Studying the neurobiology of memory with computer vision & machine learning
Zasha Benites
Undergraduate Researcher
Computer Science Major (Luddy School of Informatics, Computing and Engineering)
Sophia Sullivan
Undergraduate Researcher
Psychology Major (College of Arts and Sciences)
Ehren Newman
Faculty Mentor
Ehren Newman (College of Arts & Sciences)
Project Description
My lab studies the neurobiology of learning and memory. To understand the connection between brain activity and behavior, we track what our experimental participants--rats--are doing over time. The standard way to do this is by recording them with cameras and then watching the recordings and scoring them. Not only is this extremely laborious, it is also subject to subjective scoring differences between those watching the videos. In this project, you will help us implement a new tool that takes advantage of machine vision and learning techniques to automatically score the behavior. You will be trained in state-of-the-art techniques while contributing to research into the links between brain and behavior. As a member of my lab you will also learn about the neurobiology of memory, navigation, and the changes that accompany Alzheimer's pathology. If you are successful in this program, you will have the chance to participate more broadly in our work as it suits your interests and availability.
Technology or Computational Component
Building an automatic behavioral scoring system will involve using state-of-the-art machine vision and learning systems known as convolutional deep neural networks. We have already a network that is well suited to the task of scoring the behavior but what you will do is help us train it to label the specific behaviors we are interested in, assess the network's ability to label data it wasn't trained on, and make use of the output that it generates to make inferences about animal behavior.