4/30 Discussion
Jen Vosilla
--
Subscription settings: http://groups.google.com/group/socialneuro780/subscribe?hl=en
Stuart Daman
Camille Barnes
Lindsay Morton
Lindsay C. Morton, MS
Psychology Department
University at Albany
1400 Washington Avenue
Albany, NY 12222
(973) 534-0530
Jenny Perella
beka strock
David Dinwiddie
Jenny Perella
is a good deal of overlap between paper topics this week, I am
especially curious to read others’ responses this week (primarily to
see if others wonder the same things as me, or offer solutions). The
first part of my discussion will be summary, and then I will get to my
discussion points.
This paper described 2 studies that were identical in procedure but
varied in the stimuli used. The goals of the research were to a)
examine the neural correlates associated with racial groups, and b) to
examine both conscious and non-conscious social evaluations and their
associated brain activity. The authors’ primary ROI is the amygdala,
as it has been implicated in “the indirect expression of learned
emotional responses that have been acquired without direct aversive
experience”.
In both studies, participants completed the modern racism scale, the
IAT, an fMRI task, and an eyeblink startle response task. During the
fMRI, participants were shown pictures of either White or Black faces
and were asked to press a button to indicate whether the picture they
were viewing was the same or different from the picture previously
presented. Reaction times were recorded. During the eyeblink startle
response, participants were exposed to a loud burst of unpleasant
white noise while simultaneously viewing either Black or White faces.
An EMG (electromyogram) recorded participants’ magnitude of defensive
eyeblink response by measuring the movement of the muscles below the
eye.
In the first study, participants (n = 14, all white, 7 male) viewed 18
photos (9 Black/9 White) of unfamiliar White/Black faces. The same
stimuli were used during the fMRI task as during the startle response
task. The authors write that Black faces evoke negative emotional
evaluations in Whites, which is presumably due to societal norms and
individual experience. Because of this negative emotional evaluation,
Phelps et al hypothesize that there should be greater amygdala
activity when viewing Black compared to White faces. Similarly, they
hypothesized a greater eyeblink startle response to Black compared to
White faces, and this too should correlate with greater amygdala
activity. It was further expected that the results of the IAT would
correlate with amygdala activity.
Results of Study 1 showed that there was substantial between-subjects
variability such that there was no overall significant amygdala
activity, though the authors point out that the majority of subjects
(8) showed greater amygdala activation when viewing Black faces… of
course, this means that the ‘minority’ (6) did not show this effect,
or showed greater amygdala activation when viewing White faces.
However, responses on the IAT were as expected: participants were
slower to respond to Black/good + White/bad pairings than White/good +
Black/bad pairings, and the authors argue that this suggests an
“unconscious anti-Black, pro-White evaluation”. These IAT results
correlated positively with amygdala activity. Specifically, IAT
correlated with the left amygdala, a region extending from the right
amygdala to the inferior insular cortex, and part of the anterior
cingulate. Eyeblink startle responses showed a greater defensive
response to the noise when viewing Black faces, and this also
correlated with brain activity, specifically in the left amygdala and
2 regions within the superior insular cortex. There was no correlation
between brain amygdala activity and the modern racism scale, on which
all participants expressed pro-Black attitudes.
As I mentioned before, the procedure in the 2nd study was identical to
that of the first, expect subjects viewed familiar/well-liked White/
Black faces instead of unfamiliar. The authors hypothesize that if the
results in Study 1 were due to a learned negative evaluation of
Blacks, then the patterns of fMRI, eyeblink startle responses, and IAT
results seen in Study 1 should disappear when the stimuli represent
familiar, well-liked Blacks (and Whites). In this study, N = 13.
Results revealed that the same pattern of results emerged on the
Modern Racism Scale (participants expressed pro-Black attitudes) and
on the IAT (participants were quicker to respond to White vs. Black
faces). However, there were no significant differences in eyeblink
startle responses to White vs. Black faces, no pattern of amygdala
activity, and no correlation between amygdala activity and any other
measure of racial bias.
Now to my discussion. First, I wondered how the authors were going to
justify their results as a conclusive indication of negative Black
evaluations instead of simply a factor of familiarity or in-group/out-
group bias. In their discussion they do address this issue, but flat-
out say “we cannot distinguish responses due purely to familiarity
from those due to positive evaluation, independent of familiarity”,
though they argue that the IAT results lend support to their theory of
negative evaluations. Here, I would just like to point out that a
priori, the authors expected no differences in the IAT for study 2.
They make up some vague explanation in their discussion that the
reason they found differences and the reason the IAT supports their
theory is that “the labels [in the IAT] do not allow categorization on
dimensions other than race”… if there had been other options, they
wouldn’t have found an effect. I did wonder about their stance on in-
group/out-group bias, and by the end of the discussion felt they were
suggesting that race was a specific form of in-group/out-group bias,
and that results may generalize (though they never overtly said this).
This brings me to my second issue: the IAT. The authors (as do other
researchers who use the IAT) purport that slower reaction times on the
Black/good + White/bad compared to the Black/bad + White/good indicate
an unconscious anti-Black, pro-White evaluation. I disagree. Maybe I
am just getting hung up on the word choice of ‘evaluation’, but I
think it is more likely that the IAT is an indicator of the strength
of the cognitive associations between a group and a stereotype. Though
the authors do acknowledge that ‘unconscious racial bias’ on the IAT
doesn’t mean that this bias is consciously embraced, I think that
talking about this unconscious association as an ‘evaluation’ and as
‘anti-Black, pro-White’ sends the wrong message.
Third, I wondered why these studies only examined White participants.
Wouldn’t it be more informative to include both races when doing a
study on racial perception? The authors do acknowledge this in a
sentence or 2 in the discussion, but offer no explanation for why they
failed to address this question in their own studies.
Overall, I thought the study was well-done and the fMRI methodology
was great (I didn’t write about this in detail, though). However, I
had some problems with they way the authors conceptualized and
theorized, and also had some problems with the conclusions they drew
from their studies.
Stuart Daman
The paper by Amodio, Harmon-Jones, Devine, Curtin, Hartley, & Covert
(2004) was pretty straightforward and well conducted. Eddie Harmon-
Jones is a credible guy in his domain (neuropsych), including hot
topics like emotion and motivation. David Amodio also does a lot of
neural mechanism research on similar topics. Patricia Devine can
probably be safely called one of the pioneers of stereotyping/
prejudice research, and has been around a few years longer than the
other two guys. All three of these people are great writers, and I've
always been happy to read their papers in comparison to others. The
other three, I admit, I'd never heard of until I read this paper.
Anyways, in this paper, they were interested in isolating temporal
aspects of race-biased responses. They adopt a dual system theory of
behavior in the face of conflicting behavioral tendencies. The first
system, the conflict-detection system, monitors behavior for
conflicts. Neuropsychological evidence suggests that this is indicated
by activity in the ACC. The second system, the regulatory system,
implements intended responses while inhibiting unintended ones. Take a
moment to consider how these would go in situations in which you
(White people) might behave in a race-biased fashion. They basically
sought to see how this dual system theory might line up with brain
activity as measured by EEG. They had participants complete a task
that is somewhat similar to the IAT. In this task, people are primed
with Black or White faces then shown a picture of a handgun or hand
tool, which they are to classify as a handgun or hand tool. The
general idea is that subconscious associations between Blacks and guns
result in faster responses to guns when primed with Black faces and
more errors of classifying tools as guns when primed with Black faces
(when task is completed by White people).
They had a little over 30 White, right-handed females do the task
while having EEGs record brain activity. They chose EEG because they
were particularly interested in the temporal nature of the conflict
detection and resolution process. They found the basic stuff they were
looking for, just like the expectations mentioned above. The part they
were really interested in was the ERN (error-related negativity). This
part of the EEG record comes from the ACC and occurs when a conflict
is detected, as well as being larger when an error is made (i.e.
conflict is not resolved as expected; unintentional behavior carried
out). In this study, they emphasized quick responses to participants,
forcing them to respond quickly, which made them more vulnerable to
mistakes and hopefully generating more variability in ERNs. This was
the case. ERNs were larger when tools were mistakenly classified as
guns after Black face primes than White, suggesting higher conflict
detection (correct responses did not differ). Also, when guns were
erroneously classified as tools, ERNs were similar, regardless of face
prime (because conflict was not due to a race bias). Additional
correlational analyses showed that greater ACC activity in response to
race-biased response conflict was associated with a particular pattern
of control.
They suggest that "despite an enhanced signal for control to prevent
race bias, the successful implementation of control may be impeded by
lack of time or cognitive resources” (p. 93) In other words, by
pressuring participants to respond quickly, the regulatory system was
not able to inhibit the unintended behavior, resulting in the race-
biased response. This is important because their results suggest that
the detection process is below awareness (i.e. automatic) and very
fast, whereas earlier theorizing said it relied on conscious awareness
of the biased behavior.
I don't really have a lot of feedback about this paper. It was well
written and carried out as far as I can tell. They are careful not to
explicitly say that the people are racist, sticking to terms like
"unconscious associations" and the like (unlike Jenny's paper--"pro-
white" and "anti-black"). I think it would be interesting to show
similar effects in other situations than Black vs White faces though.
I'm also curious if there would be any differences in sex (both
targets and participants) and lefties. They chose to not include these
groups basically to eliminate that extra variance, but never talk
about what could be different--would they simply need a larger sample
to detect differences if this precaution was not taken? Their used EEG
for its temporal resolution, but still refer to specific brain areas
(ACC), citing other papers that show this evidence. Is that good
enough?
I agree with Jenny on most of her points regarding the IAT, its
definitely got its problems. I can talk more about the task used in
this study in class, but I think it eliminates some of those problems
fairly well, although others certainly still remain.
Lindsay Morton
It seems that this week’s readings largely focused on the amygdala, though I was happy that the Amodio et al. (2004) article discussed the ACC, which intuitively seemed like a relevant area when discussing race bias and stereotypes. Phelps, Cannistraci, and Cunningham (2003) provide a succinct overview of past research findings concerning the amygdala with a great definition of it as “a small, almond-shaped structure in the medial temporal lobe that is primarily known for its role in emotional learning and memory” (pp. 203). They also cover the research that has shown that the amygdala is typically linked with automatic responses, an important point considering that most of this week’s articles addressed automatic, implicit processes.
Although their literature review of different findings related to amygdala activation in response to race evaluations is quite good, Phelps et al. (2003) should probably be titled, “Read with extreme caution: We have the lowest N of any study you’ve read yet.” Basically, the researchers included one amygdala-lesion patient and a total of 2 controls to evaluate whether or not the amygdala plays a role in explicit racism and implicit racial associations. To be fair, this was done in an effort to support past imaging data, but it is reported separately for I have no idea what reason. The findings showed that the amygdala does not play a role in responses to the Modern Racism Scale (conscious, self-report measure on which high scores suggest anti-Black beliefs and attitudes) or to the IAT, which utilized Black and White male faces in this study. This is concluded because there was no difference between the patient’s responses and the other 2 participants. Like I said, interpret at your own risk.
Phelps et al. (2003) do touch upon the role of the FFA in race evaluations and how the FFA is used more often with in-group faces, which is assumed to evidence familiarity and “expertise” in own race face evaluations. This brings me to a point that Jenny mentioned – namely, in many of these studies, differential amygdala activation in the perception of Black and White faces is taken as evidence of racism. The one problem that stares me in the face as a social psychologist is familiarity with the majority (White) and the minority (in this case, Black). It seems that the majority of these studies never give participants a simple measure of how often they interact with individuals of other ethnicities and what types of experiences they have had with them. It should be noted that the Wheeler and Fiske (2005) has participants list the first 20 acquaintances that come to mind and the ethnicities of each, yet they don’t discuss it further than that (whether due to non-significance or their disinterest, it is not apparent). Their may be work out there that shows that familiarity doesn’t play a role. I’ll admit that I just don’t know, but I still wish that it was an issue discussed or touched upon by the authors.
Switching gears a bit - the Wheeler and Fiske (2005) article was designed to test how different social processing goals impact amygdala activation in response to Black and White faces. The main reason is that although stereotype knowledge and prejudice varies between individuals, work has shown that a person’s current processing goals can also impact category-based responses. In the first neural imaging study (N = 7), it was found that social categorization by age resulted in greater activation, whereas social individuation (as indexed by vegetable preference) resulted in decreased amygdala activation. Activation was in comparison to a neutral dot search task in the faces. In experiment 2, participants (N = 42) performed a lexical decision task after priming with stereotype relevant or irrelevant words. It was unclear how these words were derived and from my reading may be subject to the region’s specific prejudices. For example, “loyal” is not specifically a racially focused word in my opinion so I wish they would have included more about that. On the note of results, it was found that only after engaging in a social categorization of Black faces, rather than a social individuation or neutral search task did response times to stereotype-relevant words show greater speeds.
One of the more interesting studies for this week was Lieberman, Hairi, Jarcho, Eisenberger, and Bookheimer (2005). In contrast to past work, this study found that both Black and White participants showed greater amygdala activation in response to White faces on a perceptual processing task but not a verbal processing task. Somehow these researchers failed to report their sample size, which I would have liked to see so I could base how strongly I should consider these results. At the same time, the authors highlighted that this pattern of activation may suggest that both Black and White individuals are aware of the negative stereotypes towards Blacks in American culture.
Jen Vosilla
This week I focused on the Correll, Urland & Ito (2006) article. This study looked at event-related potentials and the role of threat perception (P200 response) and cognitive control (N200 response) in the decision to shoot targets either Black or White targets. The authors believe that due to racial stereotype linking Blacks to danger, a Black target automatically activates the idea of danger and predisposition to shoot. The term shooter bias is used throughout the paper refers to the decision to shoot armed targets more quickly and to not shoot unarmed targets less quickly if they are Black compared to if they are White.
The study included 40 right-handed participants. ERP data was collected while participants played a video game consisting of 80 randomly ordered trials. After a fixation cross, a target person appeared and participants had 850ms to press a button indicating the decision to shoot or another to indicate the decision not to shoot. The top three performers received $30, $15 & $10 that was calculated through the use of a point system. Also, participants (31 of them) completed a questionnaire assessing their perception of the personal and cultural stereotype associating Blacks with aggression, violence and danger. The authors hypothesized greater P200s and smaller N200s when participants were making shoot/don’t shoot decisions with Black relative to White targets.
The results include the following:
· Significant interaction of Target Race and Object Type supporting shooter bias (shot armed Black targets and didn’t shoot non-armed White targets the quickest). However, Race by Object interaction was not significant in the P200 or N200
· Larger P200s and Smaller N200s for Blacks relative to Whites
· In the P200 and N200, responses to unarmed Whites were significantly different than responses to all other targets and there was NO significant difference between responses to unarmed Blacks compared to armed Blacks and Whites
· Cultural stereotypes predicted both the early ERP (P200 and N200) racial differentiation and racial bias in the game
Overall, it was a fascinating study and I thought there was greater ecological validity in using a video game for the stimulus. However, in future research I would suggest giving the questionnaire (which needs to be described better, as it is very vague) to all the participants (only gave it to 31 out of 40), and assessing stereotypes implicitly (responses could be affected by social desirability). The significant relationships between the questionnaire and ERP activity were only for questions regarding cultural stereotypes, but more interesting would be whether personal stereotype can predict ERP activity and behavioral bias. An implicit measure of personal stereotype would be more appropriate. The authors also only got demographic information on the 31 participants who completed the questionnaire and failed to address whether racial differences in ERP activity might exist. Also, it would have been interesting to use both Black and White participants and compare differences in ERP activity and shooter bias.