Lecture 2: learning theory

Impact of Training on Animal Welfare

  • Training can impact animal welfare in two main directions: negatively or positively, depending on how it is used.

    • Negative impacts occur when unethical or aversive methods are used, triggering anxiety, fear, or pain in the patient, or when training (even if ethical) is too difficult and causes frustration.

    • Training can also be dangerous if reinforcement is missed or unmet expectations lead to aggression (e.g., marine mammal trainers). This can weaken the human–animal relationship.

  • Positive welfare outcomes from training when done ethically:

    • Uses methods that trigger positive emotions and minimize frustration.

    • Provides predictability and a sense of control over the environment, reducing stress.

    • Aligns with welfare principles (e.g., the direct model) that communicate owner expectations of behavior, increasing predictability and control for both owner and animal.

    • Can be enriching and facilitate learning, as animals are highly motivated to learn to adapt.

    • Strengthens the human–animal relationship.

  • Practical implications for household and veterinary contexts:

    • Clear rules in a household are not inherently bad; predictable expectations can reduce stress.

    • When done well, training can improve welfare and the quality of interactions between people and animals.

Training in the Veterinary Context

  • As family professionals, we are trusted sources of advice on animal training and training problems.

    • Owners will seek guidance online or from friends; high-quality, evidence-informed advice from veterinarians is highly valued.

  • Training can be applied to facilitate veterinary procedures and improve welfare for patients and staff:

    • Improves connection with patients, reducing risk of injuries from aggression.

    • Improves client satisfaction as owners see their animals are more comfortable, potentially increasing preventative and early intervention care.

  • Traditional restraint in practice can be emotionally negative for animals (fear, anxiety, pain) and may be justified by "for the animal's own good". There are better, increasingly used approaches:

    • Cooperative care and preparation for procedures to minimize restraint and stress.

    • These approaches are part of broader low-stress or fear-free veterinary practice, not the sole solution.

  • Example video (Dartmoor Zoo, UK): tiger training to facilitate veterinary procedures without general anesthesia for routine checks and dental work

    • Tigers learn to touch and position themselves to allow abdomen checks, mouth checks, and injections.

    • Training improves staff welfare and animal welfare by reducing stress and risk of injury.

    • Techniques demonstrated include positive reinforcement and cooperative handling.

  • Fear-Free / Low-Stress Accreditation:

    • Fear Free (USA) is well-known internationally.

    • In Europe, Dog Friendly Clinic (Dogs Trust) and Cat Friendly Clinic (BioCat Care / ISFM) are recognized accreditations.

  • Learning theory definitions:

    • Learning: the process by which an animal modifies its behavior as a result of experience.

    • Any species can learn if training is appropriate to the species and the individual.

    • Adapting training to the individual: e.g., juvenile dogs (short, playful sessions); arthritic older dogs may require gentler, longer sessions; cats may have different motivators than dogs; rabbits may require different tasks.

  • Training as enrichment and survival adaptation:

    • Learning is a survival strategy; animals are motivated to learn.

    • Properly implemented training can be enriching and support welfare.

Learning and Conditioning: Key Concepts

  • Two broad types of associative learning:

    • Classical conditioning (Pavlovian): association forms between a neutral stimulus and an unconditioned stimulus, leading to a conditioned response.

    • Operant conditioning (instrumental): learning is driven by consequences of behavior (trial and error).

Classical Conditioning: Overview

  • Basic sequence:

    • Unconditioned stimulus (US) naturally elicits an unconditioned response (UR).

    • A neutral stimulus (NS) is paired with the US until it becomes a conditioned stimulus (CS) that elicits a conditioned response (CR).

  • Terminology (example with Pavlov’s dogs):

    • US = food (unconditioned stimulus). US=extfoodUS = ext{food}

    • UR = salivation/happy response. UR=extsalivationUR = ext{salivation}

    • NS = a bell (initially neutral, no response).

    • After conditioning, CS = bell; CR = salivation/happiness in response to bell. CS=extbell, CR=extsalivation/happinessCS = ext{bell}, \ CR = ext{salivation/happiness}

  • Key requirements for conditioning:

    • The CS must precede the US (predicts the arrival of the US).

    • Temporal contiguity: the CS and US must be closely and consistently paired in time (contingency and contiguousness).

  • Important properties:

    • Emotions are central in veterinary practice (emotional conditioning matters for welfare).

    • Conditioning can occur quickly; sometimes, fear or anxiety can be classically conditioned after a single pairing.

    • Stimuli can generalize to similar contexts or situations.

  • Everyday examples:

    • Carriage/travel: a child’s fear associated with a car ride after a painful vaccination event.

    • Shoes: a dog becomes excited when owner puts on shoes because it predicts a walk.

    • Rustle of a bag: in a cat or dog, the sound becomes a cue for food or a meal.

  • Practical notes:

    • Avoid creating negative conditioning; be mindful of the emotional state when introducing cues.

  • Video example: goldfish training using lights to indicate correct behavior and deliver a food reward

    • The fish learns to associate light cues with a reward.

    • Observed indicators of welfare: engaged, anticipating, and approaching the target cue for reward.

  • Role of conditioning in practice:

    • Classical conditioning is foundational for behavior modification, desensitization, and counterconditioning.

Operant Conditioning: Overview

  • Definition: trial-and-error learning where the animal makes a conscious choice about behavior; consequences determine future repetitions.

  • Consequences and behavior:

    • If a behavior is followed by a desirable outcome, it is more likely to be repeated; if followed by an undesirable outcome, it is less likely to be repeated.

  • The four quadrants (reinforcement vs punishment; positive vs negative):

    • Positive reinforcement (PR): add something to strengthen the behavior (e.g., treat for sitting).

    • Negative punishment (NP): remove something desirable to reduce the undesired behavior (e.g., withhold a treat when the dog does not sit).

    • Positive punishment (PP): add something aversive to reduce the undesired behavior (e.g., shock collar; added discomfort for barking or pulling).

    • Negative reinforcement (NR): remove something aversive to strengthen the desired behavior (e.g., release pressure after the dog stops pulling).

  • Important clarifications about terms:

    • Positive/negative do not imply good/bad; they indicate addition or removal of a stimulus.

    • Punishment is a consequence that reduces a behavior; it is not inherently bad, but positive punishment is generally discouraged due to welfare and learning concerns.

  • Practical cautions:

    • Positive punishment is generally avoided because it can cause pain, fear, anxiety, weaken the human–animal bond, and may provoke aggression.

    • Negative punishment should be used in combination with positive reinforcement; using punishment alone can hinder learning and increase frustration.

  • Practical examples:

    • Positive reinforcement: giving a treat when a dog sits.

    • Negative punishment: withholding a treat when the dog does not sit.

    • Positive punishment: applying a shock to stop barking or too much pulling.

    • Negative reinforcement: releasing pressure when the dog reduces pulling, thereby increasing the likelihood of walking nicely.

  • Connection to practice:

    • Reinforcement schedules modulate learning speed and persistence.

    • The timing of reinforcement is critical for strengthening the desired behavior.

Reinforcers and Schedules

  • Types of reinforcers:

    • Primary reinforcers (unlearned, biologically important): include food, water, play, exploration. Some animals may find social interaction or play as primary reinforcers.

    • Secondary reinforcers (learned, via association): praise, tapping, clicker, certain sounds; often require classical conditioning to be effective.

    • The clicker is a common example of a secondary reinforcer (a conditioned cue that signals that a primary reinforcer is forthcoming).

    • The framework principle (lifestyle reward): the reward for doing the desired behavior is getting to do something the animal already wants to do (e.g., allowing a dog to go outside only after it has calmed down).

  • Individual differences in reinforcement:

    • Animals have personal hierarchies of reinforcement; preferences vary (e.g., one dog may love popcorn, others may prefer cheese or a toy).

    • Cats and rabbits may be motivated by different rewards than dogs; always determine species- and individual-specific preferences.

    • Stress can alter reinforcement effectiveness: in a stressed animal, appetite for food may be suppressed; other forms of reassurance or trust-building may be needed.

  • Practical considerations:

    • For obesity concerns, portion control and timing matter; smaller, chopped rewards can be used to maximize value per unit of food.

    • When animals are about to undergo anesthesia or procedures, alternative methods such as water rewards or non-food cues can be used to maintain motivation without overeating.

  • Reinforcement schedules:

    • Continuous reinforcement (every correct response) helps establish a new behavior quickly.

    • Intermittent reinforcement strengthens behavior more persistently (e.g., after a variable number of correct responses; similar to a slot machine effect).

    • Jackpot rewards can be used as a rare, exceptionally large reward for exceptional performance, but should be used sparingly to avoid loss of focus.

  • Practical considerations for using reinforcement:

    • The timing of reinforcement should be immediate and properly aligned with the target behavior; use cueing tools to mark the exact moment the desired behavior occurs (e.g., a clicker or a marker word).

    • If you reinforce too late or for incorrect behaviors (e.g., rewarding eye contact while the dog is not in the desired position), you can mislead the animal about what you’re rewarding.

    • A continuous reinforcement schedule can be followed by a gradual shift to partial reinforcement to strengthen behavior over time.

Tools and Techniques in Practice

  • Marking and bridging tools:

    • Clicker or audible marker (e.g., words like "yes" or "good"): used to precisely mark the desired behavior so you can deliver reinforcement promptly.

    • A target stick can be used to guide the animal to touch it with nose or paw, enabling trainers to maneuver the animal into new positions.

  • Training sequences and shaping:

    • Shaping: break a complex task into smaller, manageable steps and reinforce gradually as each step is performed correctly.

    • Capturing: reward a behavior as it occurs naturally, without prompting.

    • Chaining: link together a series of behaviors to complete a more complex task.

  • Cueing hierarchy:

    • Start with a body language cue (e.g., a hand gesture) since animals often respond better to visual cues than verbal ones.

    • Then add a verbal cue once the body cue is established.

    • The order is body language cue first, then verbal cue; allow the cue to stay distinct from the behavior before adding more cues.

  • Training plan and criteria:

    • Define the goal behavior and the steps required to achieve it.

    • Set training criteria (e.g., 80% success or 4/5 trials, or 8/10 trials) to determine when to advance to the next step.

    • If performance is around 60%, maintain the current step and practice more; if 50% or less, make the task easier.

    • Trainers should be flexible and adaptive and adjust training to different locations due to varying distractions.

  • Training environment and generalization:

    • Train in multiple locations to ensure the behavior generalizes across contexts and is not limited to a single environment.

  • Application in complex behaviors:

    • For complex tasks, plan training with a clear hierarchy of steps and a plan to increase difficulty gradually.

  • Specific practice tips:

    • Use a continuous reinforcement schedule early on for new behaviors, then gradually move to variable schedules for stability.

    • Use jackpot rewards sparingly, late in the session, to avoid disrupting concentration.

  • Ethical considerations in training:

    • Favor positive reinforcement and negative punishment over positive punishment or aversive methods.

    • Avoid relying on negative punishment alone; combine with positive reinforcement to maintain motivation and reduce frustration.

    • Remember that negative reinforcement and positive punishment often form a linked sequence and should be used with caution.

Behavioral Modification Techniques

  • Response substitution (differential reinforcement of incompatible behavior):

    • Train the animal to perform a compatible behavior to replace an undesirable one (e.g., hold something in mouth to replace barking).

  • Extinction:

    • A previously reinforced behavior fades when reinforcement is no longer provided; can be accompanied by an extinction burst (temporary increase in the undesired behavior) and potential frustration unless a motivating alternative is taught.

  • Conditioning and counterconditioning:

    • Conditioning (classical): change emotional response from negative to positive by pairing a fear-evoking stimulus with a positive outcome.

    • Desensitization: gradually expose the animal to a stimulus at a low level the animal can tolerate, increasing exposure gradually.

    • Counterconditioning: pair the stimulus with a positive outcome to change the emotional response.

  • Nonassociative learning (habituation and sensitization):

    • Habituation: a decreased response to a stimulus after repeated exposure without any consequence.

    • Desensitization is similar in approach but used for increasing exposure gradually; sensitization is the opposite process (increased response) and is typically undesirable for welfare.

  • Other learning concepts:

    • Nondiscrimination of hearing (nonassociative learning): changes in behavior due to the frequency or duration of exposure rather than the consequences.

    • Observational learning: learning by watching others perform a behavior (common in primates; less common in dogs and cats).

    • Insight learning and latent learning: behaviors appear after processing information learned earlier; often studied in research.

Ethical Framework and Best Practices

  • Ethical conditioning emphasizes a combination of reinforcement-based methods:

    • Emphasize positive reinforcement paired with negative punishment to reduce unwanted behaviors.

    • Avoid positive punishment due to risk of fear, pain, aggression, and weakening the human–animal bond.

  • Practical guidance for clinicians and students:

    • Always prioritize welfare-first approaches and avoid methods that cause distress.

    • Adapt training plans to individual animals, considering age, health, stress levels, and past experiences.

    • Use environment and enrichment to support learning and welfare.

    • Maintain clear communication with clients about training goals, expectations, and safety considerations.

Real-World Considerations and Takeaways

  • Training is not a one-size-fits-all approach; tailor to species, breed, age, health, and individual temperament.

  • The welfare benefits of training come from predictability, control, enrichment, and a strengthened human–animal bond when done with ethical practices.

  • In veterinary settings, cooperative care and fear-free practices can reduce restraint needs and improve patient and staff safety while encouraging preventative and early-intervention care.

  • The use of training tools (clickers, target sticks) and cue sequencing (body language first, then verbal) helps precision and reduces unintended reinforcement.

  • Consistent and well-planned training with appropriate reinforcement schedules builds durable skills and improves welfare for both animals and caregivers.

Quick Reference: Key Terms and Concepts

  • US, UR, CS, CR: foundational classical conditioning terms.

    • US=foodUS = \text{food}

    • UR=salivationUR = \text{salivation}

    • CS=bellCS = \text{bell} (after conditioning)

    • CR=salivation/happinessCR = \text{salivation/happiness} in response to the bell

  • Contingency and contiguousness: timing and predictability of CS–US pairing.

  • Positive reinforcement (PR): add a reward to strengthen a behavior.

  • Negative reinforcement (NR): remove an aversive stimulus to strengthen a behavior.

  • Positive punishment (PP): add an aversive stimulus to reduce a behavior.

  • Negative punishment (NP): remove a desirable stimulus to reduce a behavior.

  • Primary reinforcers: food, water, play, exploration.

  • Secondary reinforcers: praise, clicker, conditioned cues.

  • Framework principle: reward the animal with something they want in return for the target behavior.

  • Continuous reinforcement (CRF) vs intermittent reinforcement (e.g., VR, VI schedules).

  • Jackpot: a large, one-off reward for exceptional performance.

  • Shaping, capturing, cueing, and task analysis for complex behaviors.

  • Desensitization and counterconditioning: increase tolerance to a stimulus and pair it with a positive outcome.

  • Habituation and sensitization: nonassociative learning processes.

  • Observational and insight/latent learning: learning from others and processing knowledge to act later.

"}