Advancement: Quantifying Innervation using Deep Learning Algorithms for Application in Wound Healing

Sam Teymoori
Electrical Engineering Ph.D. Student
Virtual Event
Marcella Gomez

Join us on Zoom: / Passcode: 779299

Description: Wound healing is the complex and dynamic process of replacing damaged and destroyed cellular structures and tissue layers. It requires the cooperative activity of numerous cell types and information guiding cells to grow and implement tissue morphogenesis to a specific (fully regenerative) outcome. We aim to speed up the healing process by manipulating two biological processes: improving nerve growth and activating macrophages toward reparative functions. Here, we focus on developing methods to inform intervention strategies to accelerate wound healing by manipulating innervation. This is done through analysis of in vitro and ex-vivo studies. To determine how nerve growth can be manipulated, we analyze in vitro studies where neurons are treated with different drugs and dosages. However, the process of quantifying nerve growth from microscopy data is manual, which makes this process slow and labor-intensive.

In this work, I have developed a deep learning model to detect nerve growth. By using this method, the nerve growth can be quantified in a more efficient, fast, and fully-automated way with repeatable results. To analyze the effects of various drugs and dosages on nerve growth, we have applied our deep learning model on the datasets produced from in-vitro experiments. The results provide the drugs and corresponding dosages that can improve nerve growth.

In the future, we plan to use our deep learning model to determine nerve growth in real-time and combine this with a feedback control algorithm informed by the in vitro studies for real-time feedback control of innervation in vitro. In order to identify the best timing for intervention in vivo, we analyze ex-vivo data. By analyzing the results of the ex-vivo experiment, we can determine the spatio-temporal dynamics of nerve growth. Using this information, we can adjust the treatment dosage, timing, and location to speed up the wound healing process in vivo.