Neuroscience of decision-making: value-based decisions Flashcards
Value-based
Valuing and choosing prospects with uncertain outcomes
- modelled with expected utility theory
Perceptual
action selection based on uncertain evidence
- modelled with evidence accumulation threshold
Studying neural mechanisms underlying value-based decisions
Neurophysiology - single cell recording
Neuropsychology - patient studies
Neuroimaging- fMRI
Neuroeconomics
the study of neural mechanisms underlying the variables and computations assumed in psych economic theories
single cell recording - probability in expected value
cued saccade task- Lip neurons recorded. fixation point and 2 other targets appear either side. point changes colour and look at matching target when point disappears. targets cued with different probabilities (.8 and .2) so knowledge about which target cued. so even before cue neural response to target different. differences in activity disappear when cue appears
Magnitude of reward
associate different target colours with different rewards. if target in neuro receptive field associated with lower rewards then less activity then ones with higher reward
Expected value in the brain
study
recorded LIP neurons
firing rates modulated by probability of movement and magnitude of reward
EV- value x probability
so sensitive to both components of EV
Patient studies
most neuroscience correlational
natural manipulation - study patients with neurological damage
Damage to ventromedial pre-frontal cortex associated with decision making in the presence of otherwise normal cognitive function
tumour patient
removal led to bilateral lesion to orbitofrontal cortex and serve impairment of decision making
Iowa Gambling task outlined
choose a card from one of four decks and try and gain max reward. uncertain outcomes. start with $2000. Decks a and b are high risk choices as give high rewards and punishments and are used to measure learning an decision making impairment. Decks c and d are low risk decks and advantages in the long run
Iowa Gambling task results
Typical person initially tries all decks but then moves to predominately c and d
patients- kept going for high risk a and b
Control brain damage- behave like typical person
sub-dividing brain areas associated with decision making
vmPFC - assigning values to options in absence of uncertainty or learning
OFC - updating values based on reward feedback
fMRI
loss aversion
negative emotional response - amygdala
asymmetric valuation system - vmPFC
gamblers choose form asymmetric matrix where diagonal is loss is 2x gain.
accept if above diagonal.
RT show choices with larger differences between gains and losses easier
gain areas - vmPFC and OFC
loss area- vmPFC and OFC (lower activation)
no activation of amygdala so emotion not important?
network involved - vmPFC and ventral striatum in Basal Ganglia
neural and behavioural loss aversion correlate