Learning & Motiv Week 2 & 3

Schedules of reinforcement

how and when behavioral consequences will be presented

Continuous reinforcement

every single time the animal carries out the behavior it is reinforced

Partial Reinforcement

reinforcement occurs intermittently but it is response-dependent

Steady-state performance

when the pattern of responding on a schedule becomes consistent over time

Ratio schedule

reinforcement is delivered following a set number of responses

Interval Schedule

reinforcement is delivered to the first response after a certain amount of time has passed

Fixed schedule (FR or FI)

reinforcer is delivered after same ratio or interval every time

Variable schedule

reinforcer is delivered after a variable ratio or interval

What shape of responding do FR schedules produce

“pause-and-run” responding

What kind of responding do VR and VI schedules produce?

steady rate of responding, no pauses occur after reingorcement

What kind of responding do FI schedules produce?

scalloping

Negatively Correlated Cumulative Records

records in which slopes decrease over time

Positively Correlated Cumulative Records

records in which slopes increase over time

Intrinsic Reward

one with a natural relation to the responses that produce it (ex music is an intrinsic consequence of playing the piano)

Extrinsic Reward

one with arbitrary relation to the responses (ex: music teacher’s praise is an extrinsic consequence of playing piano)

Sensory Reinforcement

when sensory systems activate to reinforce a behavior

Latent learning

learning without immediate reinforcement

Extinction vs free reinforcememt

response decreases with both but continues under free reinforcement

Superstition

when a response is reinforced accidentally during free reinforcement, leading to an accidental pairing

Operant

a class of responses created by its consequences

Response classes

groups of similar responses that have the same outcome

Differential reinforcement

the reinforcement of only responses falling within a specific class

Two aspects to goal-direction

1. the action must be performed because it is causal with respect to its consequences

2. the consequences of an action must be shown to be a goal for the animal

Contingency criterion

Performance reflects the causal relationship between the action and outcome

What is a yoked control group

a control group that gets the same number and rate of reinforcement but without any response dependency

Contiguity vs contingency

the number and degree of pairings determines learning vs if-then causal statement

Probability of an outcome given a response (formula)

p(O|A) = Number of O’s during the A/number of As

Probability of an outcome given no response

p(O|no A) = Number of O’s in the absence of the A/number of no-A’s

Strength of the action-outcome association (formula)

p(O|A) - p(O|noA)

Motor Competition

when an organism is too busy with a competing task to respond

How do we test between contingency and motor competition

design an experiment with two behaviors, one reinforced and one not, if both responses fall away, it is motor competition, if only the response that is not reinforced falls away, it is contingency

Devaluation hypotheses

the signaling hypothesis, evaluative hypothesis

Devaluation: signaling hypothesis

the outcome signals that something noxious is coming like a CS signals an aversive US

Devaluation: the evaluative hypothesis

the emotional significance of the outcome itself has changed (sugar water is no longer pleasant it is unpleasant)

Which devaluation hypothesis is supported and why

the evaluative hypothesis because rats tend to try the aversive stimulus after devaluation to learn it is no longer reinforcing before ceasing responding

Incentive learning

allows animals to establish the desirability of various events or commodities

Conditioned reinforcer

a reinforcer that is learned and is not evolutionarily beneficial to the organism (ex money)

Limited hold

a schedule where a reinforcer is available only for a designated time so if no response occurs soon enough the reinforcer is lost

Differential-Reinforcement-of-Low-rate (DRL)

arranges reinforcers for responses preceded by some minimum time without responding

Differential-Reinforcement-of-High-rate (DRH)

usually specified by the number of responses required within a time window