Power Analysis. A statistical power analysis was performed for sample size evaluation, based on our data comparing H-BPD to L-BPD groups. With an alpha = .05 and power = 0.80, the projected sample size needed do detect “medium” effect size differences between groups (GPower 3.1) is approximately N = 12 for this between group comparison. Thus, our sample size is adequate for the main hypotheses of this study.
Behavioral analyses. A mixed model ANOVA was performed on ratings of untrustworthiness, with Group (2 levels: low BPD, high BPD) as the between-subjects factor and Test (2 levels: pretest, posttest) and Condition (4 levels: ambiguous, negative, positive, mixed) as within-subjects factors. We found a main effect of Group, F(1,25) = 8.32, p < .01, MSe=.53, η2 = .44; participants high in BPD features judged faces significantly higher in untrustworthiness (average = 3.23/5.0) than participants low in BPD features (2.94/5.0). There were no main effects of Test, F(1,25) = 1.91, ns, MSe=.09, η2 = .02, or Condition, F(3,75) = 1.40, ns, MSe=.60, η2 = .25, but there was a significant Test x Condition interaction, F(3,75) = 3.24, p < .05, MSe=.53, η2 = .08. As shown in Fig. 2, social learning caused participants in the ambiguous and untrustworthy conditions to judge faces as significantly less trustworthy while judging faces in the positive and mixed conditions as slightly more trustworthy. We found a trend toward a three-way interaction of Group, Test and Condition, F(3,75) = 2.42, p = .07, MSe =.53, η2 = .06, because the effects of training on condition were largely restricted to participants low in BPD features. An ANOVA of RTs yielded only one statistically significant effect: a main effect of Test, F(1,25) = 9.47, p < .01, MSe=.03, η2 = .74. Participants were significantly faster at rating face stimuli after (1036 ms) compared to before training (1112 ms). We found a trend toward a Test x Condition interaction, F(3,75) = 2.46, p = .07, MSe=.53, η2 = .08, because the speedup from training was relatively stronger in the negative condition and relatively weaker in the ambiguous condition.
ERP analyses. A mixed model ANOVA was performed separately on the P1, P2, N2 and slow wave ERP components, with Group (2 levels: low BPD, high BPD) as the between-subjects factor and Test (2 levels: pretest, posttest) and Condition (4 levels: ambiguous, negative, positive, mixed) as within-subjects factors. An effect of training was found in the P1 ERP component, F(1,25) = 5.07, p < .05, MSe=261.73, η2 = .55, with the magnitude of P1 significantly larger at posttest (1.98 µV) than at pretest (.19 µV). However, there was no effect of Group, F(1,25) = .01, ns, MSe=246.91, η2 = .001, and no interaction between Group and Test, F(1,25) = .01, ns, MSe=261.73, η2 = .001, indicating that the sensory effects of training were equivalent between groups.
A trend-level effect of training was also found in the P2 ERP component, F(1,25) = 3.90, p = .06, MSe=30.21, η2 = .28, with P2 amplitude at posttest (2.68 µV) diminished relative to pretest (3.69 µV). There was no main effect of Group, F(1,25) = .18, ns, MSe=284.51, η2 = .12. However, there was a trend-level interaction between Group and Test, F(1,25) = 3.790, p = .06, MSe=30.21, η2 = .27: The drop in P2 after training was relegated to the high BPD participants. Similar trends were found in the N2 ERP component. There was a significant main effect of Test, F(1,25) = 5.06, p < .05, MSe=68.47, η2 = .28, with N2 amplitude growing more negative from pretest (-.23 µV) to posttest (-.52 µV). There was no effect of Group, F(1,25) = 1.12, ns, MSe=395.12, η2 = .35, but Group and Test trended toward interaction, F(1,25) = 2.51, p = .13, MSe=68.47, η2 = .14, because the effects of training were seen only in the high BPD group.
A different pattern was observed in analysis of the slow wave. Here, we found no effect of Test, F(1,25) = .45, ns, MSe=107.05, η2 = .02, or Group, F(1,25) = 2.71, ns, MSe=376.18, η2 = .38, and no interaction between the two variables, F(1,25) = .24, ns, MSe=107.05, η2 = .01. However, there was a significant three-way interaction with Test, Group, and Condition, F(3,75) = 3.40, p < .05, MSe=28.02, η2 = .11. As one can see in Fig. 3, after training the slow wave decreased in magnitude for low BPD participants in the negative condition, but decreased in magnitude for high BPD participants in the positive and mixed conditions.