NORHA: A NORmal Hippocampal Asymmetry Deviation Index Based on One-Class Novelty Detection and 3D Shape Features.

Radiologists routinely analyze hippocampal asymmetries in magnetic resonance (MR) images as a biomarker for neurodegenerative conditions like epilepsy and Alzheimer's Disease. However, current clinical tools rely on either subjective evaluations, basic volume measurements, or disease-specific models that fail to capture more complex differences in normal shape. In this paper, we overcome these limitations by introducing NORHA, a novel NORmal Hippocampal Asymmetry deviation index that uses machine learning novelty detection to objectively quantify it from MR scans. NORHA is based on a One-Class Support Vector Machine model learned from a set of morphological features extracted from automatically segmented hippocampi of healthy subjects. Hence, in test time, the model automatically measures how far a new unseen sample falls with respect to the feature space of normal individuals. This avoids biases produced by standard classification models, which require being trained using diseased cases and therefore learning to characterize changes produced only by the ones. We evaluated our new index in multiple clinical use cases using public and private MRI datasets comprising control individuals and subjects with different levels of dementia or epilepsy. The index reported high values for subjects with unilateral atrophies and remained low for controls or individuals with mild or severe symmetric bilateral changes. It also showed high AUC values for discriminating individuals with hippocampal sclerosis, further emphasizing its ability to characterize unilateral abnormalities. Finally, a positive correlation between NORHA and the functional cognitive test CDR-SB was observed, highlighting its promising application as a biomarker for dementia.

VBM sensitivity to localization and extent of mouse brain lesions: A simulation approach.

BACKGROUND AND OBJECTIVES: Voxel-based morphometry (VBM) is a popular neuroimaging technique, used to detect and quantify morphological differences in brain tissues between groups. Widely used in human studies, VBM approaches have tremendous potential for neuroimaging studies in animal models. A significant challenge for applying VBM to small animal studies is the poor understanding of how the design of preprocessing pipelines impacts quantitative results. This is important because the large differences in size, resolution, and imaging parameters implies that human imaging preprocessing pipelines cannot be uncritically applied to small animal studies. In this work, we assessed and validated the performance of different VBM pipelines for the study of the mouse brain. METHODS: We applied two pipelines -namely DARTEL VBM and Optimized VBM- by varying spatial normalization used during preprocessing. Using an automatic method, we simulated varying levels of volumetric gray matter (GM) loss and sizes of tissue atrophy on specific areas of the mouse brain. We evaluated the performance of each pipeline by comparing location and extent of the differences detected by them with the simulated ones. Finally, we applied both pipelines on magnetic resonance (MR) images of the brain derived from an experimental model of growth restriction on mice. RESULTS: Our results demonstrated that some subtle atrophies were detected by the Optimized workflow but not by the DARTEL VBM workflow. Detection of less subtle atrophies was similar for the two workflows, but DARTEL VBM performed better at estimating their size and anatomical location. Both VBM pipelines had difficulties at finding atrophies with a very small level of volumetric loss and, in general, they underestimated the magnitudes of difference between groups. These results also varied across brain regions, with better performance on brain cortex than other regions such as the cerebellum. CONCLUSIONS: The analysis and quantification of VBM pipelines on different areas of the mouse brain allows a better understanding of the advantages and limitations of their results. We performed a controlled and quantitative analysis of the method providing robust evidence to interpret changes in real contexts.