S-S or S-R associations?

US -> UR	or	CS -> US -> UR
      ^
      |
      CS


response-prevention paradigm
  CS buzzer, US shock, UR leg lifting
  prevent response by temporarily damaging spinal motor nerves
  CR is displayed after recovery of nerves
  this is weak evidence for the S-S idea
  objection: S was associated to representation of R in CNS

US devaluation
  CS light, US food, UR increased activity
  satiate rats -> US doesn't evoke UR
  CS doesn't either
  evidence for S-S association

sensory preconditioning
  CS2 CS1, then CS1 US
  CS2 evokes CR
  shows two stimuli can be associated

second-order conditioning
  reverse order of preconditioning
  devalue US by satiating rats
    response to CS1 is extinguished
    response to CS2 is not extinguished!
  indicates that 
    second-order associations are S-R
    first-order are S-S

whether associations are S-S or S-R depends on whether the stimulus of
  response is more salient

what is the conditioned stimulus?
  stimulus generalization
  discrimination learning

what is the conditioned response?
  CR is the UR
  CR is preparation for the US, not a response
    blink is preparation for puff
    US electric shock
      UR is heart-rate acceleration
      CR is heart-rate deceleration
      CR is opposite of UR
      relaxation may reduce perceived magnitude of pain
    rats: US electric shock
      UR increased activity
      CR freezing
    natural situation
      CS sight of predator, CR freezing
      US predator attacks, UR flight

contiguity or contingency?
  shock with and without tone, measure bar pressing
  vary P(shock|tone) and P(shock|no tone)
  contingency viewpoint: conditioning only if P(shock|tone) > P(shock|no tone)
  tone: freeze (CER), stop bar pressing
  suppression = (R(no tone)-R(tone))/(R(no tone)+R(tone))
    zero if no conditioning
    one if conditioning is complete
  graph suppression versus P(shock|no tone)
    one line for each value of P(shock|tone): 0.4, 0.2, 0.1
    x intercept at P(shock|tone)
    y intercept at 1 for P(shock|tone)=0.4
    decreases as P(shock|tone) decreases

conditioned inhibition
  CS+ (clicker) US (shock)   -> CER
  CS+ (clicker) CS- (light)  (no shock)   -> no CER
  does CS- lead to anticipation of no shock?
  summation test
    condition tone to shock
    tone and light, CER reduced
  retardation paradigm
    takes longer to condition CS- to the US
  
associative bias
  CSflavor (saccharine water) CSav (light and click)
  US (shock) -> greater CER (reduced drinking) to CSav
  US (nausea-producing drug) -> greater CER to CSflavor
  taste aversion
    single pairing
    24 hour interval
  Bayesian viewpoint: posterior = prior * evidence
  against tabula rasa idea, behaviorists

blocking
  control: 8 trials of CS1 (noise) CS2(light) US (shock)
  experimental
    16 trials of CS1, US
    8 trials of CS1, CS2, US    no CER
                
configural conditioning
  respond to A+B but not A or B alone
  respond to either A or B but not A+B
  respond to A+B and C+D, but not A+C and B+D