Automated Behavioral Profiling Using Neural Networks Reveals Differences in Stress-Like Behavior Between Cave and...

Area of Science:

• Evolutionary biology
• Animal behavior
• Neuroscience

Background:

• Behavioral stress responses are crucial for animal survival and adaptation.
• The Mexican cavefish (Astyanax mexicanus) offers a unique model system due to divergent stress responses between surface and cave morphs.
• Ecological differences, such as predator presence, are hypothesized to drive these divergent stress responses.

Purpose of the Study:

• To comprehensively examine and quantify the complex behavioral differences in stress responses between surface and cave Mexican cavefish.
• To investigate the genetic basis of these behavioral differences using F2 hybrids.
• To highlight the utility of computational tools in behavioral phenotyping for evolutionary studies.

Main Methods:

• Automated pose-tracking and machine learning tools were employed for high-throughput behavioral quantification.
• A novel tank assay was used to measure stress-associated behaviors like freezing, bottom-dwelling, and hyperactivity.
• Analysis of F2 hybrids was conducted to assess the co-segregation of specific stress behaviors.

Main Results:

• Surface fish displayed significantly higher stress responses, including prolonged bottom-dwelling and frequent freezing.
• Cavefish exhibited reduced stress behaviors, characterized by increased exploration and minimal freezing.
• Freezing and bottom-dwelling behaviors were found to co-segregate in F2 hybrids, indicating shared genetic or physiological factors.

Conclusions:

• Computational behavioral phenotyping precisely quantifies complex traits, aiding evolutionary research.
• Divergent ecological pressures have shaped distinct stress response behaviors in Mexican cavefish morphs.
• This study provides a framework for understanding the evolution of integrated behavioral and physiological traits.