Operant Conditioning in Indian Free-Ranging Dogs: Effects of Positive and Threatening Cues on Sociability

Technical Summary: Operant Conditioning in Indian Free-Ranging Dogs

Problem Statement
Sociability toward humans is a critical adaptive trait for free-ranging dogs (FRDs), facilitating resource acquisition while necessitating risk management in human-dominated environments. While previous research has established that FRDs can learn from human interactions, it remains unclear whether operant conditioning shapes sociability in non-approachable individuals and how these learned responses generalize to unfamiliar humans. Specifically, there is a need to understand the differential learning dynamics of positive (reward-based) versus threatening (aversive) cues, including whether a "negativity bias" (faster learning to avoid threats than to approach rewards) exists in this population and whether learned behaviors are specific to the interacting individual or generalize to strangers.

Methodology
The study utilized a longitudinal experimental design involving 58 groups of Indian free-ranging dogs across four urban and semi-urban sites in West Bengal, India. Groups were defined as three or more dogs within five meters of each other, with a minimum of three healthy adults per group.

• Experimental Design: Groups were randomly assigned to one of two conditions: a Positive Cue group (receiving food rewards) or a Threatening Cue group (exposed to a wooden stick).
• Procedure: Over five consecutive days (Days 1–5), a familiar experimenter (E1) conducted an approachability test followed by the administration of the assigned cue. On Day 6, an unfamiliar experimenter (E2) conducted the same approachability test, followed by E1 repeating the test, to assess generalization.
• Data Collection: Behavioral parameters were coded from video recordings, including:
  1. Approach Proportion: The fraction of dogs in a group approaching within one body length (~0.8 m).
  2. Approach Latency: Time taken to approach (censored at 30s for non-approachers).
  3. Demeanor: Categorized as aggressive, anxious, neutral, or affiliative.
• Statistical Analysis: Data were analyzed using Bayesian Generalized Linear Mixed-Effects Models (GLMMs) in R. The study employed zero-one inflated beta regression for approach proportions, lognormal survival models for latency (accounting for right-censoring), and multinomial regression for demeanor. Random effects accounted for group and individual identity.

Key Results
The study found that operant conditioning significantly shapes sociability, with distinct patterns for positive and threatening cues:

• Approach Behavior:
  • Positive Cue: Dogs showed a significant increase in approach proportion and a reduction in approach latency over repeated days. Notably, this increased approach behavior generalized to an unfamiliar experimenter on Day 6A, although approach latency toward the stranger remained high, indicating hesitation.
  • Threatening Cue: Dogs exhibited a significant decrease in approach proportion and an increase in approach latency over time. Crucially, this reduction in approach did not generalize to the unfamiliar experimenter; dogs did not reduce their willingness to approach strangers, but they did display increased latency (caution) when approaching the familiar experimenter on later days.
• Demeanor:
  • Positive Cue: Repeated exposure led to increased affiliative demeanor and reduced neutral behavior. Anxious and aggressive behaviors remained relatively stable.
  • Threatening Cue: Repeated exposure resulted in a shift from affiliative and aggressive behaviors toward neutral demeanor. Anxious demeanor increased initially (Days 3–4) but stabilized.
• Learning Dynamics and Generalization:
  • Contrary to the hypothesis of a strong "negativity bias" in learning speed, the study did not find that threat avoidance learning was significantly faster than reward-based learning in terms of initial approach reduction.
  • Asymmetric Generalization: Positive experiences partially generalized, increasing overall sociability toward strangers (though with hesitation). Threatening experiences did not generalize to reduce approach toward strangers; instead, they increased caution (latency) specifically toward the familiar threat source.
• Sex Differences: Female dogs generally displayed higher approach latencies and a higher probability of neutral demeanor compared to males, suggesting greater caution.

Key Contributions

• Operant Conditioning in FRDs: The study provides empirical evidence that operant conditioning modifies sociability in free-ranging dogs, even in individuals that do not initially approach humans.
• Generalization Patterns: It identifies an asymmetry in how positive and negative experiences generalize. Positive reinforcement fosters a generalized increase in trust/sociability, whereas negative reinforcement leads to specific caution toward the threat source without broadly suppressing sociability toward all humans.
• Behavioral Plasticity: The findings demonstrate that FRDs possess significant behavioral plasticity, adjusting their strategies based on accumulated experience rather than fixed personality traits alone.

Significance and Claims
The authors claim that these findings reflect an adaptive strategy for free-ranging dogs to balance resource acquisition with risk avoidance. By maintaining a baseline willingness to approach humans (facilitated by positive generalization) while simultaneously increasing caution in response to specific threats (without broadly suppressing sociability), FRDs can navigate the unpredictable nature of human-dog interactions.

The paper concludes that these insights have practical implications for improving human-dog coexistence. Understanding that positive interactions can generalize to improve overall trust, while negative interactions primarily induce specific caution rather than broad avoidance, can inform the design of management strategies and public awareness initiatives aimed at reducing conflict in urban environments. The study emphasizes that behavioral modification through experience is a key mechanism in the survival and adaptation of free-ranging dog populations.