Automatic quantification of cell viability in hippocampal organotypic cultures

Traumatic Brain Injury (TBI) is a leading cause of disability with many different types of forces. Because the brain is made of up interconnected nerve cells, injury from one area of brain can easily spread throughout the rest of the brain. In order to study this effect in this paper, we use hippocampal organotypic cell cultures. To analyze how secondary and further injury is spread in the brain due to TBI, specific regions of interest are outlined using bright-field microscopy images and analyzed for cell death through fluorescent images. Because of the bias in segmenting these specific regions, and the darkening of bright-field images from necrosis, there is a need for the development of an automatic cell viability analysis method. We first create silhouette masks identifying the area of the hippocampus within the bright-field images. Then, we align these masks of these images between two time points – one before experimentation and another 24 hours after lab experimentation - using Mattes Mutual Information registration to help analyze corresponding regions between time points. We also use relative intensity gradients and contrast enhancement to help isolate the specific regions of interest within the hippocampus, and then aligned the regions of interest between time points using the transformation discovered through the registration process. Finally, we look within these extracted regions for cell viability in fluorescent images taken after experimentation and calculate the ratio between the area of the fluorescent region extracted and the area of the segmented region. Comparing the results from this method with manual segmentation by experienced users, we discover an average of 12% error in segmentation and 2% error in registration. The overall method took an average of 7 minutes per image, which compared to the time it takes complete manually, allows users to conserve time and prevent user-to-user bias in segmentation.M.S.Includes bibliographical referencesby Aparna Kanna