Markus Barth

This paper evaluates the accuracy and repeatability of quantitative susceptibility mapping using an anthropomorphic phantom, finding better performance at 3T than 7T.

This study evaluates four phantom doping materials for quantitative susceptibility mapping (QSM) by measuring molar susceptibility and test-retest repeatability at 3T and 7T.

The paper develops a multi-model consensus deep learning approach for automated gold fiducial marker segmentation in prostate MRI, achieving high sensitivity and precision.

This study uses ultra-high field fMRI to identify columnar-like functional patterns in human somatosensory cortex through vibration stimulation and cortical surface modelling.

This paper tests T1-weighted fMRI with iron oxide nanoparticles and UTE imaging at 9.4T to detect functional hyperaemia in mouse visual cortex, showing positive signal changes in both parenchyma and pial surface.

This study shows body size parameters significantly correlate with brain and spinal cord MRI metrics and should be included in neuroimaging studies.

VesselBoost is a Python tool for segmenting small blood vessels in human MRA data using deep learning and data augmentation to overcome labeling challenges.

This paper develops a deep learning model using QSM to differentiate gold fiducial markers from calcification in prostate cancer patients, achieving higher accuracy than conventional MRI methods.

The paper introduces HumanBrainAtlas, a high-resolution in vivo MRI dataset with detailed segmentations for research, clinical, and educational use.

The paper develops a model to predict depth-dependent VASO and BOLD responses in human primary visual cortex and finds that CBV changes mainly occur in small arterioles and capillaries.

QSMxT improves QSM processing by reducing artefacts and requiring minimal parameter tuning through a robust masking and two-pass reconstruction method.

LASHiS improves longitudinal hippocampal subfield segmentation using multi-contrast MRI, showing better reliability and performance than existing methods.

The paper uses Bayesian population receptive field modeling to estimate somatosensory pRFs in human primary somatosensory cortex and finds larger representations for the index finger and larger pRF sizes for the little finger.

This paper improves hippocampus subfield segmentation by using non-linear realignment to reduce movement artefacts and enhance image quality in MR scans.

DeepQSM uses deep learning to solve the dipole inversion in MRI for susceptibility mapping, improving the identification of brain substructures and their magnetic properties in diseases like multiple sclerosis.

SHARQnet improves background field removal in QSM by using a deep convolutional neural network, outperforming current methods without parameter tuning.

The paper analyzes how auditory cues during slow-wave sleep affect brain network connectivity and memory stabilization.

This paper develops a contrast matching algorithm to align MDA models of the human hippocampus with different contrasts, creating a high-resolution in vivo model for improved segmentation and disease detection.