instrumental conditioning

• instrumental conditioning


  learned contingency between behavior and consequence
• law of effect

  • behavior with positive consequence is stamped in/used a lot
  • behavior with negative consequence is stamped out
• reinforcer


  stimulus is presented after response that impacts frequency that responses are performed
• presentation of positive reinforcement (reward training)


  increases behavior
• removal of negative reinforcement (escape training)


  increase behavior
• presentation of negative reinforcement (punishment training)


  decrease behavior
• removal of positive reinforcement (omission/time-out training)


  decrease behavior
• auto-shaping

  • gradual modification of behavior by rewarding particular response
  • only works with simple behaviors without external guidance
  • shapes by successive approximation to desired behavior
  • rewards successive steps
• chaining

  • used to develop sequence of behavior
  • each behavior is reinforced with opportunities to perform next behavior in sequence
  • reinforces behavior as long as it's performed in the defined order
  • behavior and order are set prior to training
• discrimination stimulus (SD or S+)


  signal when the contingency between a particular behavior and reinforcement is (un)valid
• S[delta] or S-


  indicated when contingency relationship is not valid (not rewarded)
• continuous reinforcement (CRF)

  • response elicits reinforcement at every trial
  • rare
• partial reinforcement (PRF)

  • more common
  • fixed ratio, fixed interval, variable ratio, variable interval
• ratio

  • per number of responses
  • based on number of responses made by subject
• interval

  • time
  • based on amount of time since the last response was reinforced
• fixed

  • constant
  • conditions are consistent across trials
• variable

  • random (ratio/interval schedule)
  • rewards follow variable amount work/time
• fixed ratio (FR-#)

  • pause and run patter
  • procrastination during pause
• variable ratio (VR-#)


  more/same reward for less work creates more motivation for responses
• fixed interval (FI-#)

  • scallop pattern
  • after reinforcement there is a low period when responses stop then slowly pick up again
  • responses peak just before reinforcement
• variable interval (VI-#)


  steady rate to make sure not to miss opportunity for reinforcements