Towards the Development of an Automated Pipeline for Cilia Detection

Alzheimer's Disease (AD) is a neurodegenerative disease that is characterized by a decline in memory and cognition, and there is an urgent need to understand the mechanisms of the disease. Primary cilia have been implicated in AD research because of their potential relationship with signaling abnormalities in patients with AD. However, identifying cilia in stained postmortem human brain tissue samples through immunofluorescence microscopy is difficult and can be subjective and prone to observer bias. We aimed to automate the process to reduce the analysis time and the potential bias using a novel preprocessing pipeline and a convolutional neural network (CNN). Our research found that our preprocessing pipeline reduced noise and normalized the data. However, the CNN did not show high accuracy in most instances. A larger dataset is required to thoroughly evaluate its ability to find the cilia in the stained brain tissue samples.