Drawbacks

Due to its rather unconstrained nature, the DTD method is known to be noise-sensitive [@Reymbaut_accuracy_precision:2020]. Indeed, it merely estimates a set of diffusion components that best fits the measured signal. If the signal is noisy, the solution retrieved by the DTD method will also fit the acquisition noise. In particular, DTD tends to overestimate microscopic anisotropy in isotropic tissues. This problem is similar to the noise-induced overestimation of anisotropy observed in the context of diffusion tensor imaging [@Jones_Basser:2004].

Nonetheless, the noise sensitivity of the DTD method can be mitigated upon acquiring multiple b-tensor shapes and/or applying denoising/debiasing techniques to the acquired signals prior to data analysis.