BLP Exam 3 - Instrumental Learning

R-O associations

Rescorla and Colwill performed experiment in which they changed the ‘value’ of the outcome after learning and saw how responding changed

R-O association experiment

Phase I: R1 -> O1; R2 -> O2…Phase II: O1 -> LiCl (illness)...Test: R2

Results show that rats associate R1 with O1, because once O1 makes them ill, they stop.

S-O associations

Rescorla and Colwill performed experiment in which they changed R-O relationship depending on the stimulus that was present

S-O association experiment

Phase 1: S1: Rn -> O1; S2: Rn -> O2…Phase 2: R1 -> O1; R2 -> O2

Results: When in the presence of S1, rats performed R1; when in presence of S2, rats performed R2

hierarchical associations

made animals learn what R-O relationship is in effect during the signal stimulus

hierarchical association experiment

Phase I: S1: R1 -> O1; R2 -> O2; S2: R1 -> O2; R2 -> O1…Phase II: O2 -> LiCl

Results: During S1, rats performed R1 more frequently, in S2, rats performed R2 more frequently

Significance: Rats can learn which R -> Outcome relationship is in place during stimulus cue

concurrent schedules

two schedules are in effect at the same time, each one is reinforced according to its own schedule; the participant is free to switch from one to the other at any time

examples of concurrent schedules

slot machines at casinos, remote control for TV (choose which channel; reinforcement set by program itself), talking to people at a party (choose who to talk to; reinforcement is set by how interesting)

relative rate of responding

rate of responding on choice A / rate of responding on all choices

relative rate of reinforcement

rate of reinforcement on choice A / rate of reinforcement on all choices

Herrnstein’s Matching Law

relative rate of responding = relative rate of reinforcement

the relative rate of responding on an alternative matched the relative rate of reinforcement on that alternative

significance of Herrnstein’s Matching Law

whether a behavior occurs frequently or infrequently depends not only on its own schedule of reinforcement, but also on the rates of reinforcement of other activities the individual may perform

generalized form of matching law

incorporated two more variables to account for “undermatching” (s and b)

s

varying sensitivities to the change for the reinforcement schedule (might not notice/understand differences between two)

example of s

don’t realize that earning extra credit in one way is more difficult than the alternative

b

response bias, we prefer some activities over others, even when reinforcement schedule is lower

example of b

may prefer to work on spreadsheets over presentations even though you might receive more praise for your presentations

stimulus control

cues that are present during operant learning will begin to signal (or control) when it is (or is not) appropriate to make the operant response

example of stimulus control

bedroom (context cue) - undress (response) -> appropriate; public (context cue) - undress (response) -> inappropriate

forms of stimulus control

• intradimensional discrimination

• interoceptive stimuli

• configural stimuli

• contextual cues

intradimensional discrimination

red light v. green light

interoceptive stimuli

internal sensations of drug withdrawal serve as a signal to engage in drug seeking behaviors

configural stimuli

figure or “whole” image (not just a light or tone) signals when it is appropriate to make the response

contextual cues

learn to navigate about town using landmarks

typical stimuli control experiment steps

training phase: red circle with white triangle -> peck -> food

test: red circle alone (do they peck? some); white triangle alone (do they peck? some)

typical stimulus control experiment significance

the degree of differential responding (responding differently to each stimulus) tells us the degree of control that stimulus has over the behavior

stimulus generalization

an organism responds in a similar fashion to two or more stimuli, this is the opposite of stimulus discrimination and/or differential responding (more generalization = less stimulus control)

stimulus generalization gradient

steep - more stimulus control (less generalization), flat - more stimulus generalization (less control)

example of stimulus generalization gradient

A pigeon is trained to peck a red light for food. If that pigeon is color-blind (he cannot distinguish one color from another), then he will peck at all lights, no matter what color (flat stimulus generalization gradient). If the pigeon is not color-blind, then he will only peck at the red light and maybe some colors that are close to red.

factors that affect stimulus control

• sensory capacity and orientation

• ease of conditioning

sensory capacity and orientation

physical limits of what our sensory systems can perceive

ease of conditioning

some stimuli are easier to notice, identify, encode, and remember (overshadowing)

types of reinforcers

visual signals → responses that lead to appetitive outcomes

auditory signals → responses that lead to removal of aversive outcomes

stimulus elements

sometimes we perceive individual elements that make up the cue, sometimes cue as a whole (this affects how we generalize learning)

discrimination training (learning)

experience with stimuli (learning about them) may determine the extent to which those stimuli come to control behavior

S+

stimulus that signals reinforcement is available

S-

stimulus that signals reinforcement is not available

example of discrimination training

traffic lights: S+ = green light; S- = red light

will learn discrimination faster if…

S+ and S- are presented simultaneously

will have greater stimulus control to S+ if…

S+ and S- are close in similarity

example of how we use interoceptive cues as S+ and S-

hunger → S+ = hunger (you eat); S- = stomach full (you don’t eat)