3) Decision Neuroscience

what are the two observation of the modern field of decision neuroscience (2)

• patients with vmPFC lesions show poor judgement and risky decision making in the real world

• behavioural neuroscience studies show that during pavlovian conditioning, dopamine neurons fire to unexpected rewards and conditioned stimuli that predict reward

Damasio’s patient EVR (2)

• formerly a pillar of society, a store manager/ accountant, husband and father

• had a frontal lobe tumour that was detected early and had surgery to remove it

Damasio’s patient EVR post operative recovery (7)

• above average IQ, memory, perception

• no psychiatric symptoms

• performed well on neuropsychological assessment, including traditional tests of frontal lobe damage

• but he made many poor real life decisions

  • unreliable employee, fired from jobs

  • questionable business ventures that led to bankruptcy

  • erratic personal life, got a divorce and moved back with parents

what is brain damage to the vmPFC linked to (4)

• real life decision making deficits

• where choices offer immediate gratification but with long term negative consequences

• patients failed to learn from their mistakes

• myopia for the future, shortsighted in their choices

Iowa gambling task (4)

• given 100 choices from 4 decks of cards

• decks A and B are risky outcomes, they have high immediate reward but very high long term punishment, have a net loss (e.g. win $100 or pay $250)

• decks C and D are safe outcomes, they have low immediate reward but minimal long term punishment, have a net profit (e.g. win $50 or pay $50)

• net score = safe choice (C/D) - risky choices (A/B)

results of the Iowa gambling tasks on normal vs brain damaged patients vs EVR and EVR patient type (2)

• normal and brain damaged patients do well and choose the safe choice options

  • at the start, they are drawn to the higher rewards but after a few trials, they realise that the penalties are not worth and chooses the safe deck

• EVR and EVR patient type, the target subjects at the end, they are still choosing the risky desks

ecological validity of the Iowa gambling task (3)

• it lacks a wager as nothing it put forward

• it lacks a real incentive

• there is no pure chance as it is a learning task

arousal and risky decision making in Iowa gambling task (4)

• tested using skin conductance, sweating from fingertips to measure the somatic markers

• healthy ppts show skin conductance responses to winning and losing outcomes and develop anticipatory responses before card choices esp to the risky desks

• patients with vmPFC damage show outcome responses to wins and losses but do not develop anticipatory responses before their choices

• patients with amygdala damage struggle to mount any psychophysiological responses to outcomes or choices

brain structure in people with gambling problems

structural MRI in people with gambling disorder shows differences in morphology, sulcus and gyrus folds in the orbitofrontal cortex

developmental changes in preferences for risk (2)

• adolescents are widely assumed to be high risk takers where there is a non linear effect, driven by a neurobiological imbalance, where there is a mature system and immature frontal control system

• but research actually indicates a gradual decline and improvement in risk taking as they get older

what are the 2 dopamine pathways

• nigro striatal pathway

• meso limbic pathway

nigro striatal pathway (2)

• starts at the substancia nigra and to the striatum

• the pathway that degenerates in Parkinson’s disease, affecting movement

meso limbic pathway

starts at the ventral tegmental area (VTA) to nucleus accumbens, that is part of ventral striatum that processes reward

where are dopamine cell bodies located in the brain (2)

• substancia nigra

• ventral tegmental area

Schultz’s study on dopamine and reward learning (4)

• during appetitive pavlovian conditioning, midbrain dopamine neurons signal prediction errors

• early in the task, phasic responses follow the juice

• with learning, the phasic responses shift to the conditioned stimulus, picture that predicts reward

• from this point, the dopamine cells no longer fires to the reward itself

what is the modern view on dopamine

changed from pleasure to motivation, information, prediction, learning and driving behaviour

how do we record individual cell in brain and measure dopamine in humans

using functional MRI

functional MRI (3)

• not measuring dopamine directly but through blood flow response

• relies on oxygenated and de oxygenated blood having different magnetic properties

• has good temporal resolution as blood flow changes every 2- 7s

conclusion of fMRI research on reward system

in each research, see ventral striatum responding to different kinds of rewards

reward system in social media use (4)

• social media is a complex platforms where ppl can post, receive feedback, observe, provide feedback or compare

• interested in which part of the brain responds to positive description of yourself vs other three condition

• sees correlation with brain areas that responds to pairing of positive words with yourself and real world weekly facebook use

  • people who used facebook more show stronger response to positive reputation words and photos of themselves

study on instagram likes and brain areas

found that ventral striatum activity increases with number of likes on posted photos supplied by the participants compared to neutral and risky images

regulatory pressure on social media (4)

• concerns dating back to 2019 regarding infinite scroll, auto play and personalised recommendations

• some platforms offer de- metricating options, removing numbers and metrics

• EU digital services act require risk assessment to be submitted for any design changes to large online platforms

• recent major us lawsuit against meta

gambling in parkinson’s disease (4)

• ppl with parkinson’s disease have dopamine neurons that are degenerated

• dopamine agonist medications can trigger excessive gambling as a side effect

• most linked to 2 drugs, pramipexole and ropinirole that bind to dopamine D3 receptors

• other side effects include impulse control and reward problems

Winstanely - rat gambling task (3)

• rat is in a skinner box with several holes where the rat can poke their nose to make a choice where a light will light up when choice is made

• to translate IOWA to rats, switched money to food pellets, rats got 1-4 pellets depending on the hole, probability also exists with the holes (e.g. hole with 1 food pellet has a higher probability of reward and hole with 4 food pellets has a lower probability of reward)

• penalties is time out where the box goes dark and nothing works in the box, length depends on the choice of hole

Barrus and Winstanley - extension of rat gambling task (5)

• made changes to the tasks by adding bells and whistle

• cued condition: when it wins food pellets, there are flashing bulbs and chimes, intensity differs with number food pellets

• uncued condition: original version of rat gambling task

• found that adding audiovisual feedback reduced preference for the best “deck”, it was harder to determine the best option but in uncued condition they selected the best option

• injecting dopamine D3 agonist had a dose dependent effect of reducing choice of option 2 and increasing choice of option 3, the risky choice

value function - win $10 vs pay $10 (2)

• the function is steeper in the loss domain compared to the gain domain

• so losses dominate over gains

why is there a bias for losses dominating gains in our decision making

as species who are foraging in environments with scare resources, there is a drive to protect what we have and gain if there is an opportunity

endowment effect (3)

• a cognitive bias where people assign a higher value to objects simply because they own them

• asked students to choose how much money to sell mug after receiving it during the summer vs just asking how much they are willing to pay

• value of willingness to sell is greater than value of willingness to pay

loss aversion

a cognitive bias where the psychological pain of a loss feels roughly twice as intense as the pleasure of an equivalent gain

Shiv et al - investment task (4)

• asked 20 times if they want to invest $1 on coin flip, heads lose investments vs tails win $2.50

• EV is positive and would make a profit if they say yes

• but healthy participants say no to alot of gamble as they are loss averse and don’t want to lose the $1

• target patients, vmPFC/ amygdala/ insula damage make higher profits and accept the investment because they lack emotional response to loss aversion

De Martino - loss aversion in amygdala (5)

• looked at the effect of the amygdala on loss aversion

• looked at 2 rare patients with urbach wiethe disease which is a rare genetic disorder where calcium deposits cause selective amygdala atrophy

• does a whole series of loss aversion exercise

• both would accept gamble even when there is a lower EV value whereas control would only accept gamble when there is a high positive EV value

• amygdala damage show abolished loss aversion therefore showing its importance of amgydala in this effects

two theories of vmPFC function

• Damasio/ Bechara - somatic marker hypothesis

• Kable and Glimcher - valuation perspective

somatic marker hypothesis (2)

• vmPFC retrieves emotional memories from similar decisions made in the past to guide decisions making

• emphasises physiological signals in the body

valuation perspective (2)

• vmPFC contains information about subjective values, utilities of objects or actions

• emphasises brain signals, no view on the body

Plassmann et al - willingness to pay study (5)

• auction tasks, bidding for food they like

• free bid trial where they can bid whatever

• control condition where they are told what to bid

• looked at brain activity while deciding what to bid and how much you actually bid

• vmPFC correlates with economic willingness to pay

meta analysis of fMRI studies of subjective value (3)

• fMRI signals in vmPFC scales with subjective value as well as expected and delayed value

• vmPFC converts outcomes into a common neural currency that allows us to make our decisions

• vmPFC is also more interested in choices than outcomes