Striatum and Reflected Social Self-Evaluations

1

STRIATUM AND REFLECTED SOCIAL SELF-EVALUATIONS

Supplementary Materials

Method

S1.1 Data Analysis

In order to examine activity recruited by each age group across conditions,two4 (evaluative perspective) x 3 (domain) whole-brain, repeated-measures ANOVAs were conducted for adults and adolescents separately. Planned linear contrasts were conductedfor each age group in order to identify regions specifically recruited during direct self-evaluations, direct other-evaluations, and reflected self-evaluations (relative to control malleability-evaluations) across the three domains. To correct for multiple comparisons, whole-brain, voxel-wise and cluster-extent thresholds were calculated for each repeated-measures ANOVA, resulting in voxel-wise thresholds of p < 0.005 and cluster-extent thresholds of k = 57 and 60 voxels (corresponding to p < 0.05 FWE-corrected) for adult-only and adolescent-only analyses, respectively.

Results

S2.1 Behavioral Data

To investigate age group differences in task performance, mean RTs were entered into a 2 (age group) x 4 (evaluative perspective) x 3 (domain) repeated measures ANOVA. There was a significant main effect of age group, such that adults responded significantly faster than adolescents [Ms = 1.45 and 1.71 seconds for adults and adolescents, respectively; F(1, 35) = 9.54, p = 0.004]. There was also a significant main effect of evaluative perspective [F(2.34, 81.89) = 16.34, p < 0.001], such that participants made significantly faster direct self-evaluations and direct other-evaluations than reflected self-evaluations [t(36) = -2.68, p = 0.011 and t(36) = -2.12, p = 0.041, for direct self- and direct other-evaluations relative to reflected self-evaluations, respectively] and malleability-evaluations were significantly slower than the other three evaluative perspectives[t(36) = -5.91, p < 0.001, t(36) = -4.20, p <0.001, and t(36) = -2.87,p = 0.007, for direct self-, direct other-, and reflected self-evaluations relative to malleability-evaluations, respectively; Ms = 1.53, 1.53, 1.58, and 1.68 seconds for direct self-, direct other-, reflected self-, and malleability-evaluations, respectively]. Finally, there was a significant main effect of domain [F(2, 70) = 11.55, p < 0.001], such that participants made significantly faster academic evaluations than physical or social evaluations [t(36) = -4.39, p < 0.001, t(36) = -3.85, p <0.001, for academic relative to physical and social evaluations, respectively; Ms = 1.53, 1.61, and 1.59 seconds for academic, physical, and social, respectively].

These main effects were qualified by a significant interaction effect between age group and evaluative perspective [F(2.43, 81.89) = 10.71, p < 0.001], such that adults made significantly slower malleability-evaluations than the other three evaluative perspectives [t(18) = -9.33, p < 0.001, t(18) = -8.10, p < 0.001, and t(18) = -5.48, p < 0.001, for direct self-, direct other-, and reflected self-evaluations relative to malleability-evaluations, respectively; Ms = 1.40, 1.37, 1.42, 1.64 seconds for direct self-, direct other-, reflected self-, and malleability-evaluations, respectively], while adolescents made significantly faster direct self-evaluations than reflected self-evaluations [t(17) = -3.30, p = 0.004; Ms = 1.66, 1.69, 1.75, and 1.72 seconds for direct self-, direct other-, reflected self-, and malleability-evaluations, respectively]. There was also a significant interaction effect between evaluative perspective and domain [F(6, 210) = 2.58, p = 0.020], such that participants made significantly faster direct academic other-evaluations than direct social other-evaluations [t(36) = -3.28, p = 0.002; Ms = 14.8 and 1.57 seconds, for direct academic and social other-evaluations, respectively] and faster reflected academic self-evaluations than reflected physical and social self-evaluations [t(36) = -4.60p < 0.001 andt(36) = -3.58 p = 0.001, for reflected academic relative to reflected physical and social self-evaluations, respectively;Ms = 1.47, 1.64, and 1.61 seconds, for reflected academic, physical, and social self-evaluations, respectively]. There was no significant interaction effect between age group and domain [F(2, 70) = 1.87, ns], nor between age group, evaluative perspective, and domain [F(6, 210) = 0.26, ns]. However, to address the potentially confounding effect of age-related RT differences on brain function (suggested by the main effect of age group and the interaction effect between age group and evaluative perspective), trial durations were modeled as RTs.

As an alternative analytical method, mean RTs were entered into four 2 (age group) x 3 (domain) repeatedmeasures ANOVAs, for each evaluative perspective. For direct self-evaluations, there was a significant main effect of age group, such that adults responded significantly faster than adolescents [Ms = 1.40 and 1.66 seconds for adults and adolescents, respectively; F(1, 35) = 10.25, p = 0.003]. There was no significant main effect of domain [F(2, 70) = 1.15, ns] nor significant interaction effect between age group and domain [F(2, 70) = 0.24, ns]. For direct other-evaluations, there was a significant main effect of age group, such that adults responded significantly faster than adolescents [Ms = 1.37 and 1.69 seconds for adults and adolescents, respectively; F(1, 35) = 13.00, p = 0.001]. There was no significant main effect of domain [F(2, 70) = 3.03,ns] nor significant interaction effect between age group and domain [F(2,70) = 0.14,ns]. For reflected self-evaluations, there was a significant main effect of age group, such that adults responded significantly faster than adolescents [Ms = 1.42 and 1.75 seconds for adults and adolescents, respectively; F(1, 35) = 3.10, p < 0.001]. There was also a significant main effect of domain [F(2, 70) = 12.95,p < 0.001], however there was no significant interaction effect between age group and domain [F(2,70) = 059,ns]. For malleability-evaluations, there was no significant main effect of age group [F(1, 35) = 0.82, ns], nor main effect of domain [F(2, 70) = 0.78, ns], nor interaction effect between age group and domain [F(2, 70) = 1.60, ns].

S2.2 Imaging Data

S2.2.1 Neural Activity for Young Adult Self- and Other-Evaluations

Within the adult sample, planned linear contrasts from a 2 (evaluative perspective) x 3 (domain) repeated measures ANOVA were conducted to identify regions specifically recruited by young adults during direct self-evaluations, direct other-evaluations, and reflected self-evaluations (as compared to control malleability-evaluations), in each of the three domains (see Table S2 and Figure S1). Consistent with previous findings, adults recruited significant CMS activity during all self- and other-evaluations, with additional activity in social cognition regions during reflected self-evaluations. Specifically, when making direct academic, physical, and social self-evaluations, adults recruited significant activity in ventral and/or anterior rostral mPFC; rostral and/or subgenual ACC; and mPPC. When making direct academic, physical, and social evaluations about one’s best friend (direct other-evaluations), adults recruited significant activity in ventral, anterior rostral, and/or dorsal mPFC; subgenual or perigenual ACC; and mPPC. When making reflected academic, physical, and social self-evaluations from one’s best friend’s perspective, adults recruited significant activity in ventral and/or anterior rostral mPFC, Prec, and TPJ (except in the physical domain, where TPJ activity was not observed). For any given evaluative perspective, there were very few differences in the patterns observed between domains, and peaks in these regions were highly similar.

S2.2.2 Neural Activity for Early Adolescent Self- and Other-Evaluations

Within the adolescent sample, planned linear contrasts from a 2 (evaluative perspective) x 3 (domain) repeated measures ANOVAwere conducted to identify regions specifically recruited by early adolescents during direct self-evaluations, direct other-evaluations, and reflected self-evaluations (as compared to control malleability-evaluations), in each of the three domains (see Table S2 and Figure S1). In general, adolescents demonstrated similar patterns of CMS activity as adults during self- and other-evaluations; however, overall activity was less robust. There were also more domain-specific effects. When making direct self-evaluations, adolescents recruited significant activity in vmPFC, subgenual and perigenual ACC, and mPPC, but this activity was only associated with academic and physical evaluations. No activity associated with direct social self-evaluations reached significance after correcting for multiple comparisons. When making direct evaluations of one’s best friend (direct other-evaluations), adolescents recruited significant activity in vmPFC during physical and social evaluations. However, no activity associated with direct academic other-evaluations reached significance after correcting for multiple comparisons. When making reflected self-evaluations from one’s best friend’s perspective, adolescents recruited significant activity in vmPFC during academic and social evaluations, as well as Precduring social evaluations. However, no activity associated with reflected physical self-evaluations reached significance after correcting for multiple comparisons. It is important to note that when cluster extent thresholds were relaxed, all conditions revealed CMS activity patterns largely similar to those in the adult sample (see Table S2; coordinates marked by an * represent significant activity at relaxed thresholds).

Table S1.Task Stimuli

Academic positive / Physical positive / Social positive
Fast learner / Very good-looking / Very popular
Studies hard / Has a nice figure / Easy to talk to
Listens in class / Has beautiful eyes / Talks to classmates
Really intelligent / Has great hair / Goes to parties
Honor student / In good shape / Has a best friend
High test scores / Has clear skin / Often helps others
Academic negative / Physical negative / Social negative
Isn’t very smart / Fairly unattractive / Socially awkward
Bad student / Has a weird nose / Kind of a loner
Has failed a test / Has lots of pimples / Hard to like
Gets low grades / Pretty overweight / Has very few friends
Skips class a lot / Has crooked teeth / Tells lies
Cheats on tests / Has a strong body odor / Gets teased

The above stimuli were used in the MRI paradigm (see method section for more information).

Note: Trait phrases were piloted by 15 college students who evaluated (on a yes/no basis) whether each phrase represented its intended domain, was concrete (versus abstract), and was generally malleable. Trait phrases did not differ on mean representativeness, concreteness, or malleability across domains. A phrase was included in the final stimuli set if 80% of pilot participants agreed that it represented the intended domain and 75% of pilot participants agreed that it was malleable.

Table S2. Planned Linear Contrasts from Repeated Measures ANOVAs with Two Factors (Evaluative Perspective x Domain)- Conducted Independently for Adults and Adolescents

Adults / Adolescents
Contrast / Region / x / y / z / t / k / Region / x / y / z / t / k
DA>CA / Prec / 0 / -57 / 33 / 5.08 / 598 / vmPFC / -6 / 51 / -6 / 4.21 / 348
PCC / -3 / -48 / 6 / 3.57 / sgACC / -3 / 33 / -9 / 3.93
vmPFC / -3 / 45 / -6 / 4.6 / 548 / pgACC / 0 / 36 / 6 / 4.07
sgACC / 6 / 33 / -6 / 4.04 / Prec / -9 / -51 / 30 / 3.98 / 267
armPFC / 0 / 51 / 12 / 3.69 / PCC / -6 / -66 / 21 / 3.84
rACC / 0 / 33 / 12 / 3.51
DP>CP / vmPFC / -12 / 48 / 0 / 5.28 / 967 / Prec / -9 / -54 / 36 / 4.47 / 92
rACC / -3 / 18 / 21 / 4.59 / vmPFC / -6 / 51 / -6 / 3.9 / 235
armPFC / 0 / 51 / 12 / 4.21 / Prec/PCC / -15 / -63 / 18 / 3.44 / 74
PCC / -12 / -45 / 18 / 3.74 / 226
Prec / -6 / -45 / 48 / 3.3
DS>CS / armPFC / -3 / 51 / 12 / 5 / 1182 / Prec / -9 / -54 / 39 / 3.5 / 17*
vmPFC / -9 / 48 / 0 / 4.62 / vmPFC / -9 / 48 / -6 / 3.46 / 25*
rACC / 3 / 33 / 15 / 4.27
Prec. / -6 / -60 / 21 / 4.83 / 638
PCC / 0 / -51 / 12 / 3.75
OA>CA / Prec / -3 / -63 / 27 / 5.1 / 789 / pgACC / 0 / 33 / 0 / 3.31 / 25*
PCC / -9 / -54 / 12 / 4.15 / vmPFC / -12 / 48 / -3 / 3.21 / 27*
vmPFC / -9 / 42 / -3 / 4.45 / 226 / Prec/PCC / -9 / -51 / 30 / 3.14 / 14*
sgACC / 3 / 30 / -6 / 3.47
Thalamus / -27 / -33 / 3 / 3.54 / 70
OP>CP / pgACC / 6 / 33 / 0 / 4.22 / 262 / vmPFC / -6 / 51 / -3 / 3.33 / 66
vmPFC / -12 / 45 / 3 / 4.14 / Prec/PCC / -6 / -54 / 33 / 3.29 / 23*
Prec / 3 / -54 / 33 / 3.54 / 70 / pgACC / 0 / 33 / 0 / 3.23 / 10*
OS>CS / Prec/PCC / 3 / -60 / 27 / 6.09 / 811 / vmPFC / -12 / 45 / 0 / 4.44 / 504
vmPFC / 0 / 57 / -6 / 4.82 / 792 / rACC / 0 / 30 / 21 / 3.98
dmPFC / -12 / 45 / 27 / 4.59 / pgACC / -3 / 33 / 3 / 3.84
sgACC / -3 / 27 / -3 / 4.54 / dmPFC / 0 / 57 / 27 / 3.38
armPFC / 0 / 54 / 15 / 3.94
RA>CA / Prec / 0 / -60 / 33 / 4.74 / 305 / vmPFC / -9 / 48 / -6 / 3.33 / 60
STG/TPJ / -39 / -57 / 27 / 4.47 / 268 / pgACC / 0 / 33 / 0 / 3.13 / 17*
TPJ / -51 / -54 / 18 / 3.79
STS / -54 / -24 / -9 / 4.09 / 72
vmPFC / 0 / 57 / -3 / 3.28 / 29*
RP>CP / Prec / 0 / -57 / 33 / 4.38 / 156 / Prec/PCC / -9 / -54 / 36 / 3.59 / 34*
armPFC / 0 / 54 / 15 / 4.07 / 158 / vmPFC / -9 / 51 / -9 / 3.34 / 29*
vmPFC / 3 / 57 / -3 / 3.58
RS>CS / Prec / 0 / -54 / 36 / 6.73 / 595 / Prec / -6 / -54 / 39 / 4.58 / 201
vmPFC / 3 / 57 / -3 / 5.65 / 1114 / vmPFC / -3 / 57 / -3 / 4.46 / 472
armPFC / 0 / 54 / 12 / 5.07 / rACC / 0 / 30 / 12 / 3.62
pgACC / 3 / 30 / 3 / 3.96
STG/TPJ / -39 / -60 / 27 / 5.4 / 366
TPJ / -54 / -57 / 21 / 4.5

Note: Corrected for multiple comparisons (FWE p < 0.05) with magnitude and spatial extent thresholds at p < 0.005 and k = 57 voxels and k = 60 voxels for adults and adolescents, respectively. Minimum cluster-size thresholds were calculated using MonteCarlo simulations in AFNI. k-values and t-values are reported for peak voxels of each cluster. Additional subpeaks within larger clusters are included for descriptive purposes. * = magnitude and spatial extent thresholds at p < 0.005 and k = 10. DA = direct academic self-evaluations; DP = direct physical self-evaluations; DS = direct social self-evaluations; OA = direct academic other-evaluations; OP = direct physical other-evaluations; OS = direct social other-evaluations; RA = reflected academic self-evaluations; RP = reflected physical self-evaluations; RS = reflected social self-evaluations; CA = academic malleability-evaluations; CP = physical malleability-evaluations; CS = social malleability-evaluations; Prec = precuneus; PCC = posterior cingulate cortex; vmPFC = ventral medial prefrontal cortex; armPFC = anterior rostral medial prefrontal cortex; dmPFC = dorsal medial prefrontal cortex; rACC= rostral anterior cingulate cortex; sgACC = subgenual anterior cingulate cortex; pgACC = perigenual anterior cingulate cortex; STG = superior temporal gyrus; TPJ = temporoparietal junction; STS = superior temporal sulcus.

Figure S1. Mean Parameter Estimates in CMS Across Evaluative Conditions

Main effect of evaluative perspective. Parameter estimates were extracted from and then averaged across all voxels in the vmPFC/rACC (blue) and Prec/mPPC (red) clusters identified in Table 1 (peaks located at [-9 48 3] and [-6 -54 36], k = 749 and 537, respectively). Bar graphs represent parameter estimates from direct self-, direct other-, and reflected self-evaluations, relative to control malleability evaluations. Panel A illustrates mean parameter estimates representing activity in these two clusters for adults. Panel B illustrates mean parameter estimates representing activity in these two clusters for adolescents. Post-hoc one-sample t-tests indicated that all mean parameter estimate values were statistically greater than zero. Note: vmPFC = ventral medial prefrontal cortex; rACC = rostral anterior cingulate cortex; Prec = precuneus; mPPC = medial posterior parietal cortex.