Estimating uncertainty in read‐out patterns: Application to controls‐based denoising and voxel‐based morphometry patterns in neurodegenerative and neuropsychiatric diseases. Issue 7 (22nd March 2023)

• Record Type: Journal Article
• Authors: Blum, Dominik; Hepp, Tobias; Belov, Valdimir; Goya‐Maldonado, Roberto; la Fougère, Christian; Reimold, Matthias
• Is Part Of: Human brain mapping. Volume 44:Issue 7(2023)
• Page Start: 2802
• Page End: 2814
• Publication Date: 2023-03-22
• Subjects: controls‐based denoising -- pattern quantification -- pattern uncertainty; Brain mapping -- Periodicals; 611.81
• DOI: 10.1002/hbm.26246
• Languages: English
• ISSNs: 1065-9471

Abstract

Abstract: Quantifying pathology‐related patterns in patient data with pattern expression score (PES) is a standard approach in medical image analysis. In order to estimate the PES error, we here propose to express the uncertainty contained in read‐out patterns in terms of the expected squared Euclidean distance between the read‐out pattern and the unknown "true" pattern (squared standard error of the read‐out pattern, SE 2 ). Using SE 2, we predicted and optimized the net benefit (NBe) of the recently suggested method controls‐based denoising (CODE) by weighting patterns of nonpathological variance (NPV). Multi‐center MRI (1192 patients with various neurodegenerative and neuropsychiatric diseases, 1832 healthy controls) were analysed with voxel‐based morphometry. For each pathology, accounting for SE 2, NBe correctly predicted classification improvement and allowed to optimize NPV pattern weights. Using these weights, CODE improved classification performances in all but one analyses, for example, for prediction of conversion to Alzheimer's disease (AUC 0.81 vs. 0.75, p = .01), diagnosis of autism (AUC 0.66 vs. 0.60, p < .001), and of major depressive disorder (AUC 0.62 vs. 0.50, p = .03). We conclude that the degree of uncertainty in a read‐out pattern should generally be reported in PES‐based analyses and suggest using weighted CODE as a complement to PES‐based analyses. Abstract : We describe the level of uncertainty in read‐out patterns from brain image data using the concept of a squared standard error and estimate the error in pattern expression scores. By optimizing weights of nonpathological variance patterns, we predict classification improvement by our method in the most common neurodegenerative and neuropsychiatric disorders.