Transcript Stroop Task

Cognition-emotion interactions
Lecture # 5: October 13, 2004
Important issues to consider
• The nature of the task being performed
• Personal relevance
• Mode (functional) versus Node (spreading activation)
theories of cognition-emotion interactions
• The relationship between depression / anxiety and basic
emotions such as sadness / fear from which research
findings are extrapolated.
• Automatic versus strategic processing (implicit versus
explicit memory).
• Trait versus state variables
Experimental Tasks
•
•
•
•
•
•
Dot localization task
Stroop task and modified Stroop
Lexical decision task
Free recall- explicit memory
Word stem/fragment completion- implicit
Masked Stroop
Dot localization experiment
• Simultaneous display of one threatening word and
one neutral word on the computer screen.
• Anxious subject are quicker to respond to the dot
when it is presented in a location previously
occupied by a threatening word.
• Anxious subjects preferentially attend to cues
related to threat, and interpret ambiguous
information as threatening.
• Inference: subject was more likely to be attending
to the threatening word on the previous trial.
Dot Localization
Stroop task
• Stroop (1935) first demonstrated the effect
in which participants are slowed to name
the ink colour of words that are themselves
the names of colours, if the colour name and
the ink colour are incongruent compared to
a condition in which they are congruent, or
compared to a condition in which the ink of
regular words or colour patches is named.
Stroop Task: Colour naming
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
XXX
Stroop Task: Incongruent
Green
Blue
Red
Black
Blue
Red
Green
Blue
Red
Black
Blue
Red
Green
Red
Red
Blue
Green
Black
Green
Blue
Red
Black
Red
Blue
Green
Red
Green
Blue
Modified (Emotional) Stroop
• Participants must name the colour of ink of
ordinary words, or threat words related to the
source of their anxiety
• For example Foa et al. (1991) had people with
PTSD colour name words related to the source
of their trauma.
– Slowed more to trauma-related words
– Those who coped better showed less interference
Lexical decision task
• Subjects must decide whether words presented are valid
examples of English words, or whether they are nonwords (that follow the rules of pronunciation within the
English language).
• Has been used to show ‘positive priming effects’. For
instance, subjects would be faster to say that ‘doctor’ is
a word after seeing the related word ‘nurse’ than after
seeing the unrelated word ‘forest’.
– Automatic spreading of activation through a semantic network
Personal Relevance
• Depressed subjects selectively recall negative information only
when it has been encoded in relation to themselves. For
example, Bradley & Mathews (1983) found that depressed
subjects recall more negative trait adjectives when they have
been asked whether those adjectives describe themselves,
compared to when they must make the judgment of whether they
describe another person.
• Similarly, Butler & Mathews (1983; 1987) found that anxious
subjects rate the probability of negative events happening to
them in the future higher than non-anxious subjects do, but not
the probability of negative events happening to another person
(Personal Optimism Scale).
Personal Relevance (con’t)
• On the modified Stroop task, subjects are slowed most in their
colour naming performance when to-be-ignored words match
their current concerns:
–
–
–
–
–
–
Physical threats = “disease” or “fatal”
Social threats = “foolish” or “lonely”
Panic disorder = physical symptoms or catastrophic disease
Social phobics = socially threatening words
Eating disorders = food-related words
PTSD = words related to the particular trauma (also, those who coped
better showed less interference).
• Thus, emotionally selective processing is confined to material
that matches individuals’ current concerns (i.e., personal
relevance).
Bower’s Network Theory
• Emotions are represented as nodes in a memory
network, and vary on the basis of the content of the
information attached to that node.
• Information and experiences acquired in a particular
emotional state are connected in a memory network.
• When an emotional representation is activated in
memory, there is a spreading of activation to all other
information that was acquired in the same emotional
state.
Problems with Network / Node Theories
• Difficult to show state-dependent memory effects.
• Mood-incongruent recall has sometimes been found.
• Failure to find effects on lexical decision tasks
A mode theory: Oatley and Johnson-Laird’s
Communicative theory (1988; 1995)
• There are a limited number of basic emotions
representing solutions to problems of adaptation that
have been incorporated into our nervous systems
through evolution because they are functional and have
consequences that are better than acting randomly, or
not acting at all. Emotions are ‘heuristics’.
• The elicitation of emotion imposes a particular mode of
organization on the nervous system, consistent with the
function of that particular emotion. This simplifies and
specializes us to respond to a personally-relevant event
or stimulus in our environment in an adaptive manner.
• Is a goal-relevance theory of emotion… why?
Evidence to support mode theories
• Lexical decision task: Emotionally congruent effects when two
words are presented, but not with single words
– Competition for resources
– Processing priority
• Differential effects of anxiety and depression on measures of
attention and memory
– Selective attention to threatening material is a robust effect
that has been shown across several different tasks.
– Depressed patients recall more negative words related to
themselves, but mood-congruent recall is rarely seen in
studies of anxious subjects.
Evidence to support mode theories
• Anxiety and Depression have differential effects on measures of
implicit and explicit memory
– Completions by anxious patients more likely to match previously
encountered threatening words, while recovered and normals show
reverse trend. When asked to use stem to recall words, no group
differences (Mathews et al., 1989).
– In contrast, depressed patients show greater explicit recall of negative
words, but do not differ from normals in their completions
• Masked Stroop
– Interference effects found only in the group of anxious subjects for threat
words relevant to the concerns of either anxious or depressed subjects.
• Automatic and preattentive processes that are relatively crude and
involve only the classification of stimuli into ‘threat’ and ‘non-threat’
categories (doesn’t this sound like Caccioppo’s Rudimentary
Stimulus Evaluation?)
• Function of fear / anxiety: to identify and avoid danger
– Best served by rapid perceptual encoding of threatening
information, and this is why selective attention and implicit
memory effects are found
• Function of sadness / depression:
– Elaborative processing of negative personal information…
loss of goal… re-evaluation of personal concerns
• Retrospective memory: Previous coping strategies?
Remember events that led up to the demise for
comprehension? Put the loss in context?
 Sadness
Fear 
Assumption re: Emotion and
psychopathology (on a continuum)
• The research presented by Mathews and colleagues assumes a
continuum between basic emotions and emotional disorders.
– Sadness  Depression
Fear  Anxiety
• This is not necessarily warranted given the fact that
psychopathological conditions are characterized by various
cognitive and somatic phenomena that are beyond what is
experienced as part of an emotion episode, from which results are
being extrapolated from.
– For example, depression involves feelings of guilt, hopelessness, loss of
appetite, sleep disturbance among other symptoms.
Counterevidence
• Niedenthal and Setterlund (1994) induced happy and sad moods
by playing music throughout an experimental session (Adagietto
by Mahler; Vivaldi Concerto in C Major).
• Lexical decision task: Happy words, generic positive words, sad
words, generic neutral words, and neutral words.
• When listening to happy music, subjects were faster to identify
happy than sad words, and vice-versa. These effects did not hold
for generic positive and negative words.
• What evidence presented earlier in the lecture does this conflict
with?
• What are the implications of the evidence presented above?
• Baron (1987) brought pairs of people of the same sex in to the laboratory
together for a study in impression formation (one participant was actually a
confederate of the experimenter).
• Subject was ‘randomly’ chosen as an interviewer in a practice interview for
a job as a management trainee.
• While the accomplice ‘studied’ the interview questions, the subject was
induced into either a happy or sad emotional state by receiving false
feedback with regards to performance on a task they were given to perform.
• Interviewer had to ask a set of six prearranged questions and the responses
were prearranged but mixed in valence.
– E.g. ‘What are your most important traits’? Answer: On the positive side, I am
ambitious, reliable, friendly, but on the negative side my friends tell me that I
am stubborn, and I know I am impatient and pretty disorganized”.
• Results: Interviewers induced into a happy mood were more likely say that
they would hire the job candidate, and in general rated them more
positively.
• Interviewers were also asked to recall the things
that the job candidates had said about themselves.
• Interviewers that had been induced into a positive
mood recalled more positive traits about the job
candidate, and fewer negative traits, and vice versa.
• Which phenomenon described earlier in the lecture
does this result contradict ?
Baron (1987)
Emotional vulnerability
• Comes from a ‘Stress-Diathesis’ model of
psychopathology.
• Mathews and colleagues believe that the
tendency to selectively encode threatening cues
when under stress represents the cognitive
substrate for vulnerability to anxiety states.
• MacLeod and Matthews (1988) using the dot localization task
found that high-anxious subjects became more attentive to
threatening words just before an important examination than did
low trait-anxious subjects.
• Thus, only high trait anxiety subjects reacted to stress by
selectively attending to information likely to worsen their
emotional state ???
• Question: is selectively attending to cues that are related to the
source of the threat dysfunctional ? We are talking about traitanxious subjects, not anxiety-disordered subjects.
– According to Salovey and Mayer (1990), some people use emotion (even
fear and anxiety) instrumentally– e.g. to motivate themselves.
• ‘If it weren’t for the last minute, nothing would get done’? 
Strategic control: Overriding automatic
processing biases associated with emotion
• Mathews & Sebastian: Selected students with low or high selfreported fear of snakes and conducted the modified Stroop task.
• Emotion induction: Subjects were shown a small boa in a tank
and were told that they would have to approach it. No
significant differences, in fact a trend in the opposite direction.
Contrary to results presented earlier.
• Pre-attentive detection of threatening stimuli can be opposed by
inattention or selective ignoring at a post-awareness stage.
• Similar to the idea of secondary appraisals (Lazarus)