November 16, 2021
Written by: Lindsay Ejoh
Since prehistoric ages, humans have been pondering animal behavior. Cavemen paintings dating back to 40,000+ years ago show images of animals herding and performing other behaviors, proving that we have been curious about the minds of animals for millenia1.
Charles Darwin, famous for his evolutionary theory of Natural Selection, is a critical figure in the study of animal behavior, also known as ethology. In his book On the Origin of Species he wrote about instinct and the evolution of behaviors. Darwin pondered whether behaviors could evolve across generations, which would explain why some behaviors in humans and other animals are instinctual.
Darwin’s studies of behavior employed a method commonly used among scientists of his time: observational research. Even from a young age, he had a strong interest in nature and would record everything. Young Darwin would collect insects and birds and take extensive notes on their behavior. Observational research is a useful method of studying animal behavior, because it provides very detailed information about their behavior in a low cost, simple manner. Since it is not experimental, there are minimal ethical concerns as animals are not manipulated in any way. However, this method is time consuming and relatively subjective, leaving it prone to bias and other confounding variables3.
How do we study behavior now?
Ethological methods have progressed a lot throughout the ages, and observational research is not the gold standard anymore. Though there are many ways modern-day ethologists study animal behavior, I will highlight an emerging field that combines ethology with a branch of computer science called machine learning to automate behavioral tracking.
Machine learning is a type of artificial intelligence in which computer systems can learn from experience, identify patterns, and make decisions without human intervention4. These systems can be used to extract information from video recordings of animals in order to detect and analyze movement patterns in their behavior. With this data, researchers can study different aspects of an animal’s behavior, such as their posture, movements, mating calls, social interactions, and more.
Deep neural networks
Scientists have developed specialized machine learning systems called deep neural networks to train computer networks to recognize and classify behaviors. In this way, you can teach the computer program to recognize a certain behavior by training it with many examples of animals performing that behavior. After training, the program can learn to recognize that behavior in new videos that it has not seen before. The benefits of using machine learning to characterize behaviors are that it is automated, faster and much less subject to experimenter bias5.
Why study animal behavior in the first place?
Since we cannot exactly ask an animal how it is feeling or what it is experiencing, we measure behavior to gain insight into its lived experience, which will help us understand their minds, and our minds, better. From a Darwinian perspective, studying animal behavior may allow us to gain insight into the evolution of certain behaviors, distinguish between what is innate and what is learned, and identify the aspects of our nature that make us uniquely human6.
Create your own experiment!
Let’s play scientists for a moment. What questions do you have about how the mind works? They can be related to mood, decision-making, sleep, addiction, how certain psychiatric disorders develop, etc. Can you think of a way to ethically study animal behavior as a way to gain insight into those questions?
- Sinervo, B. (n.d.). Chapter 1: History and philosophy of behavioral analysis. History and Philosophy of the Study of Animal Behavior, Sinervo©1997. Retrieved November 8, 2021, from https://bio.research.ucsc.edu/~barrylab/classes/animal_behavior/HISTORY.HTM.
- Thierry B. (2010). Darwin as a student of behavior. Comptes rendus biologies, 333(2), 188–196. https://doi.org/10.1016/j.crvi.2009.12.007
- Hess A.S., Abd-Elsayed A. (2019) Observational Studies: Uses and Limitations. In: Abd-Elsayed A. (eds) Pain. Springer, Cham. https://doi.org/10.1007/978-3-319-99124-5_31
- Jordan, M. I., & Mitchell, T. M. (2015). Machine learning: Trends, Perspectives, and prospects. Science, 349(6245), 255–260. https://doi.org/10.1126/science.aaa8415
- Cepelewicz, J. (2019, December 10). To decode the brain, scientists automate the study of behavior. Quanta Magazine. Retrieved November 8, 2021, from https://www.quantamagazine.org/to-decode-the-brain-scientists-automate-the-study-of-behavior-20191210/.
- Smulders, T. (2021, October 26). By studying animal behaviour we gain an insight into our own. The Conversation. Retrieved November 8, 2021, from https://theconversation.com/by-studying-animal-behaviour-we-gain-an-insight-into-our-own-20001.