Animal Learning Flashcards
What is learning?
- Specific experiences result in a change in response to a given situation
- Many different ways to get from A to B => many different training methods
- Appropriate training method depends on the situation and the individual animal
What can we do with learning & training?
- Increase or decrease particular behaviors
- Improve “manners” and self-control
- Enrichment and confidence-building
- Change the affective state of an animal in response to a particular situation
- Try to reduce negativity of several stimuli
Simple Learning
- Animals learn which stimuli they need to pay attention to in their environment
- Which are important & unimportant
- Learn to make smart decisions in specific contexts -> habituation and sensitization
Habituation: repeated exposure to the stimulus results in a decrease in response over time
- Can be both positive and negative
- Ex: trying to eliminate fear in dog for specific object by exposure
- NOT the same as acclimation
- Acclimation - physiological and behavioral adjustment to environmental change
Sensitization: repeated exposure to the stimulus results in an increase in response, sometimes generalizes to other similar stimuli
- S1 -> R (might see small response)
- S1 -> R (more exposure = bigger response)
- S1 & S2 -> R
- Generalization of sensitization across similar stimuli
Systematic Desensitization
- Step-wise exposure to a gradually increasing intensity of an aversive stimulus to decrease the response
- Start around threshold for response - understand how far away the dog can be to not see a response
- Repeat such that no aggression or very low levels are seen - repeat at a level of threshold, every few days
- Increase to next level only when it has shows repeated instances over time, of neutral/calm
- Important to be able to read the animal’s behavior
- Be patient! - slowly incorporate to reduce negative response
Flooding
- high level of forced exposure, not a good learning tool & can increase fear response
- Animals often shut down, therefore may see reduced fear & aggression (learned they can’t escape the situation)
- Assess physiological parameters (pupil dilation, heart rate etc.) & rigid posture to assess for extreme fear w/ freezing behavior
- Or can amplify response = increased fear and aggression, which risks safety of people
Associative Learning
classical conditioning
- strengthen relationship b/twn 2 stimuli
- Unconditioned response - not learned response => ex: provide food & dog salivate
- Neutral stimulus - should not cause response
- During conditioning we pair the neutral stimulus with the unconditioned stimulus
- Repeat to strengthen the relationship b/twn the neutral and unconditioned stimulus
- After conditioning, neutral stimulus becomes conditioned stimulus
- ex: blow whistle -> dog salivate
Counter-conditioning (a type of classical conditioning)
- Can be used to change emotional states associated w/ a situation or procedure by pairing it with something good
- be aware of potential for opposite effects
- Something that was perceived as positive becomes negatively associated due to learned association w/ something VERY negative
- won’t see it occur if trear or reward is high value to animal (toy or attention)
operant conditioning
- strengthen relationship b/twn a behavior and the consequence of the behavior
- Outcome is good, behavior occurs more often
- Outcome is bad, behavior occurs less often
Key points to remember
- Consistency - need to do things in the same way, repeatedly
- Reward stimulus -> aversive stimulus - high value rewards are important
- Gradually build up the procedure while keeping fear/arousal sub-threshold
- Start off slow, read animal’s behavior
- Separate components where possible - breaking things down may help increase success
- Repetition - short periods of time b/twn training sessions (1 to every few days, not weeks/months)
Food training reward fear and aggression?
- Changing an animal’s emotional state w/ food
- treats/food - raise dopamine levels in brain = feelings of pleasure
- reinforces food seeking behavior
- changes negative emotional state to positive
- only reward behavior you want to see
Animal learning processes
positive reinforcement
adding something to increase behavior
- ex: giving treats
positive punishment
adding something to decrease behavior
- ex: shock collar
negative reinforcement
removing something to increase behavior
- ex: pushing on dog’s rump to make it sit
negative punishment
removing something to decrease behavior
- not giving attention to dog to stop it from jumping
strategies for deciphering reinforcement/punishment
- what is the behavior
- is the behavior increasing or decreasing
- am i applying or taking it away
ex:
- Use of halti lead collar can influence 2 main behaviors: reduce pulling, and increase walking next to you so leash is loose (positive punishment)
- Want to stop puppy from mouthing/biting by walking away when their teeth touch your hands (negative punishment)
Problems with aversive training methods
- negative emotional state
- can cause physical and mental harm
- breakdown of the human-animal bond
reduced social referencing impacts learning and training- Avoiding looking at owner for cues is problematic b/c this behavior is important for learning and training
- Communication = key aspect during training
- only changes the behavioral response, increases negative emotional response = increase chance of aggression
evidence of learning impairments
- Dog performance on a novel learning task was reported to be better for reward vs. positive punishment-based training
- Dogs at training schools using negative reinforcement based methods showed more stress behaviors and those at schools that use positive reinforcement showed more attentiveness
- Increased obedience correlated w/ use of rewards but not positive punishment, and problematic behaviors were related to positive punishment but not rewards
evidence of potential welfare issues
- More stress behaviors in dogs trained w/ shock collars
- Direct reactions to shock suggest pain & more stress behaviors during training & free walking in dogs trained w/ shock collars vs. control dogs
- suggests effects of training using shock collar negatively impact dog even after training during walks w/o shock collar
evidence of potential aggression issues
- Case series of 5 serious dog attacks linked to shock collar training
- Dogs trained w/ confrontational training methods are associated with aggressive responses during vet exams
- Use of physical punishment is associated w/ increases in various forms of aggression in dogs & human directed aggression specifically
problems with positive punishment
- doesn’t tell animal the correct behavior = don’t know how they’re supposed to behave
- must be immediately after the behavior and happen every time
- Animals like gambling - intermittent reward very reinforcing
- Ex: counter walking in cats -> positive punishment won’t work well b/c we always have to be there to train them not to go on counters & perching is a highly motivated behavior
Not always clear cut…
- aversive training tools: prong collars & shock collars
- for some animals, alternative training tools such as halters can cause discomfort & stress
- take an individual approach to decide which tools to use w/ each animal
Unwanted behaviors
- teach an alternative or incompatible behavior
- Ex: dog jumps up when doorbell rings -> train dog to sit on a mat when doorbell rings & give treat
- interrupt & redirect
- Ex: dog jumps up when doorbell rings -> re-direct and distract them as they are beginning to perform behavior, like directing them to sit position
- extinction: breaking association b/twn behavior and the reward by removing reward
- Ex: dog jumping when doorbell rings -> turn away and ignore them and don’t open the door, they’ll learn that jumping is no longer self-rewarding
unintentional learning
- Negative experiences or medical conditions can lead to unintentional learning = can evolve into behavioral problems
- ex: cat stops using litterbox b/c of bad experience
- Can be completely out of our control, but we still need to know how to deal w/ it
3 key factors affecting learning
timing of reinforcement
- the faster the better
- assists w/ making the connection between the behavior and reward
- ex: clicker training
- human error slows down process
- helps communicate w/ dog that reward is coming
- uses classical conditioning
strength of reinforcement
- food is effective for most animals
- balance internal vs. external factors for reward value
- beware of the “breakfast effect”
- train after the hours after feeding
- don’t want animals so hungry they can’t learn
- don’t have to limit food for most pets
- limiting food can be problematic for some => those w/ food aggression issues or food related behavior issues
schedule of reinforcement
- how often treats are delivered
- Initially reward every time to help solidify the association
- Extinction - reduction in behavior when no longer followed by a reinforcer
learned taste aversion
- related to neophobia
- foods that make an animal ill will be avoided in the future
- based on classical conditioning, doesn’t require conscious awareness
- one trial learning
- duration b/twn intake and illness can be delayed
- taste/smell most effective sensory cues in animals
- in birds, visual stimuli effective
- clinical relevance - medications hidden in food make pet sick, may have to switch food
3 training techniques
Luring
- using rewards (food) to physically encourage a particular behavior
- sit: push reward over dog’s head
- down: lower food to ground
- heel: hold reward at hip
targeting
- train to touch a target for rewards
- use target to move animal in desired ways = luring w/ a target stick
shaping
- training a complex behavior by rewarding successive approximations
- Start w/ rewards for behaviors that are close or can be built
- Don’t associate command or cue until behavior is complete
- Often used for behaviors that don’t occur naturally
training tips
- Start w/ reward every time, slowly reduce how often or replace
- Only pair w/ command once animal understands what is expected
- Start training in a quiet, undisturbed area and slowly increase to “real world” settings
- Don’t do too much at once - lots of breaks, multiple short sessions
- Remember anyone can call themselves a “behaviorist” - be careful what you read online
