Ivan Pavlov Conditional stimulus Unconditional stimulus Unconditional responsereflex

  • Slides: 30
Download presentation

 התניה קלאסית Ivan Pavlov = Conditional stimulus ( תלוי / )גירוי מותנה =

התניה קלאסית Ivan Pavlov = Conditional stimulus ( תלוי / )גירוי מותנה = Unconditional stimulus ( תלוי / )גירוי בלתי מותנה = ריור Unconditional response(reflex); conditional response (reflex)

CS- ל US יחסים טמפורלים בין 1. Simultaneous Conditioning 2. Delay conditioning 3. Trace

CS- ל US יחסים טמפורלים בין 1. Simultaneous Conditioning 2. Delay conditioning 3. Trace conditioning 4. Backward conditioning CS US Key Variable: The CS-US Interval (ISI)

 מתי מתרחשת למידה? שלושה ניסויי מפתח 1. Rescorla – Background conditioning Þ Temporal

מתי מתרחשת למידה? שלושה ניסויי מפתח 1. Rescorla – Background conditioning Þ Temporal contiguity is not enough, need contingency Contiguity = הופעה יחד , סמיכות Contingency = תלות

 מתי מתרחשת למידה? שלושה ניסויי מפתח 2. Kamin – Blocking (and unblocking) 3.

מתי מתרחשת למידה? שלושה ניסויי מפתח 2. Kamin – Blocking (and unblocking) 3. Reynold – Overshadowing Þ Contingency is also not enough!! Þ Kamin: The US needs to be surprising Þ Seems like the stimuli compete for learning

TD learning (Sutton+Barto ‘ 90 s) The general case: long term prediction. The true

TD learning (Sutton+Barto ‘ 90 s) The general case: long term prediction. The true predictions should be self consistent: If the predictions are imperfect, there will be an error: Temporal Difference error Updating V according to this will result in correct (optimal) predictions

Dopamine - דופמין Parkinson’s Disease Motor control + initialtion? Intracranial self-stimulation; Drug addiction; Natural

Dopamine - דופמין Parkinson’s Disease Motor control + initialtion? Intracranial self-stimulation; Drug addiction; Natural rewards Reward pathway? Learning? Also involved in: • Working memory • Novel situations • ADHD • Schizophrenia • …

Montague+Dayan מה דופמין מייצג? פרשנות של Unpredicted reward (unlearned/no stimulus) Predicted reward (learned task)

Montague+Dayan מה דופמין מייצג? פרשנות של Unpredicted reward (unlearned/no stimulus) Predicted reward (learned task) Omitted reward (probe trial) (Montague et al. 1996)

The TD hypothesis of DA (Montague+Dayan ‘ 96) The idea: Phasic dopamine encodes a

The TD hypothesis of DA (Montague+Dayan ‘ 96) The idea: Phasic dopamine encodes a reward prediction error • Precise (normative!) theory for generation of DA firing patterns • Compelling account for the role of DA in classical conditioning: prediction error acts as signal driving learning in prediction areas • Corticostriatal synapses: three factor learning rule modulated by DA (Wickens+Kotter)

Corticostriatal synapses: 3 factor learning Stimulus Representation X 1 X 2 X 3 XN

Corticostriatal synapses: 3 factor learning Stimulus Representation X 1 X 2 X 3 XN V 1 V 2 V 3 VN Cortex Adjustable Connections (“weights”) PPTN? R P VTA/SNc Striatum Prediction Error (Dopamine)

More dopamine responses • Partial reinforcement task (Fiorillo, Tobler & Schultz 2003) • Accords

More dopamine responses • Partial reinforcement task (Fiorillo, Tobler & Schultz 2003) • Accords with TD model