Logs and Metadata
Thelogs/
and metadata/
folders are companion outputs that provide run-level transparency and support reproducibility.
Logs
- Plain-text log file of the pipeline execution.
- Contains system messages, warnings, and step-by-step notes from the run.
- Useful for debugging (e.g., if a step fails or behaves unexpectedly).
- Each run generates a new log with a timestamped filename.
Metadata
-
*_autoclean_metadata.json
Machine-readable JSON summary. Captures:- Run ID, timestamps, and task name.
- User/system context (username, host, PID).
- Directories used (
bids/
,derivatives/
,final_files/
). - Entry point and preprocessing parameters.
- Status (
completed
,failed
, etc.).
-
*_autoclean_report.pdf
Human-readable QC report. Includes:- Import details (channels, duration, sample rate).
- Processing details (filters, reference, resampling).
- Full parameter table for reproducibility.
-
ica_control_sheet.csv
(Under Construction) A new feature in development.- Will track ICA components across subjects.
- Intended for manual review and component curation.
- Planned integration: the ability to reapply ICA weights and resume downstream steps.
Example Content
Pipeline Execution Log
The log file captures the complete pipeline execution trace: View Complete Pipeline Log Key pipeline steps documented include:Processing Log Summary
The CSV processing log provides a comprehensive record of each run: View Complete Processing Log Key metrics captured include:- Channel Information: Original (128) vs. post-processing channel counts
- Bad Channels:
['E56', 'E128', 'E8', 'E14', 'E17', 'E2']
- Filtering: 1-80 Hz bandpass with 60/120 Hz notch filters
- Epoching: 29 trials total, 7 rejected (75.8% retained)
- ICA Components: 16 components removed from 121 total
Why These Matter
- Logs capture the execution trace.
- Metadata provides both a human-readable report and a machine-readable JSON.
- Together, they ensure that every run is auditable, reproducible, and shareable.
- The ICA control sheet will expand this functionality by enabling semi-automatic component review pipelines.