Publications by authors named "Thomas Bielser"
Article Synopsis
- Translational magnetic resonance imaging of rodent brains is crucial for drug development, but automated image segmentation is challenging due to lower image quality compared to human MRI.
- The study explored the effectiveness of deep learning models for segmenting rat brain images into multiple regions, using a robust method that accommodates variability in animal strain and size.
- Among the models tested, the Attention U-Net performed best, providing accurate segmentation while minimizing the reliance on costly manual labels, and included features for uncertainty estimation and explainability for quality assurance.
Unconditioned fear plays an important yet poorly understood role in anxiety disorders, and only few neuroimaging studies have focused on evaluating the underlying neuronal mechanisms. In rodents the predator odor trimethylthiazoline (TMT), a synthetic component of fox feces, is commonly used to induce states of unconditioned fear. In this study, arterial spin labeling-based functional magnetic resonance imaging (fMRI) was applied to detect TMT-induced regional modulations of neuronal activity in Wistar rats.
View Article and Find Full Text PDFRationale And Objectives: In schizophrenia research, most of the functional imaging studies have been performed in psychotic patients, but little is known about brain areas involved in the expression of psychotic-like symptoms in animal models. The objective of this study was to visualize and compare brain activity abnormalities in a neurodevelopmental and a pharmacological animal model of schizophrenia.
Methods: Blood perfusion of specific brain areas, taken as indirect measure of brain activity, was investigated in adult rats following either neonatal ventral hippocampal lesion or acute administration of phencyclidine.