Decision making

Question 1

What is the prediction-choice-outcome loop?

Answer 1

We start out with a goal that we we’d like to achieve, that ultimately requires us to make a decision
If there are several options, how to achieve this goal, we make predictions about the outcome of different options
We then form a decision and make appropriate actions that should get us closer to our goal
The result of these actions will be an action outcome that we observe
This outcome will be subjected to internal monitoring processes where we evaluate, if our decision and the corresponding actions have actually achieved our goal (or at least brought us closer to it). If we haven’t reached our goal, our brain seems to generate a prediction error, i.e. a signal indicating how large the discrepancy is between what we had predicted originally and what the actual outcome was.
Importantly, the result of this monitoring process, the prediction error, can be used to update our memory based on the experience we have just made. This updated memory content can then be used to make more precise predictions the next time when we face the same or a similar choice.

Question 2

What is prediction error?

Answer 2

The result of the monitoring process where we evaluate if our decisions have actually achieved our goal. If the goal has not been achieved, it is said to be a prediction error.

Question 3

What factors need to be considered before making a decision?

Answer 3

1. The difficulty of the action (Effort)
2. The probability of success and failure
3. The value of a choice
4. Missed opportunities

Question 4

What are the biases observed in decision making?

Answer 4

1. To stick with your default option
2. To choose certain gains over gambles
3. To choose gambles over certain losses
4. Temporal discounting: The phenomenon that individuals value a reward less if they have to wait for longer

Question 5

What are the different level of decision making?

Answer 5

Simple perceptual decisions
More complex decisions

Question 6

How is perceptual decision making studied?

Answer 6

Using simple perceptual tasks

Question 7

What is evidence accumulation?

Answer 7

The more a chosen phenomenon is observed (either directly or indirectly using imaging studies) to occur, the steeper the evidence accumulation curve becomes. When it reaches the threshold, a response happens.

Question 8

What happens if a stimulus is noisy?

Answer 8

The evidence accumulation curves reach the destination faster and may lead to incorrect decisions. This especially happens if the accumulation for the less dominant process reaches the threshold first.

Question 9

What are the stages of perceptual decision making?

Answer 9

1. Detection of sensory evidence
2. Integration of evidence over time
3. Checking if the threshold has been reached (if not, get more evidence).

Question 10

Where does evidence accumulation take place?

Answer 10

Brain areas responsible for coding the relevant feature, such as Mt/V5 for motion. Or in the parietal and dorsal prefrontal cortex areas. Sensorimotor areas show evidence of evidence accumulation as well.

Question 11

What is a mental map?

Answer 11

An internal representation stored in our memory that guides us with the different options present. This model helps us predict the different outcomes of the available options. In order for these models to be as accurate as possible, new experiences that contribute to the accuracy of the models need to be regularly added.

Question 12

Edward Tolman

Answer 12

Rats were put in a maze where a path from A to B was blocked. When the blockage was removed, the rats knew to use the path, which means that they had encoded transitive relations of the different locations in their mental representations of the maze.

Question 13

In what contexts are these mental maps used?

Answer 13

Spatial tasks and problem solving

Question 14

What are problems (in this context)?

Answer 14

A series of decisions

Question 15

Hippocampus

Answer 15

Stores long-term memory and is thus important for decision making

Question 16

Medial frontal cortex

Answer 16

Can be looked at through imaging studies to determine how participants approach decision making. Seems to be linked to memory formation in decision-making.

Question 17

What is a state space?

Answer 17

A cognitive map for a specific task (preparing a particular meal, writing an essay, washing your dishes). Assumed to be grid-like representations of all the decision points along the way from the original problem to the goal.

Question 18

What is the hidden state?

Answer 18

The point in a state space that represents what decisions we are currently facing.

Question 19

What did Schuck et al demonstrate

Answer 19

Some aspects relevant for decision making are represented in the MFC. The MFC is linked to memory formation in a decision-making task.

Question 20

What brain parts are associated with the subjective values of opinions?

Answer 20

The medial, ventro-, and orbitofronal cortex have been associated with the subjective values of opinions. Subcortical areas such as the thalamus and the stratium have been identified in relation to subjective valuation as well.

Question 21

What explains increased risk taking in Parkinson’s disease?

Answer 21

The striatum and the dopaminergic system are affected. The medications that PD patients take counteract the loss of dopamine; these increased levels might cause increased risk taking behaviours.

Question 22

Where are goals represented?

Answer 22

Medial orbitofronal cortex

Question 23

What parts of the brain track the expected value of something in line with the current goals?

Answer 23

the ventromedial prefrontal cortex and ventral striatum.

Question 24

What do lesion studies in the OFC-vmPFC tell us?

Answer 24

Deficits in this region show impairments in simple preference judgements in value-based decision making tasks.

Question 25

What is the trust game?

Answer 25

It works as follows: Person A will be given a certain amount of money with the instruction that they might get to keep whatever they have left at the end of the game. They will then be told that they can decide how much of this money they want to give to Person B, a trustee. Whatever amount Person A gives to Person B will be tripled by the experimenter. So the more Person A gives to B, the more money B will receive from the experimenter.
Person A is informed that Person B can then decide how much of this money to give back to Person A. Obviously, a rational strategy from A’s perspective would be to give as much money as possible to Person B and trust that they share whatever they have, thus optimizing the reward for both of them.

Question 26

How do vmPFC lesions impact choices in the trust game?

Answer 26

They tend to trust less and choose to keep all the money instead of sharing it.

Question 27

What is reward rate maximisation>?

Answer 27

When individuals aim to maximise the rate of subjective rewards

Question 28

What impacts reward rate maximisation?

Answer 28

Cost of time, number of options, risk sensitivity, subjective effect

Question 29

What is utility?

Answer 29

Reward value. The payoff of an outcome.

Question 30

What is the reward rate maxmisation formula?

Answer 30

(utility x success probability) - cost / deliberation time + handling time + ITI (inter-trial interval)

Question 31

What is the urgency signal?

Answer 31

In the interest of reward rate maximization, we need to set some limit to our deliberation time. You might know this feeling from some of the decisions you have taken in the past: At some point, it makes sense to just go with one option instead of taking even more time to decide. Therefore, it is assumed that there is a mechanism that can push the evidence accumulation over our threshold to take action. This is the so-called urgency signal.

Question 32

Assumptions about the urgency signal

Answer 32

1. Serves to maximise the reward rate by pushing the evidence accumulation closer to the decision threshold.
2. Controlled by projections from the basal ganglia to cognitive and sensorimotor areas.
3. It grows during the deliberation time and helps optimise it. The longer the deliberation time, the stronger the urgency signal.
4. It is modulated by task context.
5. It shows inter-individual differences