Kerstin Ritter

This study uses machine learning to predict alcohol consumption from resting-state fMRI connectivity, finding weak but reproducible signals with limited predictive power.

The paper examines associations between impaired control symptoms in AUD and behavioral, neural, and drinking behavior measures, finding specific links to inhibitory control deficits.

This study shows that AI models using neuroimaging data do not outperform simple models using demographics and recording site in predicting psychiatric phenotypes in youth.

The paper introduces DeepTaskGen, a deep learning method that generates synthetic task-based contrast maps from resting-state fMRI data, outperforming benchmarks and enabling biomarker development for cognitive tasks.

This study shows simpler brain-age models with lower age prediction accuracy are more sensitive to disease-related changes, suggesting they may be better biomarkers for detecting neurological and psychiatric disorders.

This study compares transformers and CNNs in predicting brain age using MRI data, finding they perform similarly and represent different concepts of brain aging.

This study shows that prediction performance for brain-imaging-based phenotypes can improve significantly with larger sample sizes and multimodal data.

The paper investigates neural pathways linking psychological stress and brain health in healthy individuals and multiple sclerosis patients, finding similar functional connectivity patterns.