Advanced MRI Techniques and Applications

This study shows that cerebrovascular reactivity increases with age across multiple brain networks using longitudinal fMRI data from children.

This study demonstrates the feasibility of BOLD fMRI in detecting neuronal activation in the lumbosacral cord during lower extremity motor tasks and shows TE-dependent effects on BOLD signal.

This study evaluates two versions of NORDIC denoising on submillimeter auditory fMRI data, finding improved detection sensitivity and reliability but reduced average response amplitude.

The paper proposes methods to approximate bone mineral density and subcutaneous adiposity from T1-weighted head MRI images and validates their reliability and correlation with DXA and MRI measurements.

The paper introduces BrainQCNet, a deep learning model for automated artifact detection in brain structural MRI scans, achieving high accuracy and fast processing times.

This paper investigates noise in line-scan MRI and shows that temporal fluctuations can mimic neural responses, highlighting limitations in detecting the DIANA response.

The paper introduces FMRI quality control tools in AFNI, providing systematic, in-depth, and interactive QC features for preprocessing pipelines.

This paper characterizes cortical depth-dependent CBV and BOLD hemodynamic responses across stimulus durations using VASO and ME-GRE at 7T in humans, finding that longer durations improve CBV specificity but may reduce efficiency.

This study evaluates how acquisition and modeling parameters affect test-retest reproducibility of GABA+ measurements using edited MRS, finding that model selection significantly impacts reproducibility.

This study evaluates ME-ICA's effectiveness in reducing task-correlated head motion artifacts in motor-task fMRI data, showing improved signal clarity and stability in both subject and group analyses.

This study assesses the intra-individual reliability of 1H-MRS measurements in the anterior cingulate cortex over one year, finding that metabolite levels are reliably quantified with small variability.

The paper develops a resource to evaluate MRI harmonisation approaches by comparing variability across scanners, within-scanners, and biological differences using multimodal data from multiple sites.

The paper examines age-related changes in basal ganglia iron levels in neonates using rsfMRI data, finding postnatal age associations in specific subregions and preterm birth effects.

This study uses interleaved fMRS and fMRI to examine how visual contrast affects neurometabolite levels and BOLD responses in the human brain, finding non-linear relationships and consistent glutamate increases.

This paper demonstrates a method to concurrently measure BOLD and metabolite changes during a cognitive task using fMRS, showing significant Glx increases and moderate BOLD correlation.

This study attempts to noninvasively measure phenylalanine in the human brain using 7T MRS but fails to detect it in vivo, concluding that natural concentrations are below the detection threshold.

This paper introduces AGNOSTIC, a synthetic 1H MRS dataset for training deep learning models to detect and predict out-of-voxel echoes in brain MRS data.

This study characterizes the developmental trajectories of brain metabolites measured by 1H-MRS from childhood to early adulthood, identifying non-linear changes in several metabolites and their associations with cognitive development.

The paper compares preprocessing methods for task fMRI data, finding that OGRE reduces inter-individual variability and increases activation detection in primary motor cortex compared to FSL and fMRIPrep.

This study investigates how echo train length (ETL) affects microvascular specificity in SE-EPI BOLD fMRI, showing increased macrovascular contamination with longer ETL durations.

The paper develops acquisition protocols and processing pipelines for whole-brain layer-fMRI and introduces the Kenshu dataset for benchmarking and analysis.

SACNet corrects susceptibility artifacts in dMRI scans using a registration network with neuroanatomical constraints, showing improved performance across diverse datasets.

This paper characterizes cardiac-induced noise in brain R2* maps and proposes a sampling strategy to mitigate its effects.

This paper develops a deep learning method with uncertainty estimation to detect MRI artifacts in large imbalanced datasets, achieving high accuracy and outperforming previous approaches.

This study compares denoising methods for multi-echo 7T fMRI data, finding that ME-ICA combined with aCompCor and ICA-AROMA is highly effective for noise reduction and signal preservation.

The paper uses two non-contrast MRI methods to investigate the glymphatic system's response to sleep and aging in humans.

The paper introduces an open-access dataset combining EEG and fMRI data from naturalistic viewing tasks to study brain function and preprocessing methods.

This paper evaluates brain age predictions across repeated time points and finds small within-subject correlations, field strength influence, and model biases.

This study examines how brain glutathione levels change with age and their link to cognitive function in healthy adults, finding age-related declines in specific brain regions and sex differences in GSH dynamics.

The paper shows resting-state fMRI signals contain spectral signatures reflecting local hemodynamic response timing variations across brain regions.