PBSI 340 Final Exam

Extinction

Reduction in behavioral responding after non-reinforced trials

Is extinction forgetting?

No; the original association persists and can return

Spontaneous recovery

A temporary increase in responding after a rest period following extinction

Renewal

Increased responding when extinction occurs in a different context than testing

Reinstatement

Exposure to the US alone restores responding to the CS

Key takeaway of extinction

Extinction does not erase the original learning; memory persists

Partial Reinforcement Extinction Effect (PREE)

Behaviors learned with partial reinforcement are more resistant to extinction

What does PREE challenge?

The idea that more reinforcement always produces stronger learning

Frustration theory

Non-reward produces frustration that becomes part of learning and increases persistence

Discrimination hypothesis

Animals fail to detect the shift from reinforcement to extinction

Does more reinforcement always increase learning?

No; PREE shows this relationship is complex

Is resistance to extinction a good measure of learning strength?

No; it can be misleading

Consolidation

Stabilization of memory requiring protein synthesis

Reconsolidation

Updating memory after retrieval

Immediate extinction deficit

Extinction is less effective when done immediately after acquisition

Mechanisms of extinction

Pavlovian and instrumental

Transitive inferential reasoning

Using learned relationships to infer new ones (A>B, B>C → A>C)

What does transitive reasoning rely on?

Multiple learned associations

Gillan & Premack experiment

Chimps trained on A>B, B>C, C>D, D>E then tested on B vs D

Chimps performance on transitive reasoning

~89% correct choices

Analogical reasoning

Understanding relationships between relationships

Perceptual analogies

Simple analogies based on shape, color, or size

Conceptual analogies

Abstract, complex analogies

Chimps performance on perceptual analogies

~85% correct

Relative brain size

More important than absolute brain size for cognition

Is language unique to humans?

Debated

Behaviorist view of language

Language is learned through stimulus-response principles

Linguist view of language

Conditioning alone cannot explain language

Why can't chimps speak?

They lack the appropriate vocal apparatus

Alternative to spoken language in chimps

Sign language

Washoe

Chimp taught American Sign Language (ASL)

Washoe vocabulary size

~350 signs

Washoe communication ability

Up to 5-word combinations

Washoe example phrase

"METAL CUP DRINK"

Loulis

Chimp that learned ASL without direct human teaching

Koko

Gorilla famous for learning sign language

Koko notable behavior

Requested and named a pet cat ("All Ball")

Lexicon

The full set of words in a language

Phoneme

The smallest sound unit of language

Grapheme

The written representation of a phoneme

Morpheme

The smallest unit of language that contains meaning

Grammar

The rules for using a language

Semantics

The meaning of words and sentences

Syntax

The rules for organizing words into sentences

Are phonemes the same as syllables?

No

Number of phonemes in English

44 total (25 consonant, 19 vowel)

Number of phonemes in Spanish

~24-29 depending on dialect

Diphthongs

Pairs of vowels that form a unique sound

Morphemes examples

Whole words or parts of words (prefixes, suffixes, roots)

What can morphemes encode?

Grammar (tense, plurality, etc.)

Critical period

Time early in life when language is best learned

Genie case study

Child deprived of language until age 13

Genie findings

Learned vocabulary but struggled with grammar

Implication of Genie

Supports a critical period for language development

Artificial intelligence (AI)

Systems that recognize stimuli, understand relationships, and produce intelligent outputs

Machine learning

Algorithms that improve performance with experience

Connectionist modeling

The mind modeled as an associative network

What does connectionist modeling mimic?

Brain learning via synaptic changes (LTP/LTD)

What happens at a synapse?

Neurotransmitter release causes changes in the postsynaptic neuron

What determines synaptic strength?

Amount of neurotransmitter and receptor properties

Long-term potentiation (LTP)

Strengthening of synaptic connections

Long-term depression (LTD)

Weakening of synaptic connections

Two ways neural networks adjust weights

Mutation and backpropagation

Backpropagation

Adjusts weights based on prediction error

Extinction importance (cognition)

Updating previous information

Extinction importance (clinical)

Used to reduce fear, anxiety, and trauma (exposure therapy)

Instrumental (R-O) Learning

Behavior changes as a function of its consequences; the response produces an outcome

Operant Learning

A subset of instrumental learning where the organism actively manipulates the environment to obtain outcomes

Discrete-trial method

Only one opportunity to respond per trial

Free-operant method

Continuous opportunity to respond at any time

Reward

Addition of an appetitive stimulus increases behavior (positive reinforcement)

Punishment

Addition of an aversive stimulus decreases behavior (positive punishment)

Omission learning

Removal of an appetitive stimulus decreases behavior (negative punishment)

Escape learning

Behavior terminates an aversive stimulus after it begins

Avoidance learning

Behavior prevents an aversive stimulus before it occurs (often signaled by a CS)

Contiguity

Temporal closeness between response and outcome; determines which behavior is associated with the outcome among many ongoing behaviors

Secondary reinforcer

A neutral stimulus that becomes reinforcing through association with a primary reinforcer

Marking stimulus

A cue that highlights which response is being reinforced despite delays

Reinforcement contingency

The degree to which an outcome depends on a response; stronger contingency strengthens learning

Goal-directed behavior

Behavior controlled by knowledge of the R-O relationship and outcome value; decreases after devaluation

Habitual behavior

Behavior controlled by stimulus-response (S-R) associations, not outcome value; persists after devaluation

Outcome devaluation test

Train R-O → devalue outcome → test responding to determine if behavior is goal-directed or habitual

Ratio schedule

Reinforcement depends on number of responses; strong R-O contingency produces high response rates

Interval schedule

Reinforcement depends on time since last reinforcement; weaker R-O contingency produces lower, steady response rates

Ratio feedback function

Response rate increases continuously with reinforcement

Interval feedback function

Response rate reaches an asymptote

Feedback function

Relationship between rate of responding and rate of reinforcement under a schedule

Concurrent schedules

Multiple response options available simultaneously, each with its own reinforcement schedule

Matching law

When effort and switching costs are equal, response ratio matches reinforcement ratio across options

Generalized matching law

Accounts for bias, effort, and switching difficulty when matching law assumptions are violated

Concurrent-chain schedules

Initial choice locks organism into a specific reinforcement schedule for the trial

Drive-reduction theory

Reinforcement reduces biological drives (limited explanatory power)

Problem with drive-reduction theory

Learning can occur without drive reduction (e.g., saccharin consumption)

Premack principle

A more probable behavior reinforces a less probable behavior by providing access to it

Problem with Premack principle

Difficult to quantify behavior probabilities; not always predictive

Response deprivation theory

A behavior becomes reinforcing when access to it is restricted below its baseline level

Learned helplessness

Reduced responding after exposure to uncontrollable aversive events

Learned helplessness deficits

Motivational and associative deficits

Reducing learned helplessness

Behavioral therapy and prior controllable experience restore responding

Discriminative stimulus (S^D)

A cue signaling that a response will produce an outcome; in its absence, the response produces no outcome