Visual Evaluative Control Technology Of Resonance Spectroscopy (VECTORS): Automated Data Quality Control Pipeline for Spectroscopy Data

Motivation: The absence of a robust data quality control (DQC) process inhibits the widespread adoption of MRS. Standard quantitative metrics (noise, linewidth, and metabolite fitting) don't account for all DQ criteria, while manual DQC is subjective, irreproducible, labor-intensive, and requires expertise. Goal(s): Develop an automated DQC pipeline that accounts for all DQC criteria while minimizing manual labor, subjectivity, variability, and required expertise. Approach: The pipeline quantifies and replicates an expert's qualitative analysis. Results: The pipeline accurately flags additional spectra containing movement, phase, lipid, out-of-volume lipid, suppression, ghosting, and poor fit artifacts Impact: As a comprehensive, automated, and efficient MRS DQC process, the pipeline can help standardize DQC and expand the field to scientists with little MRS expertise. The pipeline should be used in addition to standard quantitative metrics.