Neuroscience and decision making (perceptual) Flashcards
making a model system
primate brain as model system
perform simple perceptual tasks based on uncertain sensory info
record key brain areas
suggests perceptual decision making based on evidence accumulation
What is perceptual decision making
selecting an action based on uncertain sensory info. Make inference about state of world
Random dot motion paradigm
look at random movement of dots. some move coherently. figure out if net movement is left or right and move eyes to indicate. can manipulate uncertainty and viewing time.
tests as evidence comes in overtime and is uncertain
behavioural results of random dot paradigm
more accurate with more coherence
decreased reaction time with more coherence
increased accuracy (coherence threshold) with increased viewing time
Schematic model
sensory unit (processing visual motion) -> sensorimotor unit (action planning) -> motor unit (action execution)
Visual motion processing in area MT
neurons large receptive field. populations of MT neurons Sensitive to direction (left/right preference)
fire when match preference
amount of firing within population estimates motion direction at that time
temporal interrogation
measure of total cumulative evidence
at certain point evidence is enough to respond to direction
more accurate the more distinct evidence there is (greater separation in cumulative evidence
effect of coherence in cumulative evidence
higher coherence means gain enough cumulative evidence quicker
eye movement planning in area LIP
Paradigm animal look at fixation point then target appears. fixation point disappears and move eyes to target
peak of LIP after target appears so involved with action planning
LIP and dot movement task
record LIP neuron that contains movement target in response field.
greater activity for good quality evidence (so accumulating evidence)
stop accumulating when sufficent evidence . decision threshold for movement
evidence to accumulation threshold
better evidence -> higher accumulation rate -> shorter RT
Accuracy instructions -> raise threshold -> slower more accurate decisions
Speed instructions -> lower threshold -> faster but less accurate decisions
speed-accuracy trade off