Fit ICA per subject and reject artifact components using ICLabel confidence scoring.
npx skills add https://github.com/dengzhe-hou/auto-eeg-analysis --skill eeg-ica
Please help me install this Agent Skill. Skill: eeg-ica Source: https://github.com/dengzhe-hou/auto-eeg-analysis/tree/main/skills/eeg-ica Command: npx skills add https://github.com/dengzhe-hou/auto-eeg-analysis --skill eeg-ica
EEG analysis pipelines often suffer from eye, muscle, heartbeat, and line-noise artifacts that are slow and error-prone to remove by manual inspection.
Run ICA artifact removal for your study by executing: /eeg-ica projects/my-study --method infomax --n_components 0.99 --threshold 0.7
High-intent search queries and answers about installing and using this skill.
Automated EEG artifact removal uses ICLabel confidence scoring to predict component types like eye blink and heartbeat, excluding those exceeding a threshold. This approach fits ICA per subject and applies weights to original bandpass data to generate cleaned outputs.
Handling ambiguous ICA components involves flagging items with high ambiguity or low confidence for human confirmation. This safeguard pauses the artifact removal pipeline, ensuring components are reviewed rather than blindly rejected during EEG preprocessing.
Fitting ICA on a 1 Hz high-pass copy improves decomposition quality by stabilizing the algorithm. The resulting ICA weights are then transferred and applied to the original analysis-bandpass EEG data to preserve the desired frequency range.
Yes, an unfrozen ANALYSIS_PLAN.md must be present to run the ICA artifact removal pipeline. The pipeline requires this file to define parameters and uses preprocess-stage outputs to fit ICA and generate per-subject labels.
Yes, MNE-Python supports repeatable EEG artifact rejection when integrated with ICLabel confidence scoring. This combination performs automated ICA decomposition, removes artifacts, and generates per-subject component documentation for paper-aligned preprocessing.
asapMakiAggregate daily summaries into a structured weekly markdown report.
ovachieverStream interactive React UI components from LLM responses.
StarlitnightlyPreprocess TCGA bulk data into annotated AnnData objects with expression and clinical metadata.
NeuroAIHubGuide EEG preprocessing pipelines for filtering, artifact rejection, and referencing.
