LECTURES 7-12 Flashcards
STIMULUS CONTROL
The extent to which stimuli that precede or accompany operant behaviour come to control the rate or probability of that behaviour
GUTTMAN & KALISH (1956)
Trained pigeons to peck a single S+(580nm light), showed decremental generalization gradients
JENKINS & HARRISON (1960)
Innate or learned -
two groups of pigeons pecking a white key, group 1 on SST, 100Hz tone VI schedule, food available always, tone sounding always
Group 2 on IDDT, 100Hz tone VI schedule, no tone extinction
Group 2 discriminated a lot more than Group 1. Experience is necessary for generalization.
PETERSON (1962)
Early rearing experiments - two groups of ducklings, one raised in a normal colour environment and the other raised in a monochromatic yellow environment. Trained both groups to peck a yellow key. Tested generalization, found normal gradients for the normal group and flat gradients for the monochromatic group. Appears as though prior experience is necessary - behaviour is learned.
GENERAL PROCESS APPROACH TO LEARNING
Behavioural principles of learning are common across all species.
EVOLUTIONARILY PREPARED
Behaviours that make sense with the environment e.g. an association of taste with feeling sick
EVOLUTIONARILY UNPREPARED
Behaviours that do not make sense with the environment e.g. an association of light and feeling sick
HEARST AND KORESKO (1968)
4 groups of pigeons on SST using line orientation. S+ was a vertical white line on a black key with food on a VI. Test done in extinction.
- 2 days training, very flat gradient
- 4 days training steeper
- 7 days training steeper, more responses
- by 14 days the gradient has shifted up the graph and is showing a large peak at the training stimulus, decremental gradient on either side
DIFFERENTIAL REINFORCEMENT
Arranging different reinforcement schedules to different stimuli and different responses
DISCRIMINATION TRAINING
One response is reinforced, the other is in extinction, at a lower rate or punished.
Expect a greater discrimination if S+ and S- are along the same dimension in the gen test
INTRA DIMENSIONAL DISCRIMINATION TRAINING
Where the different stimuli are along the same dimension in the test
INTER DIMENSIONAL DISCRIMINATION TRAINING
Where the different stimuli are not along the same dimension in the gen test
POSITIVE PEAK SHIFT
The most responding in the generalization test does not occur at S+, but instead at another stimulus that occurs on the opposite side of S+ from S-
HANSON (1959)
Control: SST 550nm
Exp: intradimensional training, 550nm S+ and 560nm S-
Both groups received a gen test, control showed large peak at S+, exp group showed peak shift away from S-
ABSOLUTE THEORY OF STIMULUS CONTROL
Relies on the absolute values of S+ and S-, predicts no peak shift, just generalization of wavelengths to both S+ and S- resulting in independent control by both
RELATIONAL THEORY OF STIMULUS CONTROL
The animals are learning something about the relationship between stimuli, i.e. learning not that S+ looks different than S-, but instead that S+ is more green than S-. They are learning a rule
KÖHLER
Chickens - simultaneous choice task. One light grey card (S+) and one dark grey card (S-). Pecking S+ knocks it down revealing food, pecking S- does nothing.
S- was then swapped for an even lighter S++ card. If the chickens were responding according to the absolute theory, they should continue to peck the S+ as this signals reinforcement. Instead they responded according to the relational theory, and pecked S++, indicating they have learnt to ‘peck the lighter card’
SPENCE (1937)
Absolute theory of stimulus control. Proposed that excitatory and inhibitory gradients form around S+ and S- respectively and the observed behaviour is the sum of these gradients. Accounts for peak shift because there is a considerable amount of generalization remaining to S+ but not a lot generalized from S- because it is so far away.
HONIG, BONEAU, BURNSTEIN & PENNYPACKER (1963)
Two groups of pigeons, interdimensional DT, stimuli were black line on a white key vs plain white key. Presence absence training. The groups had swapped S+ and S-.
Training stimulus at 90 deg line, Line S+ group show a peak in responding, decremental gradient, Line S- group show a minimum in responding, incremental gradient.
BRELAND AND BRELAND
Operant techniques - Trained pigs to pick up coins and deposit them in a piggy bank across the room, and chickens to stand on platforms for 15s.
Over time the animals developed different behaviours - chickens danced and pigs rooted/buried the coins. These are species specific, intrinsic behaviours.
INSTINCTIVE DRIFT
Initially reinforcement increased the target behaviour as expected, but eventually more learning about the predictability of food began to evoke species-typical food-getting behaviour. This instinctive behaviour interfered with performing the operant response.
Principles of Operant Conditioning
Shape the response, follow it with a reinforcer, continue reinforcement until the response happens reliably in the presence of the discriminative stimulus
STADDON & SIMMELHAG (1971)
Replicated Skinner (1948), presented pigeons with food every 12 seconds, again the pigeons engaged in repetitive behaviour. Unlike Skinner, the pigeons all ultimately engaged in the same patterns of behaviour.
AUTOSHAPING
Repeatedly pairing a neutral stimulus with an unconditioned stimulus, elicits response without directly hand shaping it
AUTOSHAPING
Repeatedly pairing a neutral stimulus with an unconditioned stimulus, elicits response without directly hand shaping it
PALYA & ZACMY (1980)
Showed that pigeons spot pecking happens directly after presentation of food - foraging that happens in the natural environment. Changing the time of feeding subsequently changes the time of spot pecking to match.
REINSTATEMENT
First the response produced the food now the food produces the response
EQUIPOTENTIALITY
Getting the same learning whatever response, stimulus or reinforcer is used.
LAWICKA (1964)
Tone location easily learned, tone frequency not so easily learned.
RANDALL & ZENTALL (1997)
Showed pigeons have a win-stay bias - they will repeat the behaviour that just received the reinforcer.
RANDALL & ZENTALL (1997)
Showed pigeons have a win-stay bias - they will repeat the behaviour that just received the reinforcer.
HERRNSTEIN (1961)
Herrnstein’s Hyperbola - prediction about the rate of responding per minute depending on the rate of reinforcers.
RATES
How frequently is one type of activity being engaged in. Not choice, but occurs as a result of choice.
B = kR/R + Re
Response rate = (maximum response rate x reinforcer rate) / (reinforcer rate + extraneous reinforcer rate)
B and R: variables
k and Re: parameters
STRICT MATCHING
Proportion of responses matches the proportion of reinforcers obtained for making that response
CONGER & KILLEEN (1974)
Students and confederates - one gave plenty of social reinforcement and the other very little. Students were found to spend more time talking to the former, i.e. allocation of response depended on the proportion of reinforcers.
VOLLMER & BOURRET(2000)
Basketball shots - 3 point attempts vs 2 point attempts. Proportion of attempts matches the proportion of successful shots made.
VOLLMER & BOURRET(2000)
Basketball shots - 3 point attempts vs 2 point attempts. Proportion of attempts matches the proportion of successful shots made.
DOUGHERTY & LEWIS (1992)
3 exp naive horses. Measured lever presses, changing proportion of reinforcers to lever 1 changes the response to that lever.
CATANIA & REYNOLDS (1968)
6 pigeons, VI schedules, rates from 10 - 300 foods per hour. VAC - variance accounted for, higher the value the more accurate. Looked at how closely the observed rate of responding matched the hyperbolic prediction.
CATANIA & REYNOLDS (1968B)
3 male pigeons, concurrent schedules. Reinforcer rate on key 1 affects responding on key 1 and 2 (choice).
