Prediction Errors Flashcards
The central construct in the conceptualization of learning that is described by Niv & Schoenbaum in their description of temporal difference theory is:
A. Positive reinforcement
B. Feedback
C. Knowledge
D. Prediction
D. Prediction
A central tenet of contemporary learning models described by Niv & Schoenbaum is that:
A. Ability to develop an accurate model of the intrinsic value of a stimulus is critical to reinforcement learning.
B. The relative frequency of positive vs. negative reinforcers is a critical signal for learning.
C. The relative frequency of rewards and punishments plays a critical role in learning.
D. Discrepancies between expected (predicted) and actual outcomes generate critical signals for learning.
D. Discrepancies between expected (predicted) and actual outcomes generate critical signals for learning.
From electrophysiological recording of neuronal activity in nonhuman primates, the “prediction error” signal appears to be encoded as:
A, changes in tonic firing of cholinergic neurons
B. changes in tonic firing of dopaminergic neurons
C. changes in phasic firing of cholinergic neurons
D. changes in phasic firing of dopaminergic neurons
D. changes in phasic firing of dopaminergic neurons
As described by Niv & Schoenbaum, the Rescorla-Wagner model is:
A. A model for how punishment influences behavior in instrumental learning paradigms.
B. A reinforcement learning model for instrumental learning of goal-directed behaviors.
C. A model for how prediction errors are used to update stimulus value in classical conditioning.
D. A model for how word learning occurs during development.
C. A model for how prediction errors are used to update stimulus value in classical conditioning.
According to Niv & Schoenbaum, what is the main difference between the Rescorla-Wagner model and temporal difference models?
A. The Rescorla-Wagner model applies to learning of stimulus-stimulus associations and the TD model applies to learning of stimulus-reward associations.
B. The Rescorla-Wagner model applies to discrete learning trials and the TD model allows for continuous updating of reward predictions over time.
C. The Rescorla-Wagner model applies to prediction of rewarding outcomes and the TD model to prediction of aversive outcomes.
D. The Rescorla-Wagner model applies to prediction of rewarding outcomes and the TD model to prediction of both rewarding and aversive outcomes.
B. The Rescorla-Wagner model applies to discrete learning trials and the TD model allows for continuous updating of reward predictions over time.
According to the “fingerprint” for prediction errors described by Niv & Schoenbaum, when a person receives her paycheck at the normal time in the normal amount, a response in her brain consistent with a prediction error signal is:
A. tonic increase in activity when she receives the paycheck
B. a phasic increase in activity when she receives the paycheck
C. phasic decrease in activity when she receives the paycheck
D. no change in activity when she receives the paycheck
D. no change in activity when she receives the paycheck
Which of the following statements is consistent with Niv & Schoenbaum’s discussion of prediction error signals to aversive outcomes?
A. That a “salience” signal would be expected to show a larger response to an unexpected aversive outcome than to an unexpected reward.
B. That any brain response to an unexpectedly aversive outcome would be more consistent with a “salience” signal than a “prediction error” signal.
C. That bursts of increased firing of dopaminergic neurons on prediction of aversive events has been well established.
D. That a “salience” signal would be expected to exhibit the same form for prediction of unexpected positive and unexpected aversive outcomes.
D. That a “salience” signal would be expected to exhibit the same form for prediction of unexpected positive and unexpected aversive outcomes.
According to Niv & Schoenbaum, which of the following kinds of learning is best accounted for by TD learning models?
A. Change in preference for a restaurant after consuming a meal there.
B. Recognition of actor’s name after viewing a movie in which she played a role.
C. Better ability to discriminate birds after birdwatching trip.
D. Association of the color red with fire engines.
A. Change in preference for a restaurant after consuming a meal there.
Which of the following statements is NOT CONSISTENT with Niv & Schoenbaum’s discussion of the “prediction error signals” that have been observed on functional imaging?
A. Prediction error type activation in structures such as ventral striatum and orbitofrontal cortex could reflect input from multiple sources, including nondopaminergic projections.
B. Prediction error type activation in structures such as ventral striatum and orbitofrontal cortex probably reflect firing patterns of dopaminergic neurons within these structures.
C. Prediction error type activation in structures such as ventral striatum and orbitofrontal cortex may be influenced by firing rates in dopaminergic neurons that project to these regions from the midbrain.
D. Prediction error type activation in ventral striatum by aversive outcomes is unlikely to reflect input from midbrain dopaminergic neurons.
Submit Answer
B. Prediction error type activation in structures such as ventral striatum and orbitofrontal cortex probably reflect firing patterns of dopaminergic neurons within these structures.
