Chapter 10

Stimulus Control

• Relationship between antecedent stimulus and behaviours that follow it
• Generalization
• Discrimination
• Gradients

Measuring Generalization Gradients

• Train on one stimulus, test on others
• Generalization?
• Tricky to test
• Reinforcement problem
• Extinction problem
• Successive discrimination task

Probe Trials

• Insert occasional unreinforced test stimulus

(Figure 1)

Extinction Blocks

• Train stimulus to asymptote
• Blocks of extinction trials
• Each stimulus presented once/block
• Extinction constant across stimuli

(Figure 2)

Generalization Gradient

(Figure 3)

Cause of Generalization Gradient?

• Pavlov
• Inherent property of nervous system
• Innate
• Lashley and Wade (1946)
• Explicit discrimination training required
• Learned

How Experience Affects the Generalization Gradient

• Nondifferential training
• Every trial the same
• Presence-absence training
• Two kinds of trials
• Reinforced and non-reinforced
• Discrimination

Intradimensional Training

• Discrimination training
• S+ and S- from same stimulus continuum
• Produces more focused generalization gradient

(Figure 4)

Absolute or Relative Stimulus Control?

• Stimulus discrimination
• Absolute theory:
• Learn individual stimuli
• Relational theory:
• Learn relationship between stimuli

Example

(Figure 5)

Transposition

• Transfer relational rule to new stimuli set
• Kohler (1939)

(Figure 6)

Absolute vs. Relational Theories

• Absolute theories
• Successive discrimination tasks
• Relational theories
• Simultaneous discrimination tasks
• Animals do whatever is easiest

Peak Shift

• Change in generalization gradient
• Peak level of responding
• 1. Shift in peak level of responding away from S+ in direction opposite S-
• 2. The closer S+ and S-, the greater the difference in responding at peak and S+

(Figures 7 and 8)

• Both shift left
• Differences in responding at peak and S+

Spence’s Theory of Peak Shift

• Opponent process theory
• Excitatory and inhibitory gradients
• S+: excitatory
• S-: inhibitory
• Net sum
• Direction of shift

Peak Shift: Shift Away

(Figure 9)

Role of S-

(Figure 10)

Support for Spence’s Theory?

• Honig et al. (1963)
• Excitatory and inhibitory gradients

(Figure 11)

Behavioural Contrast

• Reinforcement contingencies for one stimulus can affect response to other stimuli
• Positive
• Responding increases to unchanged stimulus
• Negative
• Responding decreases to unchanged stimulus

Gutman (1977)

• Rats, multiple schedule
• Positive contrast

(Figure 12)

Errorless Discrimination

• To develop discrimination between S+ and S-
• Goals
• Respond strongly to S+
• Respond very little/not at all to S-
• Strong stimulus control in minimum time
• Factors
• Introduce S- early
• Fading
• e.g., Initially, S- shown very briefly; gradually lengthen exposure

Advantages of Errorless Discrimination

• Decreases overall responding to S-
• Improves long-term discrimination
• Less
• Aggressive/frustration behaviour when S- presented
• Behavioural contrast
• Peak shift

Concept Formation

• Stimuli from natural environment
• Train/test many stimuli
• Positive and negative instances

Natural Categories

• Classification of stimuli into groups
• Central and peripheral characteristics
• Gradated scale of belongingness
• Herrnstein (1979)
• Pigeons, 80 pictures
• Tree/no tree = positive/negative instances
• Learn discrimination easily
• Generalization test
• Supports concept formation, not memorization