BIO

Austin Pauley is a senior studying Biological Sciences. He is originally from Bonifay, FL, a small town in the Florida Panhandle. Austin is an active member of the FSU community where he participated in the Marching Chiefs for three years during his freshman, sophomore, and senior years. He has also participated in multiple service roles during his time at Florida State University. Austin is a volunteer emergency medical responder on campus and has served as both an educator and leader within the Medical Response Unit. He was also involved in several local service opportunities through the FSU Center for Leadership and Social Change in addition to being a volunteer educator for local EMT students. Austin serves as an EMT providing medical care and aid to several local communities. Within the Department of Neuroscience, Austin is involved in research involving olfactory perception, machine learning, and computer vision. His current research interests include olfactory perception, olfactory-guided behavior, behavioral modification, and machine learning.

DeepLabCut as a tool to investigate olfactory-guided behaviors in mice

Abstract

Investigating aspects of olfactory perception, such as odor valence, typically involves the presentation of one or multiple odors to a freely-moving animal and examining the resultant behavior. This process can be extremely time consuming, as the animal must be observed by a human and manually assessed. DeepLabCut (DLC) is a markerless pose-estimation software package that is free, open-source, and accurate at spatially tracking laboratory animals. DLC utilizes machine learning and convolutional neural networks to train a model to recognize a user-defined set of points on laboratory animals. The model then can be utilized to process novel video recordings and output a list of coordinates tracking the position of the animal over time. This positional data can then be introduced into other analysis pipelines to correlate specific behaviors with odor presentation. Our goal is to integrate this technology with our established methods to investigate olfactory perception in a high-throughput manner. Individual mice are placed in a large chamber with specialized odor and vacuum ports. Odor is delivered via a flow dilution olfactometer and the resultant behaviors are video recorded and processed utilizing DLC to output coordinates for positioning of specific body parts throughout the experiment. We aim to utilize the spatial coordinates DLC produces to determine how odor-guided responses are influenced by the genetic deletion of specific olfactory receptors, correlated with neural activity in olfactory brain regions, and altered by the inactivation of olfactory cortical regions.

Keywords: DeepLabCut Behavior Neuroscience