Analysis of the spatiotemporal dynamics of behavior in domestic dogs with free locomotion under appetitive Pavlovian contingencies

This study demonstrates that continuously tracking the locomotion of freely moving domestic dogs reveals distinct spatial behavioral patterns, such as increased proximity to food dispensers and extended trajectories, which are specifically modulated by the presence or absence of appetitive Pavlovian tone-food contingencies.

The Gist

Imagine you are watching a dog in a room. Usually, when we study how dogs learn, we just ask: "Did the dog sit when I said 'sit'?" or "Did the dog look at the treat?" But this study asked a much bigger question: "How does the dog move through the room when it knows something good is coming?"

The researchers wanted to see if the space itself becomes part of the lesson. They treated the room like a giant map and watched how the dogs drew their paths on it.

Here is the story of what they found, explained simply:

The Setup: The Magic Speaker and the Food Machine

The researchers set up a room with two main characters:

1. The Owner: Sitting quietly in one corner (like a calm anchor).
2. The Magic Machine: A food dispenser on the opposite side.

They used a simple rule: A Tone = Food.

• Phase 1 (The "Magic" Phase): A buzzer sound would play, and 1.5 seconds later, a tasty treat would drop from the machine. This happened every minute, whether the dog moved or not.
• Phase 2 (The "Boring" Phase): The buzzer would play, but nothing happened. No food. Just the sound.

They did this back and forth to see how the dogs' walking patterns changed.

The Results: Two Very Different "Dance Styles"

Think of the dogs' movement as a dance. The type of music (the contingency) changed the dance completely.

1. When the Food Came (The "Hunt" Dance)

When the dogs learned that the tone meant food was coming, their movement became expansive and energetic.

• The Metaphor: Imagine a detective searching a crime scene. They aren't just standing still; they are pacing back and forth, checking the corners, and circling the area where the clue (the food) might appear.
• What happened: The dogs covered more ground. They walked longer distances. They didn't just sit near the machine; they moved between the machine and the owner, creating long, winding paths.
• The "Conditional Approach": When the tone played, the dogs would suddenly head toward the food machine. It was as if the space near the machine became a "magnet" for their attention.

2. When the Food Stopped (The "Huddle" Dance)

When the tone played but the food never came (the "Extinction" phase), the dogs' movement became restricted and calm.

• The Metaphor: Imagine a group of friends at a party who realize the party is over and no one is bringing out the cake. They stop wandering the room and just huddle together near the host, waiting for the next cue or deciding to leave.
• What happened: The dogs stopped pacing. They stayed closer to their owners or the edges of the room. They walked shorter distances and their paths became tight little loops. They stopped treating the food machine like a destination.

The Big Takeaway: Space is a Character, Not Just a Background

The most important discovery is that space isn't just an empty box where behavior happens; space is part of the learning process.

• In the "Magic" Phase: The area near the food machine became a "special zone." The dog's brain linked the sound, the food, and the location together. The dog learned, "When I hear that sound, I need to be in this specific part of the room."
• In the "Boring" Phase: That special zone lost its magic. The dog's brain reorganized, and the "safe zone" shifted back to being near the owner.

Why Does This Matter?

This study is like upgrading from a black-and-white photo to a 4K video of how animals learn.

1. Better Science: It proves that we can't just look at whether an animal "responds" or not. We have to look at how they move through space to understand what they are thinking.
2. Better Dog Welfare: The study suggests that giving dogs "predictable" but active environments (where they have to move and explore to get rewards) might be better for their mental health than just keeping them in a small, boring cage. It turns a boring day into an engaging "hunt," which reduces stress and boredom.

In a nutshell: The dogs didn't just learn a sound; they learned a map. When the map promised a reward, they explored the whole territory. When the map went blank, they retreated to the safety of their owner's side.

Technical Summary

1. Problem Statement

While the spatiotemporal dynamics of behavior are recognized as crucial for understanding organism-environment relationships, their specific interaction with Pavlovian (classical) contingencies in freely moving domestic dogs remains understudied.

• Historical Context: Previous research on spatial behavior in Pavlovian conditioning (e.g., Kupalov, 1969) often relied on "discretized" trials where the presentation of stimuli depended on the animal's location (an operant interaction), making it difficult to isolate pure Pavlovian effects.
• Methodological Gap: Most existing studies on dogs rely on qualitative observations or pictorial representations rather than continuous, quantitative tracking. There is a lack of automated, high-resolution data regarding movement intensity, distribution, and variability in space relative to stimulus events.
• Objective: To quantitatively analyze how the spatial dimension of behavior organizes itself under appetitive Pavlovian contingencies (tone-food pairing) versus extinction (tone only) in domestic dogs, using continuous tracking tools to separate Pavlovian effects from operant interactions.

2. Methodology

Experimental Subjects

• Participants: Three experimentally naive domestic dogs (two mixed breeds, one pit bull; ages 4–7 years).
• Conditions: Food-deprived (~12 hours) but clinically healthy; owners present during sessions to minimize isolation stress.
• Ethics: Approved by the Internal Committee for the Care and Use of Laboratory Animals; followed Mexican standard NOM-062-ZOO-1999.

Experimental Setup & Equipment

• Environment: A domestic-naturalistic room (2.80m × 2.66m) with a fixed food dispenser and a seated owner.
• Stimuli:
  • CS (Conditioned Stimulus): 3-second tone (23 kHz, 75 dB).
  • US (Unconditioned Stimulus): One kibble of premium dog food.
• Tracking: An overhead RGB camera (1920×1080, 10 fps) recorded the session. Software (Ethovision XT) tracked the center of mass (X, Y coordinates) at 10 Hz resolution.

Design & Procedure

• Design: Single-case reversal design (B–A–B–A) replicated across three subjects.
  • Condition B (Pairing): Tone (CS) presented, followed 1.5s later by food (US) on a Fixed-Time 60s (FT 60s) schedule. (10 sessions, 16 trials/session).
  • Condition A (Extinction): Tone presented, but no food delivered. (10 sessions, 16 trials/session).
• Familiarization (CST): Initial session with tone only to establish baseline orientation.

Data Analysis Metrics

The study utilized Motus and R for analysis, calculating:

1. Movement Routes: Continuous trajectories connecting X,Y coordinates.
2. Total Distance Traveled: Sum of Euclidean distances between consecutive coordinates.
3. Location Entropy: Shannon entropy ($H$) calculated on a 9×8 grid (72 cells) to measure spatial variability/distribution. High entropy = uniform distribution; low entropy = concentrated in few areas.
4. Distance to Objects: Mean Euclidean distance to the Dispenser and the Owner.
5. Conditional Approach Index (CAI): Ratio of time spent near the dispenser during the CS presentation vs. total CS duration (0 to 1 scale).

• Statistical Model: Linear Mixed Models (LMM) with Condition (Pairing/Extinction) as a fixed effect and Subject ID as a random intercept.

3. Key Results

Spatial Dynamics & Movement Patterns

• Pairing (B): Dogs exhibited extended trajectories, covering significantly greater total distances. Movement patterns involved back-and-forth trajectories between the dispenser area and peripheral areas (including the owner's location).
• Extinction (A): Dogs showed restricted movement patterns, traveling shorter distances and staying closer to the owner or room periphery.
• Statistical Significance: LMM analysis confirmed a significant difference in total distance traveled ($F(1, 116) = 129, p < .001$) and location entropy ($F(1, 116) = 90.4, p < .001$), with both metrics being higher during pairing.

Spatial Organization & Proximity

• Distance to Dispenser: During pairing, dogs maintained a significantly shorter distance to the dispenser. During extinction, this distance increased.
• Distance to Owner: Conversely, proximity to the owner increased during extinction and decreased during pairing.
• Conditional Approach: The Conditional Approach Index (CAI) showed a clear divergence. In pairing, trials showed extreme values (approaching 1), indicating a strong conditional approach to the dispenser during the tone. In extinction, values clustered near 0, indicating a lack of approach.
  • Result: Significant effect of condition on CAI ($F(1, 1916) = 228, p < .001$).

Visual Evidence

• Heatmaps of cumulative time showed high saturation (density) near the dispenser during pairing, shifting to the owner/periphery during extinction.
• Entropy plots confirmed that pairing induced a more distributed spatial exploration, while extinction led to spatial confinement.

4. Key Contributions

1. Methodological Innovation: The study successfully applied continuous, automated 2D tracking to Pavlovian conditioning in dogs, moving beyond qualitative observation or discretized trials. This allows for the isolation of Pavlovian effects from operant behaviors.
2. Integration of Space into CS-US Relations: It demonstrates that spatial segments (specifically the area near the dispenser) acquire a conditional function. The space itself becomes part of the contingency, organizing the organism's behavior even when the animal is free to move.
3. Differentiation of Contingency Types: The study clearly distinguishes behavioral organization driven by Pavlovian contingencies (tone-food) from those driven by extinction or operant history, showing that the type of contingency dictates the spatial topology of behavior.
4. Quantitative Metrics: Validated the use of Location Entropy and Conditional Approach Index as robust metrics for analyzing spatiotemporal dynamics in freely moving subjects.

5. Significance & Implications

• Theoretical Impact: Supports the hypothesis that space is not a passive backdrop but an active participant in learning. It validates the concept of "situational conditional reflexes" where specific locations become conditioned stimuli themselves.
• Animal Welfare Applications:
  • The findings suggest that Pavlovian arrangements (predictable, non-contingent food delivery) can serve as a form of environmental enrichment.
  • The increased distance traveled and higher spatial variability (entropy) observed during pairing indicate that such setups can reduce confinement and increase physical activity, potentially mitigating stress and compulsive behaviors in domestic dogs.
• Future Directions: The authors suggest that incorporating 3D tracking and multipoint analysis (to capture posture and orientation) could further refine the understanding of how organisms integrate temporal and spatial dimensions in learning.

Conclusion

This study provides robust empirical evidence that spatiotemporal dynamics are fundamental to Pavlovian learning in domestic dogs. By utilizing continuous tracking, the authors demonstrated that the presence of a food contingency organizes the dog's movement into specific, high-variability patterns centered on the reward source, whereas extinction leads to a reorganization of behavior toward social (owner) or peripheral safety zones. This bridges the gap between classical conditioning theory and spatial behavior analysis.