Summary Task 2 - signal detection

TASK 2: SIGNAL DETECTION
SIGNAL & NOISE DISTRIBUTIONS

SIGNAL DETECTION PARADIGM

 If there are 2 discrete states of the world (signal &
noise) that cannot easily be discriminated
 Signal must be detected by human operator
 Two response categories (yes & no) 2 states in the
world + 2 response categories  2 x 2 matrix
 Hits – signal is present & subject says “yes”
 Misses – signal is present & subject says “no”
 False alarms – signal is absent & subject says “yes”
 Correct rejections – signal is absent & subject says “no”
 Perfect answers usually do not occur  there is always data in all 4 cells


Signal  Stimulus presented to the subject, usually very faint


Noise  All other stimuli in the environment, might be mistaken for the signal
 Always present


SIGNAL DETECTION THEORY (SDT)

number of occurences per cell
 Values are expressed as probabilities:
total number of occurences∈a column
 Assumes 2 stages of information processing
 Sensory evidence is aggregated concerning the presence / absence of the signal
 A decision is made about whether this evidence indicates a signal or not
 External stimuli generate neural activity in the brain
 There is more sensory / neural evidence in the brain when a signal is present than
when it is absent
 Neural evidence increases with stimulus intensity
 Evidence variable X
 Enough neural activity, X exceeds critical threshold X C  operator decides yes
 XC = criterion value chosen by the operator – vertical line
 All X values to the right of XC (X > XC)  response: YES
 Too little neural activity  operator decides no
 Noise = average activity when no signals are present, fluctuates
 External noise = from the environment

,  Internal noise = neural responses are noisy, even if stimulus is the same on
each trial
 Total area under the curve = 1
 P ( H )+ P ( M )=1
 P ( FA ) + P (CR ) =1
 X varies considerably even without stimuli being present because of random
variations in environment & operator’s own baseline level of neural firing
 Smaller difference in intensity between signals & noise  bigger error probabilities,
because the amount of variation in X resulting from randomness increases relative to
the amount of energy in the signal
 “Perceptual effect” the subject experiences on each trial can be seen on the x-axis
 Probability distributions tell us what the chances are that a given perception of the
stimulus is due to (N) or to (S + N)

CRITERION SETTINGS & SIGNAL STRENGTH

CRITERION SETTING

 2 main components to decision-making-process: stimulus
strength & criterion
 Subject needs to pick a criterion location along
internal response axis
 Internal response greater than criterion 
response: “yes”
 Internal response less than criterion  response:
“no”
 Can adjust the kind of errors they make by
manipulating their criterion
 Liberal / risky criterion setting – more prone to saying “yes”, many hits but also many false
alarms
 XC places to the left
 Low criterion setting
 Conservative criterion setting – more prone to saying “no”, making few false alarms but
also many misses
 XC placed to the right
 High criterion setitng
 Behaviour is determined by placing decision criterions XC

STIMULUS STRENGTH

 Affects probability density functions in obvious way
 Strong signal will shift S + N curve to the right  internal response strength is
stronger