fMRI Preprocessing Pipeline Guide

fMRI preprocessing requires an adaptive decision tree to select steps like slice timing correction and normalization based on whether your analysis is task-based, resting-state, or MVPA. This ensures artifacts are avoided before statistical modeling by tailoring motion correction and smoothing to your specific study design.

fMRI quality control thresholds typically enforce framewise displacement below 0.2 mm for connectivity analysis, alongside DVARS and tSNR metrics. You should use visual inspection checklists and quantitative exclusion criteria tailored to your study type to justify preprocessing decisions and remove high-motion outliers.

Skipping spatial smoothing for MVPA is recommended to preserve fine-grained spatial patterns needed for decoding. The preprocessing guide provides domain-informed justifications for omitting smoothing in multivariate analyses, ensuring your pipeline maintains the localized voxel information critical for pattern classification.

Distortion and motion correction in an fMRIPrep pipeline are handled through step-by-step protocols that integrate BIDS-compliant inputs. The workflow guides coregistration and confound handling, utilizing fMRIPrep tooling to generate precise alignment and correct susceptibility distortions before handing off to GLM stages.

BIDS-compliant inputs are required to run fMRI preprocessing workflows using fMRIPrep. Ensuring your neuroimaging data is converted to the BIDS format before starting allows the pipeline to automatically parse acquisitions and apply the correct motion correction and normalization parameters.

Slice timing correction should be applied based on your TR and acquisition pattern, determined via an adaptive decision tree. This preprocessing step corrects temporal offsets from interleaved or sequential slice acquisition, which is critical for event-related task studies before accurate statistical modeling.