The LONI QC System: A Semi-Automated, Web-Based and Freely-Available Environment for the Comprehensive Quality...

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Data Quality Control

Background:

• Ensuring neuroimaging data quality is crucial for valid and reproducible research.
• Standardized quality control (QC) procedures are essential for high-quality neuroimaging data.
• Existing QC methods may lack comprehensive metrics and automated assessment capabilities.

Purpose of the Study:

• To introduce the Laboratory of Neuro Imaging Quality Control (LONI-QC) system, a novel web-based platform for comprehensive neuroimaging data quality assessment.
• To provide a standardized workflow for evaluating multi-modal and multi-contrast brain imaging data.
• To develop and validate an automated QC procedure for structural MRI.

Main Methods:

• Development of a web-based system for anonymous upload and processing of brain imaging data.
• Computation of an exhaustive set of QC metrics, including scalar and vector statistics, using a compute cluster.
• Implementation of an automated QC procedure for structural MRI, classifying metrics as 'good' or 'bad'.
• Validation of QC metrics' reproducibility and the automated QC's sensitivity and specificity against visual inspection.

Main Results:

• The LONI-QC system successfully computes numerous QC metrics for diverse brain imaging data.
• The automated QC procedure for structural MRI demonstrates high sensitivity and specificity in identifying poor-quality images.
• Validation confirms the reproducibility of QC metrics across single-site and multi-site datasets.

Conclusions:

• LONI-QC is the first online system offering extensive QC metrics and automated assessment for multi-modal brain imaging.
• The system supports large-scale neuroimaging studies like TRACK-TBI and ADNI, promoting data quality standards.
• Free global access to LONI-QC is expected to enhance data quality and consistency across the neuroimaging community.