5.1 signal detection theory

what does the signal detection theory state

stimulus depends on various factor and helps us predict how such factors affect the detection of stimuli

what are some factors that affect stimulus detection

stimulus strength, environmental factors, observer’s psychological state, physiological factors

examples of stimulus strength that affects stimulus detection

intensity of a stimulus, how bright or dim a stimulus is

examples of an environmental factor that affects stimulus detection

crowding

examples of an observer’s psychological state that affects stimulus detection

boredom, motivation, tiredness, observer’s criterion

examples of an observer’s psychological state that affects stimulus detection

boredom, motivation

example of a physiological factor that affects stimulus detection

random firing or neurons in visual system or noise

what does detectability depend on

neural activity that the stimulus generates and how this activity compares to the noise of visual system that is always present

If stimulus generates stronger signal than noise, is it easier or harder to detect stimulus

easier to detect stimulus

when will detection of noise be difficult

when the neuronal signal a stimulus generates is comparable to the noise of the visual system

if detection of noise is difficult, what does that say about the threshold of the stimulus

closer to threshold

black dot against white background. strong or weak stimulus

strong

grey dot against white background. strong or weak stimulus

weak

define environmental factors that affect noise

Presence of other stimuli that might surround the stimulus can be considered noise because they might mask the presence of the stimulus

definition of lax criterion

tend to say they detect a stimulus when they are not sure

lax criterion is biased towards what

YES responses

definition of strict criterion

tend to say they didn’t detect stimulus if they are unsure and only say yes when they are certain

define noise in the visual system

randomly fluctuating neuronal activity within the observer’s nervous system in the absence of an external stimulus

define external noise

external sources that compete with or mask the target/stimulus

why does noise occur in our visual system

due to the spontaneous firing of photoreceptors, ganglion cells

if we cannot get rid of noise, how do we detect it in our visual system

if stimulus is strong enough (high intensity) it will be detected as distinguishable from background noise

if detection of noise is difficult, what does that say about the stimuli

weak stimuli or stimuli having low intensity

what is the signal detection theory

Help us understand what is going on when we make decisions under conditions of uncertainty

What did Wolfe et al look at

noise from an auditory perspective but same concepts apply to the visual system

if stimulus and noise similar to each other, what does that say about the stimulus

closer to threshold

how does the observer’s biases affect their decision-making when noise is involved

their decision to react to the stimulus or not can be dependent on the motivation to detect stimulus

if an observer is not expecting a phone call in the shower, what criterion will they adapt

strict

if an observer is expecting a phone call in the shower, what criterion will they adapt

lax

According to Signal Detection Theory, why can a stimulus never be perceived in complete isolation?

A stimulus is always presented in the presence of internal or external noise

When a stimulus is presented, what does the resulting neural activity represent?

The combined neural activity generated by the noise plus the stimulus

what does this graph show

action potential generated in the auditory system (and at what frequency) in the presence of noise or noise plus a signal

if the signal is very strong, how does the distribution of action potentials look

they do not overlap and the noise signal will be separated from the noise

what is Detectability parameter (d’)?

difference between two distributions (difference between peaks of noise and noise + signal), and is measured from the peaks of distributions

what does large d’ signify

no difficulty detecting a stimulus because two distributions are well separated

what does small d’ signify

neural activities are comparable to one another and distributions overlap between noise and nose + signal

if it is difficult to detect stimulus, how is d’

smaller d’

if d’ is small, why do we need to set criterion level for response

because it allows us to decide if stimulus is present or not

possible outcomes of signal detection theory

hit, false alarm, miss, correct rejection

what outcome is this describing: stimulus corrected identified (left shower and phone rang)

hits

what outcome is this describing: stimulus incorrected identified (left shower and phone didn’t ring)

false alarms

what outcome is this describing: fail to identify stimulus (stay in shower and phone rings)

miss

what outcome is this describing: correctly identified that there are no stimulus present (stay in shower and phone didn’t ring)

correct rejection

When there is overlap between noise and noise + stimulus, what does stimulus detection depend on

observer’s biases

what does this describe: high hits and high false alarms rate

low/lax criterion

what does this describe: high correct rejections but high misses

strict/high criterion

what does this describe: low hits and low false alarms rate

strict/high criterion

what does this describe: low correct rejections and low miss

low/lax criterion

very major criterion for signal detection

very lax, lax, strict, very strict

what criterion is this?
hit: 100. miss: 0. false alarm: 80. correct reject: 20

very lax

what criterion is this?
hit: 98. miss: 2. false alarm: 60. correct reject: 40

lax

what criterion is this?
hit: 45. miss: 55. false alarm: 10. correct reject: 90

strict

what criterion is this?
hit: 20. miss: 80. false alarm: 0. correct reject: 100

very strict

how is the d’ for this?
hit: 20. miss: 80. false alarm: 0. correct reject: 100

large

how is the d’ for this

hit: 100. miss: 0. false alarm: 80. correct reject: 20

smalll

a larger d’ correspond to what criterion

very strict

a smaller d’ correspond to what criterion

very lax

if we are coming up with a drug that treats blindness, what criterion should we use

lax so that there are no misses

if an optometrist is diagnoising a life threatening disease, what criterion should we use

strict so that they are no false alarms

purpose of the ROC curve (what is it plotting for)

to illustrate relationship between hit rate and false alarm rate

If criterion at 15 in ROC curve, how is the hit and false alarm rate

hit and false alarm rate is 0

If criterion at 1 in ROC curve, how is the hit and false alarm rate

hit and false alarm rate is 100