Abstract
In the past few years, advances in machine learning have fueled an explosive growth of descriptive and generative behavior models of animal behavior. These new approaches offer higher levels of detail and granularity than has previously been possible, allowing for fine-grained segmentation of animals' actions and precise quantitative mappings between an animal's sensory environment and its behavior. How can these new methods help us understand the governing principles shaping complex and naturalistic behavior? In this review, we will recap ways in which our ability to detect and model behavior have improved in recent years, and consider how these techniques might be used to revisit classical normative theories of behavioral control.
Original language | English (US) |
---|---|
Article number | 102549 |
Journal | Current opinion in neurobiology |
Volume | 74 |
DOIs | |
State | Published - Jun 2022 |
Funding
I am grateful to Jennifer J. Sun, Yisong Yue, Pietro Perona, and David J. Anderson for helpful discussions on the organization of behavior and the application of machine learning to behavior analysis, and to Brady Weissbourd for discussions on theories of social behavior. This manuscript was supported by the National Institute of Mental Health of the National Institutes of Health under Award Number R00MH117264. The content is solely the responsibility of the author and does not necessarily represent the official views of the National Institutes of Health.
ASJC Scopus subject areas
- General Neuroscience