FijiRelax: Fast and noise-corrected estimation of MRI relaxation maps in 3D + t

Source Agritrop Cirad (https://agritrop.cirad.fr/604609/) * Autres projets (id;sigle;titre): ;VITIMAGE-2024;(FRA) Utilisation d'imagerie non destructive pour un suivi dynamique de la colonisation du bois par les pathogènes responsables des maladies du bois// ;APLIM;(FRA) Advanced Plant Life Imaging and Metrology//

International audience

Inglés. Quantitative Magnetic Resonance Imaging (MRI) is a relevant technique to measure water availability and binding in animal or vegetal tissues. Among others, MRI is useful to decipher the impact of stresses on living samples (Sidi-Boulenouar et al., 2019). Reproducible measurements of anatomical and physiological changes can be obtained by quantitative mapping of biomarkers (see Figure 1-a) such as proton density (PD) and magnetization relaxation times (T1, T2) in 2D images (Borisjuk et al., 2012 ; Dusschoten et al., 2016). However, the application of quantitative MRI to 3D + t microscopic imaging is still challenging. Detection of weak signals associated with small voxels requires longer acquisition times, which favors spatial drift and tissue deformations (Han et al., 2015). In addition, low signal-to-noise ratio (SNR) environments dramatically reduce the estimation accuracy of the parameters (Raya et al., 2010). Non-specialists facing these issues with multi-echo spin-echo data may be discouraged when using available open implementations of T1-T2 relaxometry: there is a lack of a user-friendly tools capable of processing large data in a reasonable time. We propose FijiRelax, an efficient Fiji plugin for the estimation of relaxation parameter maps from 3D + t multi-echo spin-echo sequences, integrating Rician noise and spatial drift corrections. The plugin is distributed and documented as an ImageJ plugin (Plugin page), available in a graphical interface and scriptable in BeanShell to allow batch processing for high-throughput experiments.