AllenNeuralDynamics / AllenNeuralDynamics/aind-ephys-pipeline
Support for recordings with subset of Neuropixels channels saved (e.g., 200 of 384)
- Dominant language
- Nextflow
- Stars
- 41
- Forks
- 10
- Avg merge
- 25m
- Merged PRs (30d)
- 2
Description
Hello Alessio and team,
First off, great work!
I'm using the aind-ephys-pipeline with Neuropixels 1.0 recordings from Open Ephys where I only saved 200 channels (0-199) to disk, not all 384. The pipeline fails at job_dispatch because the Open Ephys metadata specifies 384 channels but the binary only contains 200:
```
ValueError: The given Probe either has 'device_channel_indices' that does not match channel count
384 vs 200
device_channel_indices are: [0, 1, 2, ... 383]
recording channels are: ['AP0', 'AP1', ... 'AP199']
```
I tried providing a custom 200-channel probe via probe_paths in si_config.json, but the error occurs before the custom probe is applied - the Open Ephys reader attempts to set the 384-channel probe from metadata first.
Would it be possible to add an option to skip automatic probe loading when using openephysbinary reader, similar to how skip_stream_substrings works for filtering streams? Something like:
```
{
"reader_type": "openephysbinary",
"reader_kwargs": {
"folder_path": "/path/to/session",
"stream_name": "Record Node 101#OneBox-100.ProbeA-AP",
"load_probe_info": false
},
"probe_paths": "/path/to/custom_probe.json"
}
```
For now I'm using SpikeInterface directly with BinaryRecordingExtractor, but would prefer the full pipeline for QC outputs.
Thanks!
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.