Different preprocessing strategies lead to different conclusions: A [¹¹C]DASB-PET reproducibility study.

Area of Science:

• Neuroimaging
• Pharmacology
• Psychiatry

Background:

• Positron emission tomography (PET) enables in vivo study of biological processes.
• PET data quantification relies on preprocessing pipelines, but their impact on outcomes is often unclear.
• The choice of data analysis methods can influence research findings.

Purpose of the Study:

• To evaluate the impact of 384 preprocessing strategies on [11C]DASB-PET data.
• To assess how preprocessing affects the association between serotonin transporter binding and depressive symptoms after hormone manipulation.
• To identify critical preprocessing steps influencing study outcomes.

Main Methods:

• Utilized a dataset from a double-blind, randomized, placebo-controlled [11C]DASB-PET study in 30 women.
• Applied 384 distinct preprocessing pipelines to dynamic PET data.
• Analyzed the influence of motion correction and kinetic modeling on reported findings.

Main Results:

• Only 36% of all tested preprocessing strategies replicated the original positive finding (p < 0.05).
• Preprocessing strategies incorporating motion correction achieved a 72% replication rate.
• Strategies lacking motion correction failed to replicate the finding (0% replication).

Conclusions:

• The selection of preprocessing strategies critically influences the outcomes of [11C]DASB-PET studies.
• Motion correction is a vital component for reproducible results in this neuroimaging modality.
• Standardizing preprocessing pipelines is essential for reliable interpretation of PET neuroimaging data.