Fig. 5 | Scientific Reports

Fig. 5

From: Flexible and cost-effective deep learning for accelerated multi-parametric relaxometry using phase-cycled bSSFP

Fig. 5The alternative text for this image may have been generated using AI.

Influence of off-resonance-related phase accumulation within TR (\(\theta\)) on DNN and MIRACLE relaxometry in silico depending on the number of bSSFP phase cycles. DNNs were trained on noise-free data from a uniform distribution of \(T_1\) and \(T_2\) and for a varying number of phase cycles (\(N_{pc} \in \left\{ 12, 6, 4 \right\}\)). The noise-free test data were simulated based on reference \(T_1\) (62 ms) and \(T_2\) (939 ms) white matter relaxation values at 3 T55. The relative error between the parameter predictions \(\hat{T}_i\) and the simulated ground truth value \(T_i\), \(\epsilon _{rel}(T_i) = (\hat{T}_i-T_i)/T_i\cdot 100\) with i = 1, 2, is shown for MIRACLE and the standard magnitude-based DNNs (left column) as well as the complex-based DNNs (right column). Dashed lines indicate the 0 \(\%\) error for reference.

Back to article page