Classical Conditioning and Associative Strength free essay sample

The experiment conducted was a 22 within subjects experiment. The 2 cue types were over-expectation and control whereas the 2 independent variables were salience. Subjects were randomly presented varying levels of salient stimuli in the form of food pictures. These pictures were then to be associated with an allergy score between 1 and 9. In the next phase, the subjects were given each cue type 8 times along with some filler cues arranged to some simple rules. The allergy ratings were also to be determined by the subject. The third phase combined the stimuli into a compound CS conditioning and subjects were asked to determine the allergy rating. The test phase involved participants being shown each of the previous foods one at a time. They would then determine the allergic reaction rating. Overall, the results of the experiment used an analysis of variables to show that the likelihood rating for the high salience cues were, on average, lower than those of the lower salience. A significant main effect with cue-type was found, indicating an over-expectation effect, of F(1,271)=25.  758, plt;0. 001. This means that participants rated the compound cues in phase 3 as less allergenic than those that were elementally reinforced in phase 3. While there was no significant difference found between the elementally reinforced stimulus cues I and J (pgt;0. 05), there was found to be a significant difference between the compound cases involving stimulus cues B and F (plt;0. 05). Question 2: With compound CS, the associative strength is now dependant on the combined associative strength of all stimuli present and the total amount of learning. Generally in the Rescorla-Wagner model, conditioning is dependent on how effectively surprising an unconditioned stimulus is. With compound conditioned stimuli, this characterisation of surprise is altered. Instead of just a single CS being linked to a US, it is now the combined effects of all stimuli present which collectively predict the US. Over-expectation occurs when the collective associative strength of the various CS is greater than the total capacity to learn. To use the equation, ? V= (? -V) Note: We ignore ? and assume it a constant. The value of V would be  greater than ?. This occurs when the stimuli have been well presented and the associative strength of each individual stimulus is quite high. Because of this, when the target cue for, which the stimuli are conditioned for, is presented the subject expects the combined response of each individual stimulus. When the subject only receives the normal amount of reinforcement, it begins to expect less until there is no mor e association that can be gained from the combined effects of the condition stimuli. To use an analogy, a mouse is conditioned separately with a tone and a flashing light. Each of these stimuli provides one piece of cheese, with the mouse salivating when each stimulus is presented. If both the stimuli are presented together, the mouse expects two pieces of cheese and will salivate more than usual. However, only one piece of cheese is presented. As a result, the mouse begins to expect less of the compound stimuli until no further association can be gained. This consequently lowers the overall association by a certain rate until the total combined association is at total learning potential. Question 3: To get the over-expectation result, we must first have two fully conditioned CS’s which were separately conditioned. For example, let us use the A+ and B+ case. If we fully conditioned A+ and B+ separately, that means the associative strength of both would be 100%. Now, when we combine these two into a compound CS, the Rescorla-Wagner model states that the overall associative strength of the compound CS is equal to the sum of the all the stimuli presented in the compound trial. So with A+ and B+, their total associative strength would be 200%. However, the Rescorla-Wagner model only allows the maximum value of associative strength that can be conditioned to be 100%. The case of A+ and B+ is twice as much as this maximum value. Therefore, after repeated trials of the compound stimulus of A+B+ being paired with their US, the Rescorla-Wagner model predicts that the total associative strength of 200% will fall at the rate of their salience until it reaches a strength of 100%, the maximum level which can be conditioned. To show how an extinction event of a fully conditioned CS operates in the Rescorla-Wagner model. ?V= (? -V) Note: We ignore ? and assume it a constant. With a fully conditioned CS, ? , the total amount of learning, would be 0 as there is no more possible connection to make with a fully conditioned CS. V would be 100% as all possible learning has been made. Therefore, ? V=-? would be what is left over. This means that, in extinction, the rate at which the association strength, ? V, is lost at is –?, the salience. Now, in the case of B+, it was a much more salient stimulus whereas F+ was a lower salient stimulus. Therefore, B would be lost at a faster rate than F. This is shown in the results with the likelihood rating of B being lower than that of F. Therefore, salience of each individual stimulus not only effects how fast the rate of association is increased but also determines, in the case of the over-expectation scenario, how fast the association of each individual, fully co nditioned stimulus is lost.