasets and seem to point to a shared root cause:
ROI merge operations do not fully update all ROI metadata and indices.
Question
Is this a known issue with the ROI merge/update pipeline?
And is there a recommended workaround to ensure that merged ROIs have consistent indices and complete metadata?
Happy to provide a small example or Fall.mat snippet if needed.
Thanks!
Reproduce the code example:
Example input dataset used to reproduce the issue:
No response
Error message:
Version information:
0.14.4
Context for the issue:
No response