PSYC 335: Decision Neuroscience

key concepts of decision neuroscience

• human neuropsychology: studying patients with brain injuries show how specific brain regions for cognition/decision-making

• behavioural neuroscience: DA neurons fire to cues that predict rewards, not just rewards themselves

patient EVR

• frontal lobe tumour, underwent surgery and recovered IQ, memory, etc.

  • performed well on neuropsychological assessments, including traditional frontal lobe damage

• however, made poor real-life decisions, preferring immediate gratification with long-term negative consequences

  • failed to learn from mistakes

what does EVR tell us about the function of PFC

brain damage of the PFC leads to real life decision-making deficits, prefer immediate gratification despite long-term negative consequences, myopia for the future

Iowa gambling task (IGT)

• 4 decks, A-B have large short-term wins but long-term losses, but C-D have smaller wins but more long-term gains

• findings: healthy people shift to safe decks, while vmPFC patients persist with risky decks despite losses

EVR performance IGT

preferred risky decks throughout the task, thinking it was the best strategy/used reasoning

healthy control performance on IGT

originally prefer risky, but as the task goes on, they realize the safe deck is the better choice

EVR results on IGT tell us…

have issues with emotional reward punishment integration

• cannot use past emotional experiences of rewards/punishments to guide decisions

ecological validity of IGT

• mimics real gambling, with unknown probabilities, delayed feedbacks, gains/losses

• however, no monetary value to risk

somatic marker

emotional/physiological signals (eg: sweating) is linked to past outcomes that bias future decisions

• people with vmPFC damage fail to retrieve these

skin conductance on IGT - healthy participants

• show SCRs when they win/lose

• show SCRs before risky choices

  • this uses past emotional experience to warn against choosing he bad deck

skin conductance on IGT - vmPFC damage

• show SCRs when they win/lose - can react in the moment

• lack anticipatory SCRs before risky choices - can’t retrieve emotional memories of past outcomes to guide future choice

skin conductance on IGT - amygdala damage

• no SCRs at all

• body cannot be turned on by emotional events, no physiological markers guide decisions

brain issues with gambling

• no observable injury/tissue loss, but subtle brain differences (Li et al. 2019)

• instead, due to developmental disturbance and inability to use certain brain regions

developmental changes in risk-taking

• adolescents have higher risk taking due to “neurobiological imbalance”: mature reward system, immature control system

• gradual decline in risk-taking with age

location of DA cell bodies

substantia nigra, ventral tegmental area

where does substantia nigra secrete DA to?

striatum, via nigro-striatal pathway (for motor control)

where does the VTA secrete DA to?

NAc on striatum, via mesolimbic pathway (for addiction reward)

mesolimbic pathway origin

DA neurons in the VTA

mesolimbic pathway projects to…

NAc (part of ventral striatum)

nigro-striatal pathway origin

substantia nigra

nigro-striatal pathway projects to…

dorsal striatum

Schultz procedure

• recorded APs fired by DA cells

• gave monkey a reward-learning task

  • visual stimulus, then 3 seconds later gives grape juice

  • DA emits AP right after juice is given

• eventually, monkey figures out squiggle means fruit juice

  • DA increases firing in response to squiggle

  • DA is baseline when fruit juice is given

Schultz implications

DA is not interested in pleasure, but in creating expectations and predictions

• learning what leads to the reward, not the pleasure itself

prediction errors in Schultz (2006)

DA neuron responds when there’s a difference between what the monkey expects versus actually happens

fMRI function

• measures blood-oxygen changes, 5 seconds lag

• identifies active brain regions during an activity, but cannot isolate individual neurotransmitters

value function

steeper for losses than gains

endowment effect

people value something they already own more highly than if they don’t own it

amygdala function

necessary for normal fear and loss aversion

somatic marker hypothesis

vmPFC retrieves body-based emotional memories to guid future decision-making

valuation hypothesis

vmPFC encodes the subjective value of options, serving as a commun currency library

transitivity

if A > B, and B > C, then A > C

transitivity in vmPFC patients

show more transitivity errors, supporting the valuation account (unstable preference library)

Shiv - gambling with vmPFC damage

• used IGT, comarping vmPFC damage, amygdala damage, and healthy controls

• vmPFC made disadvantageous choices, lacking SCRs before risky moves

• shows that vmPFC is crucial for integrating emotional experience into future decision

  • somatic marker hypothesis

Plassman et al - willingness to pay

• hungry participants in free bid or forced bid

  • free = choose how much to pay, forced = assigned a bid value

• activity in reward-related areas/vmPFC increased when participants valued items more, decreasing for items they liked less

• vmPFC activity correlates with willingness to pay

Sherman - social media likes

• participants viewed their own posts, manipulated to be either a large or small number of likes

• ventral striatum activity increases with number of likes

dar Meshi - social media and money

• participants were shown photos of themselves paired with a positive or negative adjective

  • ventral striatum was activated when paired with positive words

  • people who spent more time on social media had stronger ventral striatum activation to positive feedbacks

• participants won money unpredictably

  • also activates ventral striatum

shows the brain interprets different kind of rewards in the same region - ventral striatum

Shiv - investment task

• participants decided whether to bet on a coin toss with a positive EV

• healthy participants often rejected bets due to loss aversion, losing out on profit

• patients with damage to amygdala, vmPFC, or insula lack somatic markers, are less averse to loss and are open to more bets to earn profit

which brain regions drive loss aversion?

amygdala, vmPFC, insula

de Martino - amygdala and loss aversion

• study on risky gambles, which bets people would take

• control: strong loss aversion, didn’t want ot take risky gambles

• AP: would accept neutral gambles

• SM: accepted negative EV gambles, actively sought out risks