Project 460573

Alzheimer's disease (AD) is the most common cause of dementia. Lots of evidence suggests that AD stems from the abnormal aggregation of some proteins in the brain. This leads to a widespread neuronal circuit impairment beyond the endogenous compensation capacities and consequently cognitive decline and memory impairment. The Entorhinal Cortex (EC) - Hippocampus (HP) circuit, which plays a crucial role in memory, is selectively vulnerable in early AD and its impairment is associated with memory and cognitive impairment. While recent advances in imaging systems and signal recording techniques enable whole-brain mapping of cell distributions and neural activity in the EC-HP circuit, current computational pipelines are limited in their ability to quantitatively assess neuronal connectivity in the EC-HP network. Indeed, there is a lack of computational pipeline which is able to map these neuronal connectivity originating from EC to HP and analyze neurons' activity/firing patterns in this network. The overall goal of this project is to establish a computational platform based on state-of-art deep learning and signal analysis techniques for understanding the EC-HP network dysfunction in AD. We will employ a transgenic rat model which recapitulates many hallmarks of AD. Advanced fluorescence microscopy systems and electrophysiological signal recordings will be utilized to acquire needed data. We will develop state-of-the-art deep convolutional neural networks for mapping neuronal connectivity originating from EC to HP at a cellular level. Subsequently, to investigate the activity of neurons in the EC-HP circuit, different time-frequency techniques will be employed using electrophysiological recordings. Having completed the proposed experiments, we will have established secondary outcome measures in the EC-HP network, i.e., beyond behavioral readouts to better understand the network dysfunction in AD, that are keys for developing more effective treatments.