RT Journal Article
SR Electronic
T1 Magnetic Resonance Imaging-based Deep Learning Predictive Models of Brain Disorders: A Systematic Review of Modelling Practices, Transparency, and Interpretability
JF medRxiv
FD Cold Spring Harbor Laboratory Press
SP 2021.11.20.21266620
DO 10.1101/2021.11.20.21266620
A1 Shane O’Connell
A1 Dara M Cannon
A1 Pilib Ó Broin
YR 2021
UL http://medrxiv.org/content/early/2021/11/21/2021.11.20.21266620.abstract
AB Brain disorders are characterised by impaired cognition, mood alteration, psychosis, depressive episodes, and neurodegeneration, and comprise several psychiatric and neurological disorders. Clinical diagnoses primarily rely on a combination of life history information and questionnaires, with a distinct lack of discriminative biomarkers in use for psychiatric disorders. Given that symptoms across brain conditions are associated with functional alterations of cognitive and emotional processes, which can correlate with anatomical variation, structural magnetic resonance imaging (MRI) data of the brain are an important focus of research studies, particularly for predictive modelling. With the advent of large MRI data consortiums (such as the Alzheimer’s Disease Neuroimaging Initiative) facilitating a greater number of MRI-based classification studies, convolutional neural networks (CNNs), which are multi-layer representation-based models particularly well suited to image processing, have become increasingly popular for research into brain conditions. Despite this, modelling practices, the degree of transparency, and considerations of interpretability vary widely across studies, making them difficult to both compare and/or reproduce. Modelling practices here refers to issues surrounding the data splitting procedure, the presence or absence of repeat experiments, the critical appraisal of performance metrics, and the overall reliability of the modelling approach. Transparency refers to how detailed the authors’ methodological descriptions are, and the availability of code. Finally, interpretability refers to the attempt made by the authors to identify structural brain alterations driving model predictions – this is particularly important as the application of deep learning systems becomes more widespread in clinical settings. Here, we conduct a systematic literature review of 55 studies carrying out CNN-based predictive modelling of brain disorders using MRI data and critique their modelling practices, transparency, and considerations of interpretability; we furthermore propose several practical recommendations aimed at promoting comprehensive, clear, and reproducible research into brain disorders using MRI-based deep learning models.Competing Interest StatementThe authors have declared no competing interest.Funding StatementThis work was conducted with the financial support of Science Foundation Ireland under Grant number 18/CRT/6214.Author DeclarationsI confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.YesThe details of the IRB/oversight body that provided approval or exemption for the research described are given below:All papers analysed in this review are available on either PubMed or Web of Science.I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.YesI understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).YesI have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable.YesAll papers analysed in this review are available on either PubMed or Web of Science.