Exercises

These exercises can completed by adapting the examples in simulate_models.m. Solutions can be found in exercses.m.

About notation: A+ means "stimulus A paired simultaneously with reward", A- means "stimulus A presented without reward", AB means "stimuli A and B presented simultaneously", A->B means "stimulus A followed by stimulus B".

Exercise 1: latent inhibition

• Implement the latent inhibition paradigm with the RW and KRW models: A- (pre-exposure) followed by A+ (conditioning).
• How does the effect change when a delay is interposed between pre-exposure and conditioning?
• How does the effect change with different parameter values?

Exercise 2: delay conditioning

• Implement a delay conditioning experiment with the TD model, where stimulus A is on for some duration (e.g., 5 timesteps) with reward delivered at the final timestep.
• Show the behavior of the prediction error signal over the course of conditioning.
• What happens when a reward is unexpectedly omitted?
• How does the prediction error signal change with different parameter values?