Fast and Reliable Myelin Water Fraction Estimation Using Neural Network Informed Non-Linear Least Squares Fitting

Motivation: Non-Linear Least Squares (NLLS) is commonly used for fitting data in Myelin Water Fraction (MWF) imaging but suffers from slow fitting times and sensitivity to experimental factors. Goal(s): To accelerate and improve the performance of the NLLS fitting method for MWF imaging. Approach: To set the optimum initial value and bounds for NLLS fitting, we employ Artificial Neural Networks (ANNs) with dropout trained on NLLS results to estimate value and uncertainty of estimation for each parameters. Results: In the simulations, the error decreased, and the fitting time was reduced by 51.82% for simulations and 57.10% for in-vivo experiments. Impact: By combining ANN and NLLS, the proposed method enhances speed and accuracy in MWF estimation, enabling clinicians to reliably assess more patients with neuroinflammatory diseases in the same amount of time, potentially improving diagnosis and treatment outcomes.