What problem does it solve? Raw fMRI time series contain low-frequency drift and linear trends that distort connectivity, decoding, and statistical analyses. This Skill provides model-level guidance for applying classical detrending to clean BOLD signals before downstream neuroimaging workflows. ## Core Features & Use Cases - Drift and Trend Removal: Removes constant and linear temporal trends from voxel-wise or ROI-wise fMRI time series. - Confound-Aware Cleaning: Optionally combines detrending with confound regression and standardization using a confounds TSV and brain mask. - Delegation-Based Workflow: Coordinates with fmri-skill for preprocessing planning and nilearn-tool for concrete implementation via nilearn's clean_img. - Use Case: A researcher has preprocessed resting-state BOLD images and wants to remove scanner drift before computing functional connectivity matrices, producing cleaned BOLD images and QC summaries. ## Quick Start Ask the assistant to detrend a preprocessed BOLD image with a given TR and optional confounds file, delegating execution to the nilearn denoising script after confirming the preprocessing plan.
