PSYC 340 1st Edition Lecture 13 Learning About S-S RelationsWhat you missed last class..I. Attentional AccountA. For learning to occur, you have to attend to it. This model has two assumptions:1. An organism has a limited attentional capacity. It can only attend to a certain number of elements at a time. 2. The CS must be attended to for learning to occur.B. This model only accounts for overshadowing and blockingII. Informational AccountA. Group 1: Random. There is no relationship between the occurrence of the CS andshock.B. Group 2: Informational. Shock only occurs during the CS.1. Both groups receive the same number of CS and shock pairings, but only the informational group learns to fear the CS. a. This is because the contiguity is insufficient for learning.C. The informational account is stated in terms of probabilities. If the probability of shock given the CS equals the probability of shock without CS, then no learning will occur. 1. This model accounts for conditioned inhibition.III. Rescorla-Wagner ModelA. Builds upon the basic notion that learning only occurs when you are surprisedB. Whether or not the US is surprising will depend on the CS-US associative strengthC. Variables:1. Λ = magnitude of the US2. V = the degree to which you expect the US or the CS-US associative strength3. VT = V1 + V2 +… + VN Accounts for multiple CS and equals the net expectation of the US. All CS elements being concurrently presented determine the net expectation of the US4. α = salience of the CS, it can only between zero and one5. β = learning rate parameter for the US, also only between zero and one, accounts for biological constraints on learningD. The goal of this model is to predict the change in associative strength observed for a particular CS in a particular trial. To do this, we use the Rescorla-Wagner Model equation: ΔV = αβ(Λ – VT)These notes represent a detailed interpretation of the professor’s lecture. GradeBuddy is best used as a supplement to your own notes, not as a substitute.Exam 2 I. Neuroboxes 3-4 will NOT be on the exam! However, neurobox 2 will beII. Will be covering everything from 2/10 on syllabus and the very beginning of 3/3 (Pavlovian conditioning in Aplysia) Rescorla-Wagner Model (1972) ΔV = αβ(Λ – VT)I. lambda = magnitude of the USII. V = the degree to which you expect the US or the CS-US associative strengthIII. VT = V1 + V2 +… + VN Accounts for multiple CS and equals the net expectation of the US.All CS elements being concurrently presented determine the net expectation of the USIV. Lambda – Vt = surprise; you will learn if you surprise V. DeltaVx – change in associative strength to x; proportional to αxβVI. ΔV = αβ(Λ – VT)A. Learning stops when VT is equal to lambda (you fully expect the US) VII.How does the model account for spontaneous recovery?A. It really doesn’t; can’t accurately account for it VIII. ApplicationA. Blocking B. Overshadowing 1. if A is twice as big as X, A gets 2/3 lambda and X gets 1/3 lambdaC. Extinction 1. If there is no US (like in A-,) lambda should be zero 2. You will lose associative value when DeltaVa = negative 3. Accurately represents extinctionD. Conditioned inhibition E. US pre-exposure effect1. Environment = c = context 2. CS = x3. Blocking experiment! IX. Novel Predictions A. Can the model handle the basic phenomena? Yes, it can! 1. However, it must make novel predictions if it is to be valuableB. Over expectation 1. Trained A+ and X+ to lambda; in phase 2, you expect 2 lambda but only get 1; you need to bring your expectations down to what you get; lose associative value C. Contingency Effects1. Put animals on a random trial, early on you will blame a cue, though it is totally
View Full Document
Unlocking...