Title: Amygdala segmentation using deep learning
Legend: Image segmentation of the amygdala in a child with severe traumatic brain injury using deep learning. The segmentation of small brain structures like the amygdala is quite challenging in the presence of distorted and abnormal anatomy from major brain injuries. Our group is developing machine learning methods to segment brain structures like the amygdala in cases with traumatic brain injury. Machine learning based segmentations of the amygdala are shown in brown and yellow overlays. Images courtesy of Yilin Liu in the Alexander lab.
About the Lab: Alexander’s research focuses on the use of magnetic resonance imaging (MRI) for mapping and measuring the functional and structural organization of the human brain. These techniques are used to investigate the brain in both typically developing individuals and subjects with developmental disorders including autism. Functional MRI (fMRI) is used to assess brain regions associated with cognition and affect and their dysfunctions in these populations. Diffusion tensor MRI (DT-MRI) is used to study the patterns of structural connectivity between brain activity regions. Anatomic imaging methods are used to assess longitudinal structural changes in brain regions. These measurements are ultimately compared with measures of affect, behavior and cognition in specific population groups.