Nowadays, MRI is becoming widely used to image human body from fetus to adults for diagnostic purposes and also in research studies to understand how human brain works. Thanks to its non-invasive and non-ionizing radiation nature that we can also image newborns without giving any harm. Since MRI provides details of fine structures even a small motion of subject can cause blurry images. Imaging uncooperative subjects such as newborns it is difficult to circumvent motion without using any sedation. Mostly, structural scans like T2 weighted scans have to be repeated due to motion corrupted images which prolongs the total scan time, adds discomfort to babies and increases costs. There are many studies in the literature either tracks and corrects motion or circumventing it by acquiring shorter scans and applying reconstruction methodologies.

In our study, we implemented the super resolution reconstruction (SRR) method on 2D T2 weighted low resolution images of newborn subjects and Free Induction Decay (FID) navigated motion tracking and correction method in phantom and volunteer experiments. SRR is especially well adapted to thick slice 2D acquisition as it mitigates the risk of motion artifacts on the images. It reduces the acquisition time without any impact on the total duration of the exam. Therefore, mutually orthogonal in-plane high-resolution and through-plane low-resolution stacks were acquired for the reconstruction. SRR resulted in good quality high resolution (HR) volumes compared to standard HR acquisition. The results were validated by morphometry assessment using automatic segmentation as well by diagnostic assessment support the reconstruction results.

FIDnavs are an attractive approach for imaging babies as they do not require any external hardware. The method does not add any additional acquisition time and spatial characteristics on the image in terms of resolution and contrast. FIDnav is especially suited to 3D acquisitions as all the image information is kept, and motion corrections is a matter of k-space points displacement together with non-uniform Fourier transformation. In order to apply FIDnav in newborn imaging, it is required to avoid choreographic calibration and to restrict the imaging volume to the brain that moves rigidly. Thus, FIDnav approach has been improved by motion simulation and by adding slab selection excitation. Volunteer experiments showed that retrospective correction with FIDnav motion estimates improved the structural image quality in the presence of continuous head-shaking motion.