Consensus Recommendations to Establish Reporting Standards in fMRI of Migraine: A Delphi Study.

BACKGROUND AND

OBJECTIVES

Migraine is a multifaceted primary headache disorder. In neuroimaging of migraine, fMRI has been used to elucidate pathophysiology or monitor treatment effects. The current literature, however, is highly heterogeneous regarding reported variables and methodologies. This begets a lack of comparability and complicates synthesis of results across studies. We developed a framework for standardized reporting of fMRI studies in migraine.

METHODS

Experts on fMRI in migraine were identified from the literature and subjected to structured questionnaires in 2 iterations of 3 rounds according to the DELPHI method. A total of 157 statements across 17 reporting domains were rated on 5-point Likert scales (strong support to strong opposition). The first iteration covered demographic data, migraine-specific factors, medication, scan timing, healthy controls (HCs), participant sampling/recruiting, standardized forms, study preregistration, region of interest (ROI) analyses, validation data sets, data sharing, preprocessing documentation, and analysis software. The second iteration of the questionnaire covered scanner-related factors, sequence-related factors, physiology monitoring, and stimulation-related factors. Items showing strong consensus/consensus (≥90%/≥75% of participants indicating scores 4 or 5) were included as standard reporting items.

RESULTS

All 3 rounds of the first/second iteration were completed by 29 and 26 researchers (age 46 ± 11 years; 38% female/age 46 ± 12 years; 44% female) from 23 and 21 institutions. Across both iterations, strong consensus and consensus was achieved for 34 (3 scanner-related factors, 9 sequence-related factors, 1 stimulation-related factor, 2 demographic factors, 7 migraine-specific factors, 2 medication-factors, 2 scan timing factors, 4 HC factors, 1 preregistration factor, 1 analysis software factor, and 2 ROI analyses factors) and 33 (1 scanner-related factors, 4 sequence related factors, 1 factor related to physiology monitoring, 1 stimulation-related factor, 3 demographic factors, 6 migraine-specific factors, 4 medication factors, 3 HC factors, 2 sampling factors, 1 standardized form, 1 preregistration factor, 1 data sharing factor, 2 analysis software factors, and 3 ROI analyses factors) items, respectively. From these, a checklist covering 63 items from 14 reporting domains was created.

DISCUSSION: We present an expert-based framework for reporting standards in fMRI studies of migraine, which can be used for future studies to homogenize cohort characterization, fMRI acquisitions, and analysis protocols.