Pavlovian Learning II

• Standard view of Pavlovian learning

  • The CR is always the same as the UR.

  1. Pairing the US and CS is a necessary condition for acquiring a CR – i.e., they must occur together if acquisition is to occur.
  2. Pairing the US and CS is a sufficient condition for acquiring a CR – i.e., if they occur together, then a CR will (eventually) be acquired.
  3. The US-UR reflex is always a pre-existing, innate reflex; the CS-CR reflex is an acquired reflex
• Not the case
• CS-US pairing necessary?


  If pairing is necessary, then there can be no situations in which a CR is acquired when the CS and US are not paired

  1. Consider a conditioning procedure in which the US is only presented if the CS is not presented (no pairing)
  2. Are there any such situations?
  3. If the CS (e.g., tone) is presented, then no US (e.g., shock) is presented. The US & CS are both presented, but not paired
• CRs can be acquired even when CS and US are not paired
• Unpaired procedure


  If CS then no US, called an inhibitory conditioning procedure
• When US is aversive

  CR = approach to the location of the CS
• When US is appetitive

  CR = avoidance of the location of the CS
• Conclusion: CS-US pairing is NOT necessary
• CS-US pairing sufficient?


  Sufficiency: all that's needed for CR acquisition is that the US and CS are paired (nothing else is needed)

  1. If pairing is sufficient, then a CR will be acquired in all situations in which the CS and US are paired
• Are there any situations in which they are paired, but CRs are not acquired?
• Different CS-US pairings


  Simultaneous conditioning: CS and US presented at the same time

  1. Trace conditioning: CS and US separated by a trace interval
  2. Delay conditioning: CS and US separated by a delay
  3. Backward conditioning: US presented before CS
• Eye blink reflexes


  • The US (puff) elicits a blink with a short latency (< 100 ms)

  1. The US is (usually) an air puff to the eye of the subject, the UR is a blink
  2. In most animals the puff elicits a blink of the 'third eyelid' (nictitating membrane)
• People have no 'blinkable' third eye-lid, but you can see what's left of it
• Rabbits blink their third eyelid in response to air puffs and are the most frequently used experimental animals
• Delay conditioning of eyeblinks


  Many experiments have been conducted on eye blink conditioning with various species using delay and trace conditioning procedures

  1. These involved a tone CS and a various delay intervals
  2. Delays in different experiments ranged from –50 milliseconds (backward conditioning) to several seconds
• The eye blink conditioning results are typical
• If pairing is sufficient, then the only thing that would matter would be the time between stimuli – the longer the delay or trace interval, the less effective the procedure
• It matters how the CS and US are paired: backwards and simultaneous conditioning are ineffective as are positive delays (forward conditioning) that are too short or too long
• Similar results are obtained using the trace conditioning procedure, but trace intervals more than 2 or 3 seconds are ineffective in eye blink conditioning
• Does previous conditioning matter?


  In stage 1 a CS (call it CS-A) is paired with a US using an effective procedure (e.g., delay conditioning)

  1. Stage 2 involves presentation of a compound CS comprising CS-A and another CS (CS-B) presented together (CS-A + CS-B) together with the US used in stage 1
  2. After the training stages, a test is given that consists of CS-B presented alone. The control is stage 2 only and test
• The result of this training procedure is that CS-B fails to elicit a CR at test (or only weakly) in the experimental condition
• Prior experience of the (CS-A)-US relationship, blocks subsequent learning of the (CS-B)-US relationship when CS-A is present (so previous conditioning matters)
• Does previous experience with the CS matter?


  In stage 1 a stimulus is presented on its own (no USs are presented)

  1. Stage 2 involves using the stimulus presented in stage 1 as the CS in an effective conditioning procedure (i.e., paired with a US)
  2. After the training stages, a test is given that consists of CS presented alone. The control group are not pre-exposed to the CS
• CS may fail to elicit a CR at test; if a CR is acquired it is weak and/or very slowly acquired
• Prior exposure to the CS makes it less effective in a subsequent conditioning procedure, previous experience with the CS matters
• Case 2: previous conditioning experience matters – CRs are not acquired to a CS if that CS is presented in combination with 'pre-conditioned' CS
• Case 1: the US must come after the CS if CRs are to be acquired, but not too soon after and not too long after
• Case 3: previous experience with a stimulus matters – if a stimulus is familiar, it is ineffective when used as a CS in a conditioning procedure
• Innate and acquired reflexes


  • The US-UR reflex is a pre-existing, innate reflex; the CS-CR reflex is an acquired reflex
  • It's certainly true in the original experiments of Pavlov, but it is not generally true
  • Many conditioning experiments have been successful
• Group X


  • Result
  • Control group
  • P
• These three cases show that CS-US pairing is not a sufficient condition for CR acquisition
• The final assumption of the standard view of Pavlovian learning and conditioning is: The US-UR reflex is a pre-existing, innate reflex; the CS-CR reflex is an acquired reflex
• It's certainly true in the original experiments of Pavlov, but it is not generally true
• Many conditioning experiments have been successfully conducted (CRs acquired) using stimulus-elicited behaviours that do not fit the definition of reflex
• The idea that the behaviour that mediates the US-UR relationship must be innate can be easily tested
• Two stage experiment


  1. Starting behaviour

  2. Conditioning procedure

  3. Acquired behaviour
• Stage 2 procedures have been found to be effective and are called second order conditioning
• Standard view of Pavlovian learning

  • The CR is always the same as the UR
  • Pairing the US and CS is a necessary condition for acquiring a CR
  • Pairing the US and CS is a sufficient condition for acquiring a CR
  • The US-UR reflex is always a pre-existing, innate reflex; the CS-CR reflex is an acquired reflex
• Not true
• CRs are not always the same as URs