A Wireless Wearable Ecosystem for Social Network Analysis in Free-living Animals

Abstract

Studying animal social systems requires understanding variations in contact and interaction, influenced by factors like environmental conditions, resource availability, and predation risk. Traditional observational methods have limitations, but advancements in sensor technologies and data analytics provide new opportunities. We developed a wireless wearable sensor system, "Juxta," with features such as modular battery packs and a smartphone app for data collection. A pilot study on free-living prairie voles (Microtus ochrogaster), a species with complex social behavior, demonstrated Juxta's potential for studying social networks and behavior. We propose a framework for merging temporal, spatial, and event-driven data, which can help explore complex social dynamics across species and environments.

Keywords: Wireless sensors; behavioral tracking; contact tracing; neuroethology; social networks.

Fig. 1.

Juxta. (a) The battery is regulated to provide a stable power source to the Bluetooth Low Energy (BLE) Microcontroller Unit (MCU). A battery voltage sensor is implemented using two passive components. The memory and inertial sensor read/write over central communication lines. (b) The physical circuit board (c) is packaged in a 3D-printed case.

Fig. 2.

Smartphone Application. (a) The main interface connected to a device. The upper third shows device information, the middle third reads/writes device settings, the lower third contains a data terminal. (b) The advanced Options panel. (c) Terminal output for log data.

Fig. 3.

Exemplar Data from Single Enclosure. (a) Movement events with 1-minute resolution (black) and device temperature data (red) over roughly 24 hours for Vole 06 (V06) in 20-minute bins. Movement was normalized where 1 means there was a movement event for every 1-minute interval within the bin. (b) Scanning activity of unique (colored) nearby devices (“other voles”) over time for V06. The size of the circle represents the RSSI, a proxy for physical distance (legend upper left) to another vole. E.g., Shortly after 00:00, the data indicate V06 (main subject), V13, and V29 were together. (c) Adjacency matrix for 37 voles where connecting lines indicate bidirectional detection of two voles; line width represents pair strength. Vole V06 (main subject) is highlighted with an arrow; other connections are set to 20% opacity. (d) Relationship between movement and device temperature for each vole as the slope of the linear fit between those data; negative slopes indicate anticorrelation. Signification (p < 0.05) correlations marked with red *. V06 slope marked in blue.