Using AI to Detect Loss of Connections Between Neural Cells in Glaucoma

We will develop AI models able to recognize billions of synapses in the inner retina of animal models of glaucoma.

Death of retinal ganglion cell in glaucoma is not equal across the retina, with neurons in some regions able to survive longer than others. The reason behind this is still largely unknown. We propose to create new AI tools to investigate large portions of the retina and use them to assess structure and function of synapses early after ocular pressure elevation in animal models of glaucoma

This project leverages AI to assess a large number of synapses in the inner retina, and uses this tool to detect large patterns of rearrangement across different animal models of glaucoma. Such approach allows to overcome longstanding biases caused by sampling small number of retinal ganglion cells and their synapses, and highlights shared alterations of retinal circuits associated with glaucoma across multiple animal species.

This study will provide a novel, free and open-source tool for other laboratories studying connectivity of neural cells in the retina during glaucoma, and will expand our knowledge of what synapse rearrangements are most commonly associated with loss of function in retinal ganglion cells during glaucoma.