Reward-related brain activity Flashcards
Explain the famous endless lever-pressing task.
In 1954 James Olds and Peter Milner found that rats will press a lever until exhaustion, sleep, and press it some more if it stimulates septal nucleus.
It was later discovered that this was true in many places along the medial forebrain bundle (MFB)
MFB
medial forebrain bundle.
What is being stimulated during pleasure in the medial forebrain bundle?
Fibers from dopamine nuclei.
What does the medial forebrain bundle contain?
axons going from dopaminergic cell groups in the midbrain to the basal ganglia and cerebral cortex.
What are dopaminergic cell groups in midbrain the project through the MFB
ventral tegmental area and substantia nigra pars compacta
VTA
Ventral tegmental area
SNpc
substantia nigra pars compacta
Optogenetic
Use of light to control cells
What happened when midbrain dopaminergic cells were stimulated in rats through optogenetics?
Rats worked to get the stimulation (nose poking). If the stimulation was gone, they promptly quit working. They started again if they were made aware of the availability.
How were hedonic and learning hypothesis teased apart in testing?
Looking at neuronal activity, they could see pleasure at reward onset. However, during conditioning, they found pleasure at cs onset and not at reward onset. Additionally, the extent of the firing was related to the amount of juice the cue predicted.
It’s a phasic response
How did Schultz further show that learning was at play for these dopamine cells?
Monkey either learned a cue (click) for a food item or discovered food when he opened the door and touched it. Both acted as conditioned stimuli and both prompted response. In the click condition, touching the food elicited no response.
The prof refers to these are surprising or unpredicted incidences when they got the cue.
What is a reward prediction error?
There’s a difference between what is expected from CS and what is received.
Scenario 1: Cell response to unexpected reward
Scenario 2: Cell response only to unexpected cs
Scenario 3: Cell response to unexpected cs and then suppression of response when reward doesn’t come.
The difference between expectation and reward serves as a training signal for refining prediction
Why is the prediction error important?
According to Rescorla-Wagner, classical condition requires that the initial pairing be surprising or unexpected and therefore generate a prediction error. For example, in a test animals do not learn that a new stimulus will predict reward during compound cues if one of the cues has already predicted a reward. Likewise, if a cue has not previously yielded a reward and then is paired with a compound cue, the animal learns that the new cue predicts reward.
How is value of the reward affected?
by timing and probability of reward
How is it clear that DA neurons are sensitive to probability of reward?
Neurons in the VTA fire more strongly for cues that predict a higher probability of reward. Also, some neurons in this area fire more strongly as the reward might approach during maximal uncertainty of reward
What does the random dot discrimination task show?
As the level of coherence of the motion of the dots goes down, probability that the monkey will get the trial correct declines and that means less chance of reward. For a small reward, DA neurons don’t show much difference between probabilities, but they fire according to probability for large rewards.
What is a weakness of the learning hypothesis?
DA neurons do not appear to encode punitive events comparably to rewarding events. When getting juice, air puff, or bitter taste, the DA neurons only signal prediction errors for juice.
Unpredicted is just the reward. Predicted is using a CS. Prediction error is not related to value of reward or lack of reward.
What is the incentive salience hypothesis?
It’s all about wanting.
DAT knockdown mice learn at a similar rate, but their hyperdopaminergic brains make them more energized to work harder for the reward.
Also, their liking of sweet or aversive to bitter is no more common contrary to Hedonic hypothesis.
What are the three hypothesis of reward?
Hedonic, learning, incentive salience
Which hypothesis is most valid?
Trick questions. There doesn’t need to be one answer.
Describe the midbrain projections
Midbrain dopamine neurons project to multiple targets including the dorsal striatum, the ventral striatum, prefrontal cortex, the amygdala and the hippocampus. They have multiple modes of action including via phasic and tonic activity and via d1 and D2 receptors and via neurons of different types.
How does dopamine relate to decision making?
Dopamine neurons encode the value of a chosen response relative to the value of the unchosen cue, but this is all post-choice. They don’t participate in making the decision
What is a fourth interpretation of reward-related activity?
representation of anticipated value in the service of economic-decision making in OFC.
How do orbital frontal neurons act in response to probability vs value of juice?
Offer value: Some neurons signaled the value of a juice of a particular flavor.
Chosen value: Value of chosen juice regardless of flavor
Only fire if selection for cells preferred juice is chose.
Are decisions in orbital frontal goods based or action based?
Goods based (He chooses the flavor and then the direct).
Action based (He associates a flavor and a direction and chooses the action.)
