Cramér–Rao bound-informed training of neural networks for quantitative MRI

Xiaoxia Zhang, Quentin Duchemin, Kangning Liu*, Cem Gultekin, Sebastian Flassbeck, Carlos Fernandez-Granda, Jakob Assländer

Research output: Contribution to journalArticlepeer-review

Abstract

Purpose: To improve the performance of neural networks for parameter estimation in quantitative MRI, in particular when the noise propagation varies throughout the space of biophysical parameters. Theory and Methods: A theoretically well-founded loss function is proposed that normalizes the squared error of each estimate with respective Cramér–Rao bound (CRB)—a theoretical lower bound for the variance of an unbiased estimator. This avoids a dominance of hard-to-estimate parameters and areas in parameter space, which are often of little interest. The normalization with corresponding CRB balances the large errors of fundamentally more noisy estimates and the small errors of fundamentally less noisy estimates, allowing the network to better learn to estimate the latter. Further, proposed loss function provides an absolute evaluation metric for performance: A network has an average loss of 1 if it is a maximally efficient unbiased estimator, which can be considered the ideal performance. The performance gain with proposed loss function is demonstrated at the example of an eight-parameter magnetization transfer model that is fitted to phantom and in vivo data. Results: Networks trained with proposed loss function perform close to optimal, that is, their loss converges to approximately 1, and their performance is superior to networks trained with the standard mean-squared error (MSE). The proposed loss function reduces the bias of the estimates compared to the MSE loss, and improves the match of the noise variance to the CRB. This performance gain translates to in vivo maps that align better with the literature. Conclusion: Normalizing the squared error with the CRB during the training of neural networks improves their performance in estimating biophysical parameters.

Original languageEnglish (US)
Pages (from-to)436-448
Number of pages13
JournalMagnetic resonance in medicine
Volume88
Issue number1
DOIs
StateAccepted/In press - 2022

Keywords

  • deep learning
  • magnetic resonance fingerprinting
  • magnetic transfer
  • objective function
  • parameter estimation
  • quantitative MRI
  • Magnetic Resonance Imaging
  • Neural Networks, Computer
  • Phantoms, Imaging

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Fingerprint

Dive into the research topics of 'Cramér–Rao bound-informed training of neural networks for quantitative MRI'. Together they form a unique fingerprint.

Cite this