Week 3 Lecture: Classical Conditioning II

Factors Influencing Conditioning Effectiveness

Contiguity

• Contiguity refers to the temporal proximity of the CS and US.

• Example: "When Thunder Roars, Go Indoors!" - illustrating a clear temporal relationship.

CS Pre-exposure

• Group Ctrl: X leads to CR (Conditioned Response).

• Group Exp: X leads to no CR (weaker conditioned response).

• CS pre-exposure reduces attention to the CS, resulting in weaker responding at test.

• The table represents the contingency:

  • Outcome (US) Present: a

  • Outcome Absent: c

  • CS Present: b

  • CS Absent: d

• Before, During, After: x x

• Contingency (Necessity). Better contingency=better conditioning

US Pre-exposure

• Group Ctrl: X leads to CR.

• Group Exp: X leads to no CR (weakened conditioned response).

• US pre-exposure habituates the subject to the US, weakening responding to the CS at test.

• The table represents the contingency:

  • Outcome (US) Present: a

  • Outcome Absent: c

  • CS Present: b

  • CS Absent: d

• Before, During, After: x x

• Contingency (Necessity).

Partial Reinforcement

• Group Ctrl: X leads to CR.

• Group Exp: X leads to no CR.

• Note: The total number of US presentations must be equal for both groups.

• Group Exp experiences partial extinction on CS- trials, resulting in weaker responding to the CS at test.

• The table represents the contingency:

  • Outcome (US) Present: a

  • Outcome Absent: c

  • CS Present: b

  • CS Absent: d

• Before, During, After: x x

• Contingency (Necessity).

Degraded Contingency

• Group Ctrl: X leads to CR.

• Group Exp: X leads to no CR.

• Note: The total number of US presentations must be equal for both groups.

• Group Exp habituates to the US, weakening responding to the CS at test.

• The table represents the contingency:

  • Outcome (US) Present: a

  • Outcome Absent: c

  • CS Present: b

  • CS Absent: d

• Before, During, After: x x

• Contingency (Necessity).

CS Post-exposure

• Group Ctrl: X leads to CR.

• Group Exp: X leads to no CR.

• Group Exp undergoes procedural extinction, leading to weaker responding to the CS at test.

• The table represents the contingency:

  • Outcome (US) Present: a

  • Outcome Absent: c

  • CS Present: b

  • CS Absent: d

• Before, During, After: x x

• Contingency (Necessity).

US Post-exposure

• Group Ctrl: X leads to CR.

• Group Exp: X leads to no CR.

• Group Exp experiences retroactive interference (repeated presentations of US alone interferes with original learned CS-US) = weakened responding to the CS at test.

• The table represents the contingency:

  • Outcome (US) Present: a

  • Outcome Absent: c

  • CS Present: b

  • CS Absent: d

• Before, During, After: x x

• Contingency (Necessity of one for the other).

Overshadowing

• Demonstrates reduced responding despite perfect contiguity and contingency with US.

• Group OV: AX+ leads to X elicits little cr.

• Group Ctrl: X+ leads to X elicits CR.

• Salience: Intensity, physical characteristics, biological relevance, naturalistic of stimuli.

Blocking (learning is being blocked)

• Challenges the notion of contiguity alone being necessary and sufficient for learning.

• Group Exp: Phase 1 (A+), Phase 2 (AX+), Test (X leads to little cr). X is seen as a neutral stimulus as A overshadows or blocks x when put together

• Group Acq Ctrl: Phase 1 (A+), Phase 2 (X+), Test (X leads to CR).

• Surprisingness: Is there any new information to learn about?

Latent Inhibition (CS-pre-exposure) and US pre-exposure

• Latent inhibition is context specific.

• Group LI: Pre-exposure (X-), Acquisition (X+), Test (X leads to little cr).

• Group Ctrl: Pre-exposure (A-), Acquisition (X+), Test (X leads to CR).

• Group US-Pre: Pre-exposure (+), Acquisition (X+), Test (X leads to little cr).

• Group Ctrl: Pre-exposure (-), Acquisition (X+), Test (X leads to CR).

• Group LI-Diff: Pre-exposure ((X-)1), Acquisition ((X+)2), Test ((X)2 leads to CR).

• Group LI-Same: Pre-exposure ((X-)2), Acquisition ((X+)2), Test ((X)2 leads to little cr).

• Novelty.

Latent Inhibition (CS-preexposure) and US Preexposure

• Not truly inhibitory.

• The CS does not pass negative summation test (but does pass retardation test).

• Attention deficit explanation.

• Protective benefit from irrelevant stimuli and later acquisition of phobias.

• Group NegSum: Preexposure (X-), Acquisition (B+), Test (XB leads to CR).

• Group NegSumCtrl: Preexposure (X-), Acquisition (B+), Test (B leads to CR).

• Group Ret: Preexposure (X-), Acquisition (X+), Test (X leads to little cr).

• Group RetCtrl: Preexposure (A-), Acquisition (X+), Test (X leads to CR).

• Novelty

Garcia and Koelling’s bright and noisy (and flavored) water experiment

• Phobias and fear conditioning.

• Evolutionarily based genetic predispositions.

• Conditioning Test:

  • Flavor + audiovisual -----> shock: Audiovisual yields greater CR.

  • Flavor+ audiovisual -----> sickness: Flavor yields greater CR.

• CS-US Relevance or Belongingness.

Factors Influencing the Nature of the CR

The US Identity

• Pavlov’s Stimulus Substitution Model described by Preexisting neural connections, pathway established by conditioning and response pathway.

The CS Identity

• Timberlake & Grant (1975):

• Does not fit with stimulus-substitution model.

CS-US Interval

• The time between the CS and US presentation affects the strength and type of CR.

Higher-Order Conditioning

Higher-Order Conditioning

• Learning without a direct CS-US experience.

• Second-order conditioning:

  • Metronome (CS) followed by Black square (Second-order stimulus).

  • Black square (Second-order stimulus) followed by Food (US).

  • Metronome (CS) elicits Salivation (CR).

  • Black square (Second-order stimulus) elicits Salivation (CR).

Second Order Conditioning

• Group SOC: First Order Cond (A+), Second Order Cond (X- à A-), Test (X leads to CR).

• Group Ctrl: First Order Cond (A+), Second Order Cond (X- / A-), Test (X leads to little cr).

What causes responding to the second order conditioning (SOC) CS?

• Stimulus-Response (S-R): response is elicited directly by the CS; reflexive learning (response occurs without needing a US to be present).

• Stimulus-Stimulus (S-S): response is elicited by the activation of the US representation (response occurs in anticipation of an outcome or US).

• Differentiate between learning mechanisms by US devaluation.

Holland & Rescorla (1975)

• Group Ctrl: Phase 1 (T à Food), Rotation, Test 1 (T à CR), Phase 6 (--), Test 2 (T à CR)

• Group Exp: Phase 1 (Food à Rotation), T à little cr, Food alone, T à CR

• Weakened the contingency.

• US Devaluation.

Holland & Rescorla (1975)

• Group Ctrl: Phase 1 (T à Food), Food deprived, Test 1 (T à CR), Phase 3 (Food deprived), Test 2 (T à CR)

• Group Exp: Phase 1(Satiated ), Test 1 (T à little cr), Test 2 (T à CR)

• Made food valued again.

• US Devaluation.

Holland & Rescorla (1975)

• Group Phase 1 Phase 2 Test 1 Phase 3 Test 2

• Ctrl T à Food Rotation T à CR -- T à CR

• Exp Food à Rotation T à little cr Food alone T à CR

• Group Phase 1 Phase 2 Test 1 Phase 3 Test 2

• Ctrl T à Food Food deprived T à CR Food deprived T à CR

• Exp Satiated T à little cr T à CR

• Evidence of S-S learning. But not all instances of CC involve S-S learning. There are some cases of S-R learning.

Returning to SOC, what causes responding to the second order conditioned stimulus that was never directly paired with the US?

• Chaining of two S-S associations:

  • SOC CS -> FOC CS -> US -> CR (S-S, S-S)

• Single S-S association:

  • SOC CS -> FOC CS -> US -> CR

• Single S-R association:

  • SOC CS -> FOC CS -> US -> CR

• Extinction US devaluation

  • SOC CS -> FOC CS -> US -> CR

Two papers by Holland & Rescorla (1975) investigated extinction and US habituation in SOC.

• Extinguishing the first-order stimulus had no effect on responding to the second-order stimulus.

• Group Exp: Phase 1 (L+), Phase 2 (T à L), Phase 3 (L-), Test (T à CR).

• Group Ctrl: Phase 1 (--), Phase 2 (T à CR).

Second-order conditioning

• Second-order conditioning was observed early in training, but after many nonreinforced compound presentations, conditioned inhibition was observed.

  • Group Exp Phase 1 L+, Phase 2 T à L

  • Group Ctrl --

US devaluation had no effect on responding to the second-order stimulus.

• Group 1-Ctrl: Phase 1 (L à Food), Phase 2 (T à Food), Rotation, Test (T à CR).

• Group 1-Exp: Phase 1 Food à Rotation, Test (T à little cr).

• Group 2-Ctrl: Phase 1 (T à L), Rotation, Test (T à CR).

• Group 2-Exp: Phase 1 Food à Rotation, Test (T à CR).

SOC is the result of a simple S-R association.

• Group SOC: First Order Cond (A+), Second Order Cond (X à A), Test (X à CR).

• Group Ctrl: First Order Cond (A+), Second Order Cond (X / A), Test (X à little cr).

Sensory Preconditioning

• Same as SOC but Phases 1 and 2 are switched.

• Second-order stimulus: Black Square

• Conditioned stimulus (CS): Metronome

• Unconditioned stimulus (US): Food

• Sensory preconditioning: Black square is followed by Metronome.

• Metronome is followed by Food which elicits Salivation (CR).

• Black square elicits Salivation (CR).

Rizley & Rescorla (1972)

• Showed that extinction of the first-order stimulus does affect responding to the second-ordered stimulus.

• This suggests that sensory preconditioning (SPC) works through S-S associations rather than an S-R association, which means that these two seemingly similar procedures operate through different learning mechanisms.

• SPC and SOC offer potential reason for likes and dislikes with no apparent reason.

• Group SPC: First Order Cond (X- à A-), Second Order Cond (A+), Test (X à CR).

• Group Ctrl: First Order Cond (X- / A-), Second Order Cond (A+), Test (X à little cr).

Control by Conditional Relations

Conditional Relations

• Involve a third event, a modulator, to modulate the validity of an association between two events.

• In Classical Conditioning, modulation also called occasion setter/occasion setting.

Occasion Setting

• Refers to the ability of an event (occasion setter) to modulate the association between a CS-US pair.

• Conditioned Stimulus (CS)

• Unconditioned Stimulus (US)

• Occasion Setter

Occasion Setting

• Positive occasion setting (OS): A--X+ / X-

Occasion Setting

• Negative OS: A-X-/X+

Control by Conditional Relations

• OS are not simply conditioned stimuli.

• OS do not necessarily have direct (excitatory or inhibitory) associations with the US.

• OS can have direct association with the US independent of its modulatory properties.

• The direct conditioned association can be of opposite valence of its modulatory potential – orthogonality.

Stimulus Control by Contextual Cues

Discrete cues

• Are stimuli that have a clear beginning and end and can be easily characterized.

• Examples: tones, lights.

Contextual cues

• Are diffuse with no clear beginning or end.

• Examples: visual, auditory, or olfactory features of a room or place where the discrete discriminative stimuli are presented.

• Conditioning contextual cues does not refer to one characteristic in particular, but rather the overall composition and arrangement of the features of a physical space.

Stimulus Control by Contextual Cues

• Contextual cues are often critical in determining whether or not to produce an instrumental or Pavlovian conditioned response.

• Contextual cues can control behavior in the same way as discrete cues.

Stimulus Control by Contextual Cues

• Contextual cues can become conditioned excitors.

• Common in drug-conditioned place preference studies in which the contextual cues form a direct association with the US.

Stimulus Control by Contextual Cues

• Contextual cues can become conditioned inhibitors.

• Polack, Laborda, & Miller trained Context B as a conditioned inhibitor:

• (X+)D | (Z+)A | (Z-)B

• Test 1 negative summation in ext ctx, (X)B versus in neutral ctx, (X)C

• Test 2 retardation of behavioral control in ext ctx (+)B versus neutral ctx (+)C

Stimulus Control by Contextual Cues

• Contexts can serve as conditioned excitors and conditioned inhibitors that can pass retardation and summation tests.

• Contextual cues do not have to directly signal reinforcement/non-reinforcement to gain behavioral control.

• They can also control behaviour through occasion setting, just like a discrete cue.

Summary

• Variables important for training efficacy include contiguity, contingency, salience, novelty, informative value, evolutionary predispositions.

• Variables important for determining behavior include the CS, the US, and the CS-US interval.

• Higher-order conditioning (SOC, SPC) will also affect behaviour.

• Higher-order conditioning can also take the form of modulation or occasion setting in which a third stimulus (or context) modulates the associative validity between a CS and the US.