NEURAL AND BEHAVIORAL AGE DIFFERENCES IN EMOTIONAL BIASES A Thesis Submitted to the Faculty in partial fulfillment of the requirements for the degree of Master of Arts in Experimental Psychology by Kristin Flanary DARTMOUTH COLLEGE Hanover, New Hampshire May 2011 Examining Committee: Catherine J. Norris, Ph.D. (Chair) Paul J. Whalen, Ph.D. Jeff T. Larsen, Ph.D. _____________________ Brian W. Pogue, Ph.D. Dean of Graduate Studies
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
NEURAL AND BEHAVIORAL AGE DIFFERENCES IN EMOTIONAL BIASES
A Thesis
Submitted to the Faculty
in partial fulfillment of the requirements for the
degree of
Master of Arts
in
Experimental Psychology
by
Kristin Flanary
DARTMOUTH COLLEGE
Hanover, New Hampshire
May 2011
Examining Committee:
Catherine J. Norris, Ph.D. (Chair)
Paul J. Whalen, Ph.D.
Jeff T. Larsen, Ph.D.
_____________________ Brian W. Pogue, Ph.D. Dean of Graduate Studies
Abstract Research has demonstrated a “positivity effect” in aging, such that older adults remember
and attend to more positive than negative information, compared to their younger
counterparts. The underlying mechanisms of this effect, however, remain largely
unknown. One possibility is that the positivity effect arises from changes in the
negativity bias (the tendency for negativity to outweigh positivity at high levels of
emotional input) or the positivity offset (the tendency for positivity to outweigh negativity
at low levels of emotional input) over the lifespan. The current study aimed to compare
age differences in these emotional biases in order to assess whether the positivity effect is
related to a decreased negativity bias, an increased positivity offset, or both. To do this,
we examined younger (18-25) and older (65-80) females’ behavioral and neural
responses to pictures ranging from neutral and mildly emotional to moderately and
extremely emotional in content. The age groups demonstrated a similar negativity bias in
picture ratings, but older adults showed a larger positivity offset, which was driven by
higher positive and lower negative ratings of the mildly emotional pictures. Older adults
also took more time to rate the mildly unpleasant pictures, and the longer they took to
rate these pictures, the more positive they reported feeling about them. Neural responses
to the emotional pictures appeared to be characterized by prefrontal control regions acting
to down-regulate negative and up-regulate positive emotional responses to both mildly
and very emotional pictures. Results suggest that older adults use multiple emotion
regulation strategies with varying degrees of effectiveness, but appear to successfully
regulate their responses to mildly emotional stimuli in order to more fully appreciate the
simple things in life.
ii
Acknowledgments
First and foremost, many thanks are due to my advisor, Catherine Norris, PhD, for
the generosity, patience, support, and instruction she has extended throughout my time at
Dartmouth. I will always be grateful for all that I have learned from you. I would also
like to thank the members of my committee, Paul Whalen, PhD, and Jeff Larsen, PhD,
whose invaluable assistance, guidance, and support I have truly appreciated.
Thanks to the members of the Norris Lab for three great years of invigorating
discussions and patient hours of data entry and reduction. Special thanks are due to the
many people who have provided technical assistance and helpful discussions regarding
data analysis and interpretation of these data over the last three years: Zachary
Ingbretsen, Lisa Sprute, Ross O’Hara, Dylan Wagner, Kristina Caudle, and especially
Katie Powers, who has been my biggest cheerleader.
Many of my fellow graduate students have made the last three years remarkable
and full of smiles: John Kingsbury, Ross O’Hara, Megan Roberts, Lisa Sprute, Kristina
Caudle, Andrea Robinson, Katie Powers, and Robert Chavez. They have all been
amazing classmates and confidantes, and I am honored to call them friends.
Thanks to my parents, Phil and Eileen Wood, and my family, Ron, Candace,
Scott, and Shane for always believing in me and supporting my every decision. Most of
all, I would like to thank my husband, William Flanary, who has been my biggest
supporter and who reminds me daily why life is so much fun. I can’t imagine a better
partner to stand by my side.
iii
Table of Contents
Introduction ....................................................................................................................... 1 Methods .............................................................................................................................. 6 Participants ............................................................................................................... 6 Materials .................................................................................................................. 7 Procedure ................................................................................................................. 8 Behavioral Task ............................................................................................ 8 fMRI Image Acquisition .......................................................................................... 9 fMRI Preprocessing ................................................................................................. 9 Results .............................................................................................................................. 10 Emotional Biases in Picture Ratings ...................................................................... 10 Emotional Biases in Brain Function ...................................................................... 15 Discussion......................................................................................................................... 17 Emotion Regulation and Emotional Biases in Older Adults ................................. 20 Alternative Explanations ........................................................................................ 23 Limitations ............................................................................................................. 25 Conclusions and Future Research .......................................................................... 26 References ........................................................................................................................ 28 Tables ............................................................................................................................... 35 Figures .............................................................................................................................. 39
iv
List of Tables Table 1: Group activations for the mildly pleasant – mildly unpleasant contrast ............ 35 Table 2: Group activations for the mildly pleasant – mildly unpleasant contrast ............ 36 Table 3: Group activations for the older positivity offset – younger positivity offset
contrast .......................................................................................................................... 37 Table 4: Group activations for the older negativity bias – younger negativity bias contrast
Figure 1: Emotional biases predicted by the Evaluative Space Model. ............................ 39 Figure 2: Younger and older adults’ ratings of neutral and mildly, moderately, and
extremely emotional pictures. ....................................................................................... 40 Figure 3: Standardized reaction times to mildly unpleasant and mildly pleasant pictures
by age group. ................................................................................................................. 41 Figure 4: Negative and positive ratings of mildly unpleasant and mildly pleasant pictures
by age group. ................................................................................................................. 42 Figure 5: Z-scored response times as predictors of positive ratings of mildly unpleasant
pictures in older and younger adults. ............................................................................ 43 Figure 6: Figure 7: Figure 8: Figure 9:
vi
Introduction
American society often paints a dismal picture of aging, with images of
physiological systems shutting down, memories fading, and life ending. Despite
declining physical and cognitive functioning in later life, however, emotional functions
seem to remain intact, or even improve, with age. Across the life span, goals and
motivations shift, with younger adults focusing on broadening their horizons and seeking
new life experiences, and older adults focusing on strengthening close relationships and
matched on normative arousal and extremity (i.e., distance from the midpoint of the
normative valence rating scale), and all categories were matched on RGB values, hue,
saturation, luminance, and spatial complexity.
Procedure
Once situated in the scanner, participants were instructed to view each picture for
the entire time it was displayed and think about whether it was positive or negative.
Pictures appeared in one of two predetermined pseudorandom orders, counterbalanced
across participants. An angled mirror attached to the head coil allowed participants to
view the pictures projected onto a panel at the end of the scanner bore; a PC using E-
Prime version 2.0 Professional software (Psychology Software Tools, Inc., Pittsburgh,
PA) presented the pictures. Each trial consisted of a white fixation cross displayed on a
black background for 1 s, followed by a picture displayed for 4 s; jittered intertrial
intervals consisting of a black background ranged from 1 to 13 s (mean = 5 s), which
allowed for deconvolution of the hemodynamic response.
Behavioral Task. Participants viewed each picture again outside of the scanner,
this time rating how they felt about each picture using the Evaluative Space Grid (ESG;
Larsen, Norris, McGraw, Hawkley & Cacioppo, 2009). The ESG is a 5x5 grid that
allows for independent assessments of negativity and positivity. Negativity is measured
on the y-axis and positivity on the x-axis, with ratings on each axis ranging from not at all
negative/positive (coded as 0) to very negative/positive (4). Participants used the grid to
select the cell that most appropriately represented their positive and negative feelings
about each picture. A PC using E-Prime version 2.0 Professional software (Psychology
Software Tools, Inc., Pittsburgh, PA) presented the pictures and recorded negative
8
ratings, positive ratings and response times for each picture. On each trial, the ESG
appeared immediately below the picture. Pictures appeared in random order at a rate
controlled by participants, such that each picture remained visible until participants made
a rating.
fMRI Image Acquisition
Functional magnetic resonance images were acquired on a 3.0 Tesla Philips
Achieva Intera scanner (Philips Medical Systems, Bothell, WA), with an eight-channel
SENSE head coil. Functional T2*-weighted echo-planar images were acquired in six runs
with 36 contiguous 3.5mm axial interleaved slices with a 0.5mm gap (TR = 2000 ms, TE
= 35 ms, flip angle = 90°, field of view = 24 cm, 80x80 matrix size, fat suppressed).
High-resolution T1-weighted anatomical images were acquired in 160 1-mm sagittal
slices (TR = 9.8 ms, TE = 4.6 ms, flip angle = 8°, field of view = 24 cm).
fMRI Preprocessing
We used AFNI software (Cox, 1996) for processing and analysis of fMRI data.
For each participant, the data were despiked, corrected for slice timing, realigned using a
six-parameter, rigid-body transformation, and converted to percent signal change.
Individual-participant deconvolution analyses then removed variance due to motion as
well as constant and linear trends, and the voxel-wise standard deviation of the residual
signal was calculated in order to assess the signal-to-noise ratio. Output from the
deconvolution analyses underwent spatial smoothing using a 6-mm FWHM Gaussian
kernel.
A second individual-participant deconvolution analysis isolated contributions of
each condition to the fMRI signal by generating impulse response functions (IRFs) of the
9
signal for each condition of a 2 (valence: unpleasant, pleasant) x 2 (extremity: mild,
extreme) design, with a filler condition comprised of responses to the remaining
pictures2. The analysis used a sine function, with 15 regressors per condition (i.e., one
regressor for each second of the estimated hemodynamic response) to model the
hemodynamic response to each condition. Each participant’s data were then normalized
to Talairach stereotaxic coordinate space (Talairach & Tournoux, 1988), and a measure
of the area under the curve (AUC; seconds 5-10) was calculated for use in group analysis;
each condition’s AUC was used as a measure of neural activation to that condition. All
group analyses were thresholded at p < .001, uncorrected, with a minimum cluster size of
50 voxels.
Results
Emotional Biases in Picture Ratings
The positivity offset refers to the tendency for positivity to outweigh negativity at
low levels of input to the affect system. Conversely, the negativity bias refers to the
tendency for negativity to outweigh positivity at high levels of input (Figure 1). We
examined the positivity offset and negativity bias by comparing positive ratings of
pleasant pictures with negative ratings of unpleasant pictures (i.e., “dominant ratings”).
Specifically, we sought to determine whether older and younger adults differed in the
positivity offset and negativity bias. To this end, we conducted a 2 (Age: younger, older)
x 2 (Picture Valence: unpleasant, pleasant) x 4 (Picture Extremity: neutral, mild,
2 The model was limited to these conditions in order to increase power in group analyses designed specifically to investigate neural responses underlying the positivity offset and negativity bias. The mildly emotional pictures were used instead of neutral pictures to investigate the positivity offset, as the positivity offset must be calculated by comparing two conditions (i.e., pleasant – unpleasant) at relatively low levels of emotional input. The mildly emotional pictures offered two such conditions (mildly pleasant, mildly unpleasant), while the neutral pictures did not.
10
moderate, extreme) repeated measures general linear model (GLM) analysis on dominant
ratings (i.e., negative ratings of unpleasant pictures and positive ratings of pleasant
pictures).3 This analysis allowed several manipulation checks in addition to addressing
our primary question of interest.
First, there was a main effect of Extremity, F(3, 33) = 280.50, p < .001, ηp2 = .96;
this was a manipulation check that showed that dominant ratings increased as a function
of the extremity of the picture (i.e., distance from the midpoint of the normative valence
scale); all ps < .001. There was no main effect of Age, F(1, 35) < 1, p = .99, ηp2 < .001,
indicating that older and younger adults did not differ in their overall ratings of the
pictures. There was also no main effect of Valence, F(1, 35) < 1, p = .41, ηp2 = .02,
demonstrating that participants rated unpleasant and pleasant pictures equally intensely.
More importantly, we examined whether participants demonstrated a positivity
offset and a negativity bias in response to the pictures. A significant Valence x Extremity
interaction, F(3, 33) = 50.31, p < .001, ηp2 =.82, verified that participants responded with
both a positivity offset and a negativity bias. Pairwise comparisons revealed that
participants showed higher positive (M = 1.07, SE = .09) than negative ratings (M = .57,
SE = .06) of neutral pictures, p < .001, ηp2 = .49, as well as higher dominant ratings of
mildly pleasant (M = 1.91, SE = .08) than of mildly unpleasant (M = 1.65, SE = .08)
pictures, p =.02, ηp2 = .16 (i.e., both findings are consistent with the functioning of a
positivity offset). They also showed higher dominant ratings of moderately unpleasant
(M = 2.72, SE = .07) than moderately pleasant (M = 2.28, SE = .09) pictures, p < .001, ηp2
= .41, and higher dominant ratings of extremely unpleasant (M = 3.28, SE = .07) than of
3 Positive and negative ratings of neutral pictures were used as a measure of unpleasant and pleasant valence.
11
extremely pleasant (M = 2.70, SE = .09) pictures, p < .001, ηp2 = .56 (i.e., both findings
are consistent with the functioning of a negativity bias; Figure 2).
Finally, our primary question of interest concerned whether younger and older
adults differ with respect to the positivity offset and negativity bias. A significant Age x
Valence x Extremity interaction, F(3, 33) = 9.53, p < .001, ηp2 = .46, revealed an age
difference in these emotional biases. Compared to younger adults, older adults rated
neutral (older: M = .46, SE = .09; younger: M = .70, SE = .08) and mildly unpleasant
pictures (older: M = 1.51, SE = .11; younger: M = 1.79, SE = .11), as marginally less
negative, ps < .07, and mildly pleasant pictures as more positive (older: M = 2.10, SE =
.11; younger: M = 1.72, SE = .11), p = .02. These ratings show a larger positivity offset
among the older adults. Specifically, younger adults demonstrated a positivity offset only
in response to neutral pictures, but older adults showed a positivity offset in response to
both neutral and mildly emotional pictures. However, there were no age differences in
responses to moderately or extremely emotional pictures, ps > .21, resulting in no age
differences in the negativity bias. In sum, the age groups did not differ in the negativity
bias, but older adults showed a larger and extended positivity offset as compared to
younger adults (Figure 2).
Since this analysis revealed significant age differences involving only the mildly
emotional pictures, further analyses focused on that category. We next sought to explore
the nature of the processes underlying the extended positivity offset in older adults.
Specifically, we wondered whether this effect appeared to be the result of more automatic
or effortful processes. For example, if older adults’ extended positivity offset occurs by
relatively automatic processes, then older adults should rate the mildly emotional pictures
12
as quickly, or even more quickly, than younger adults. However, if older adults’
extended positivity offset occurs by more effortful processes, then they should rate the
mildly emotional pictures more slowly than younger adults, since longer response times
may suggest that effortful processes have been recruited.
To investigate these questions, we conducted a 2 (Age: younger, older) x 2
scored response times (RTs), trimmed to exclude the fastest and slowest 10% of
responses (Bush, Hess, & Wolford, 1993). This analysis yielded no main effects of Age,
F(1, 35) = .62, p = .44, ηp2 = .02, or Valence, F(1, 35) = 1.74, p = .20, ηp
2 = .05, but did
reveal a significant Age x Valence interaction, F(1, 35) = 5.95, p = .02, ηp2 = .15.
Pairwise comparisons showed that older adults (M = .24, SE = .05) took longer than
younger adults (M = .06, SE = .05) to rate the mildly unpleasant pictures, p = .02, but RTs
to the mildly pleasant pictures did not differ by age (Molder = .02, SEolder = .06; Myounger =
.12, SEyounger = .06), p = .20. Furthermore, older adults took longer to rate mildly
unpleasant (M = .24, SE = .05) compared to mildly pleasant pictures (M = .02, SE = .06),
p = .01, while younger adults showed no differences in RTs, p = .43 (Figure 3). In sum,
older adults took longer to respond to mildly unpleasant pictures than did younger adults.
One possible explanation for this pattern of RTs is that older adults were
distracted by the unpleasant pictures, and subsequently took longer to rate them.
However, this explanation seems unlikely given previous research demonstrating greater
attention to and memory for positive (i.e., pleasant) than negative (i.e., unpleasant)
stimuli in older adults (e.g., Mather & Carstensen, 2005). A second possibility is that this
pattern reflects an effortful process on the part of older adults to feel more favorably
13
toward mildly emotional stimuli. If this is accurate, then older adults should a) assign
lower negative and/or higher positive ratings to mildly emotional pictures than younger
adults, and b) rate mildly emotional pictures as increasingly more positive and/or less
negative as picture viewing time increases.
To examine these hypotheses, we first looked at whether older adults rated mildly
emotional pictures less negatively and/or more positively. Four independent-samples t-
tests were conducted to compare older and younger adults’ positive and negative ratings
of mildly unpleasant and mildly pleasant pictures. 4 These analyses revealed that older
adults rated mildly pleasant pictures as less negative (M = .26, SE = .06) and more
positive (M = 2.10, SE = .13) than younger adults (negative: M = .47, SE = .06; positive:
M = 1.72, SE = .09), ts(35) = -2.41 and 2.41, respectively, ps = .02. There was also a
trend for older adults to rate mildly unpleasant pictures as less negative than younger
adults (older: M = 1.50, SE = .13; younger: M = 1.79, SE = .08), t(35) = -1.89, p = .07
(Figure 4). Taken together, these data support the conclusion that older adults did indeed
respond more favorably to mildly emotional pictures than younger adults.
Finally, we examined whether older adults’ more positive and less negative
responses to the mildly emotional pictures may be the result of effortful processes. As
outlined above, we expected that if older adults were engaging effortful self-control
processes, we should see that longer RTs predict better regulation (i.e., more positive,
less negative ratings) in older but not younger adults. Since older adults showed longer
RTs only to the mildly unpleasant pictures, analyses focused on those pictures. Negative
and positive ratings of the mildly unpleasant pictures were separately regressed on Age
4 This is in contrast to the omnibus GLM, which included only dominant ratings (i.e., negative ratings of unpleasant pictures and positive ratings of pleasant pictures).
14
(younger, older), RTs to the mildly unpleasant pictures (RTs), and the interaction (Age x
RT). For negative ratings, the model was not significant, R2 = .11, F(3, 33) = 1.35, p =
.28; that is, none of the variables predicted negative ratings of the mildly unpleasant
pictures, ps > .08. For positive ratings, however, the model was significant, R2 = .28,
F(3, 33) = 4.37, p = .01. There was a trend for RTs to predict positive ratings, b = .47,
t(33) = 2.03, p = .05. However, this was qualified by a significant Age x RT interaction,
b = -.69, t(33) = -2.94, p < .01. Bivariate correlations revealed that longer RTs predicted
higher positive ratings of the mildly unpleasant pictures among older, r = .58, p = .01, but
not younger adults, r = -.20, p = .42. That is, the more time that older adults took to rate
the mildly unpleasant pictures, the more positively they rated them (Figure 5). Taken
together, these data suggest that older adults may have effortfully attempted to feel more
positive when viewing mildly unpleasant pictures.
Emotional Biases in Brain Function
We next examined whether participants exhibited a positivity offset and
negativity bias in neural activity by comparing neural responses to pleasant pictures with
neural responses to unpleasant pictures across all participants. To examine the positivity
offset, a whole-brain paired-sample t-test compared neural activation to mildly pleasant
pictures with activation to mildly unpleasant pictures (i.e., mildly pleasant – mildly
unpleasant). No regions demonstrated greater activation to the mildly pleasant than the
mildly unpleasant pictures. However, several regions of the prefrontal cortex
demonstrated greater activation to the mildly unpleasant pictures (Table 1, Figure 6).
These regions included bilateral clusters in the dorsomedial prefrontal cortex (dmPFC;
BA 9), right ventrolateral prefrontal cortex (vlPFC; BA 47), left medial prefrontal cortex
15
(mPFC; BA 9), right supplementary motor area (SMA; BA 6), and left frontal eye fields
(BA 8).
To examine the negativity bias, a second whole-brain paired-sample t-test
compared activation to very unpleasant pictures with activation to very pleasant pictures
(i.e., very unpleasant – very pleasant). No regions demonstrated greater activation to the
very unpleasant than the very pleasant pictures. However, clusters in the left medial
prefrontal cortex (mPFC; BA 10) and left parahippocampal gyrus demonstrated greater
activation to the very pleasant pictures (Table 2, Figure 7).
As our primary question of interest concerned whether younger and older adults
differ with respect to the positivity offset and negativity bias, we next directly compared
the age groups’ emotional biases in neural activity. To examine age differences in the
positivity offset, we first calculated contrasts comparing neural activation to mildly
pleasant pictures with activation to mildly unpleasant pictures (mildly pleasant – mildly
unpleasant; i.e., a positivity offset) for each individual participant. These contrasts were
then submitted to an independent-samples t-test comparing older and younger adults’
neural positivity offset. This analysis revealed three regions exhibiting significant age
differences (Table 3, Figure 8). Older adults demonstrated a greater positivity offset than
younger adults in the vlPFC (BA 47) and fusiform gyrus. That is, older adults showed
greater activation of these areas in response to the mildly pleasant pictures than the
mildly unpleasant. In contrast, younger adults demonstrated a greater positivity offset in
the subgenual anterior cingulate cortex (sgACC; BA 25). That is, younger adults showed
greater activation of the sgACC to mildly pleasant than unpleasant pictures, while older
adults showed the opposite pattern of activation (Figure 8).
16
To examine age differences in the negativity bias, we calculated contrasts
comparing neural activation to very unpleasant pictures with activation to very pleasant
pictures (very unpleasant – very pleasant; i.e., a negativity bias) for each individual
participant. These contrasts were then submitted to an independent-samples t-test
comparing older and younger adults’ neural negativity bias. Older adults showed a
greater negativity bias in two regions of prefrontal cortex (Table 4, Figure 9). The largest
cluster extended from the left frontal eye fields into the SMA (BA 8 / BA 6); a second
cluster was located in the right dorsolateral prefrontal cortex (dlPFC; BA 9). That is,
older adults showed greater activation of these prefrontal regions to the very unpleasant
pictures than the very pleasant. Younger adults did not show a greater negativity bias
than older adults in any region.
Discussion
In the current study, older adults showed a larger and extended positivity offset in
ratings of emotional pictures compared to younger adults. Both age groups demonstrated
a positivity offset to neutral pictures, as predicted by the ESM; however, older adults
continued to show a positivity offset in response to mildly emotional pictures, while
younger adults did not. This extended positivity offset in older adults was driven by
higher positive and lower negative ratings of the mildly emotional pictures. Older adults
also took more time to rate the mildly unpleasant pictures, and the longer they took to
rate these pictures, the more positive they reported feeling about them. In contrast, there
were no age differences in the negativity bias. Taken together, these findings suggest that
older adults may effortfully regulate their responses to mildly emotional stimuli in order
to feel more favorably about them. That is, older adults may regulate their emotions to
17
feel more positively (and less negatively) toward a wider range of emotional stimuli than
younger adults.
The current study also investigated the neural mechanisms underlying the
positivity offset and negativity bias across the life span. Neural response patterns were
consistent with the behavioral response patterns in that several prefrontal control regions
demonstrated greater activation to the mildly unpleasant pictures than the mildly pleasant,
including the dmPFC, vmPFC, and vlPFC. Previous work has implicated the dmPFC in
the appraisal and experience of negative emotion (Etkin, Egner, & Kalisch, 2011), as well
as in making inferences about one’s feelings (Wager et al., 2008), whereas the vmPFC
and vlPFC have been shown to be involved in the regulation of negative emotion (Etkin
et al., 2011; Wager, Davidson, Hughes, Lindquist, & Ochsner, 2008) and the selection or
inhibition of information (Aron & Poldrack, 2005). These results suggest that the
dmPFC may detect the valence of the mildly unpleasant pictures and signal the need for
the vmPFC and vlPFC to down-regulate any negative emotional responses.
In contrast to this down-regulation of negative emotional responses to the mildly
unpleasant pictures, neural responses to the more extreme pictures suggested that
participants may have been attempting to engage with the very pleasant pictures or even
up-regulate their responses to them. Specifically, the mPFC showed greater activation to
the very pleasant pictures than the very unpleasant. Previous work has demonstrated that
this region is activated when increasing positive emotion (Kim & Hamann, 2007) and is
sensitive to the self-relevance of a stimulus (Kelley et al., 2002). Additionally, the
parahippocampal gyrus, a region involved in emotion perception (Wager et al., 2008),
showed greater activation to the very pleasant pictures than the very unpleasant. Taken
18
together, these results suggest that participants may have utilized the mPFC in an attempt
to upregulate positivity by focusing on the perceptual and self-relevant aspects of the
very pleasant pictures. In sum, neural responses to the emotional pictures appeared to be
characterized by the down-regulation of responses to the mildly unpleasant pictures, and
the up-regulation of responses to the very pleasant pictures.
This approach may have been particularly effective among the older adults, as
they demonstrated these sorts of patterns to a greater extent than did younger adults. For
example, Urry, van Reekum, Johnstone, & Davidson (2009) demonstrated that the
sgACC is involved in the cognitive reappraisal of emotion. In the current study, older
adults (compared to younger adults) showed greater activation of the sgACC in response
to the mildly unpleasant than mildly pleasant pictures, suggesting that older adults may
have been down-regulating negative responses to the mildly unpleasant pictures.
Moreover, older adults showed a greater positivity offset (i.e., greater activation to the
mildly pleasant than mildly unpleasant pictures) in the vlPFC and fusiform gyrus than did
younger adults. As previously discussed, the vlPFC works to generate and regulate
emotional responses (Wager et al., 2008), and the fusiform gyrus is a relatively low-level
perceptual region that is sensitive to emotional content (Norris, Chen, Zhu, Small, &
Cacioppo, 2004). These patterns of vlPFC and fusiform gyrus activity suggest that older
adults may have utilized the vlPFC in an attempt to upregulate positivity, in part by
focusing on the positive aspects of the mildly pleasant pictures. Taken together, these
results suggest that older adults may have been down-regulating their responses to the
mildly unpleasant pictures and up-regulating their responses to the mildly pleasant
pictures to a greater extent than younger adults.
19
Finally, older adults showed a similar pattern in neural responses to the very
Williams, L. M., Brown, K. J., Palmer, D., Liddell, B. J., Kemp, A.H., Olivieri, G.,
…Gordon, E. (2006). The mellow years? Neural basis of improving emotional
stability over age. The Journal of Neuroscience, 26, 6422-6430.
doi:10.1523/JNEUROSCI.0022-06.2006
Wood, S., & Kisley, M. A. (2006). The negativity bias is eliminated in older adults: Age-
related reduction in event-related brain potentials associated with evaluative
categorization. Psychology and Aging, 21, 815-820. doi:10.1037/0882-
7974.21.4.815
35
Table 1 Group activations for the mildly pleasant – mildly unpleasant contrast
Region of activation BA Hemisphere Coordinates Volume (mm3) t-value
x y z Mildy pleasant > Mildly unpleasant None
Mildly unpleasant > Mildly pleasant Medial frontal gyrus 9 R -5 -44 29 200 -4.29 Superior frontal gyrus 8 L 3 -22 49 165 -4.03 Superior frontal gyrus 9 L 7 -50 30 111 -4.05 Superior frontal gyrus 6 R -2 -18 54 110 -4.03 Lentiform nucleus/claustrum/putamen R -27 -12 -3 93 -4.12 Inferior frontal gyrus 47 R -49 -18 2 80 -4.32 Medial frontal gyrus 9 L 3 -52 20 59 -4.15
Note. Clusters listed consist of 50 or more contiguous voxels with a t-threshold of 4.00, p < .001 uncorrected. Coordinates are in Talairach stereotaxic space (RAI format).
36
Table 2 Group activations for the very unpleasant – very pleasant contrast
Region of activation BA Hemisphere Coordinates Volume (mm3) t-value
x y z Very unpleasant > Very pleasant None Very pleasant > Very unpleasant
Superior frontal gyrus 10 R -4 -61 -2 69 -4.06 Parahippocampal gyrus 28 L 19 13 -14 54 -4.21
Note. Clusters listed consist of 50 or more contiguous voxels with a t-threshold of 4.00, p < .001 uncorrected. Coordinates are in Talairach stereotaxic space (RAI format).
37
Table 3 Group activations for the older positivity offset – younger positivity offset contrast
Region of activation BA Hemisphere Coordinates Volume (mm3) t-value
x y z Older > Younger Inferior frontal gyrus 47 R -47 -28 -14 90 4.06 Fusiform gyrus 37 R -59 50 -20 78 4.58
Note. Clusters listed consist of 50 or more contiguous voxels with a t-threshold of 4.05, p < .001 uncorrected. Coordinates are in Talairach stereotaxic space (RAI format).
38
Table 4 Group activations for the older negativity bias – younger negativity bias contrast
Region of activation BA Hemisphere Coordinates Volume (mm3) t-value
x y z Older > Younger Superior frontal gyrus 8 L 2 -38 44 1119 4.23 Precentral gyrus 9 R -42 -16 34 399 4.15
Younger > Older None
Note. Clusters listed consist of 50 or more contiguous voxels with a t-threshold of 4.05, p < .001, uncorrected. Coordinates are in Talairach stereotaxic coordinate space (RAI format).
39
Figure 1. Emotional biases predicted by the Evaluative Space Model. The positivity
offset is reflected by greater positivity than negativity at low levels of emotional input;
the negativity bias is reflected by greater negativity than positivity at high level of
emotional input. Adapted from Norris, Larsen, Crawford & Cacioppo (in press).
40
Figure 2. Younger and older adults’ ratings of neutral and mildly, moderately, and
extremely emotional pictures. Dominant ratings refer to negative ratings of unpleasant
pictures and positive ratings of pleasant pictures. Compared to younger adults, older
adults show a larger positivity offset observed in responses to both neutral and mild
pictures.
41
Figure 3. Standardized reaction times to mildly unpleasant and mildly pleasant pictures
by age group. Older adults responded more slowly than younger adults to mildly
unpleasant pictures; they also responded more slowly to mildly unpleasant than to mildly
pleasant pictures.
42
Figure 4. Negative (a) and positive (b) ratings of mildly unpleasant and mildly pleasant
pictures by age group. Compared to younger adults, older adults rated mildly pleasant
pictures as less negative and more positive, and exhibited a trend to rate mildly
unpleasant pictures as less negative.
43
Figure 5. Z-scored response times (separated into short and long RTs, representing 1 SD
below and above the mean, respectively) as predictors of positive ratings of mildly
unpleasant pictures in older and younger adults. As response times increased, older
adults rated mildly unpleasant pictures more positively, while younger adults did not.