Flocks in focus: Automated video analysis of spatial behavior for stress detection in aviary-housed laying hens

TL;DR

This study uses automated video analysis to detect stress in laying hens by tracking their spatial behavior in aviary systems.

Contribution

The study introduces automated spatial behavior monitoring as a novel method for detecting stress in commercial laying hens.

Findings

Vertical movement decreased during predator exposure but increased during thunder and feeding delays.

Litter use declined under all stress conditions, with the lowest use during predator and thunder exposure.

Vertical movement consistency increased during recovery, suggesting synchronized vigilance in stressed flocks.

Abstract

Understanding stress responses in laying hens is crucial for improving welfare in commercial systems, yet real-time behavioral indicators remain underexplored. This study evaluated the use of automated spatial behavior monitoring to detect stress-induced behavioral changes in commercial-density laying hen flocks housed in aviary systems. The objectives were to (1) visualize and quantify vertical movement patterns and litter use over time and across flocks, and (2) assess behavioral changes in response to 3 stress contexts: visual (predatory bird), auditory (thunder sound), and frustrative (delayed feeding). Four flocks of 38-week-old hens were exposed weekly to each stressor over a 10-week period. Video data were analyzed using Python-based algorithms to assess vertical movement and litter use 1 hour before and after stress exposure. The effects of different stressors on vertical movement were analyzed using a Generalized Linear Mixed Model (GLMM), treating the experimental groups as biological replicates and including week as a random effect to account for temporal variation. Vertical movements significantly decreased during predator exposure (−8 movements/min, P < 0.001), but increased during thunder exposure (+6 movements/min, P < 0.001), delayed feeding (+1 movement/min, P < 0.005), and during the first 5 minutes of delayed feed supply (+3 movements/min, P < 0.001). GLMM analysis showed that litter use declined across all stress conditions, with the lowest densities recorded during predator and thunder exposure. Consistency of temporal patterns in spatial behavior varied across days, behavioral measures (i.e., vertical movement and litter use), and within the observation period. Significant differences in within-period consistency were confirmed using Friedman tests (P < 0.001), suggesting a potential link with induced stress. Vertical movement was most consistent in the afternoon in the absence of external disturbances. Stress exposure increased vertical movement consistency during recovery, potentially indicating synchronized vigilance. In contrast, litter use was more variable and sensitive to environmental changes. While this makes litter use less reliable as a standalone stress indicator, it might be a useful health and welfare indicator during predictable periods of the day, such as post-laying morning hours. These findings suggest that vertical movement consistency may serve as a promising behavioral indicator of stress recovery, while litter use patterns could inform welfare assessments during stable daily periods.

Figures