Behavioural Results
The mixed factors ANOVA for arousal revealed significant effects for instruction type (F(3, 138) = 57.0, p < .001, η2p = .553), with negative-view eliciting greater arousal (M = 6.84, SD = 0.24), followed by reappraisal (M = 6.42, SD = 0.25), distraction (M = 6.02, SD = 0.25), than neutral-view (M = 3.95, SD = 0.24); but a group effect (OB, HC) was not present, F(1, 46) = 0.3, p = .58, η2p = .007. Also, for valence, there was a significant effect for instruction type, F(3, 138) = 145.8, p < .001, η2p = .760, with negative-view eliciting greater unpleasant ratings (M = 2.34, SD = 0.17), followed by reappraisal (M = 2.74, SD = 0.19), than distraction (M = 3.02, SD = 0.21), than neutral-view (M = 6.34, SD = 0.16); however, a group effect (OB, HC) was not found, F(1, 46) = 0.75, p = .09, η2p = .002 as well (Table 2).
ANOVA analyses according to view Instruction showed a significant effect for type of image, F(1, 46) = 88.0, p < .001, η2p = .657, with greater arousal rates for the instruction negative-view images (M = 6.84, SD = 0.24) compared to neutral-view images (M = 3.95, SD = 0.24). However, a group effect was not found, F(1, 46) = 0.44, p = .50, η2p = .012. Same effects involving valence were significant. There was a significant effect for type of image, F(1, 46) = 222.1, p < .001, η2p = .828, with more negative valence ratings for negative-view images (M = 2.34, SD = 0.17) than for neutral (M = 6.34, SD = 0.16). Again, a group effect was not detected for valence ratings, F(1, 46) = 0.07, p = .78, η2p = .002.
Electrophysiological Results
Mixed factors ANOVA for mean LPP amplitudes for the entire image duration (300-5000ms) showed no significant effects for instruction type, F(3, 147) = 1.65, p = .18, η2p = .033, nor a group effect, F(1, 49) = 0.02, p = .86, η2p = .001.
For time segments for each type of instruction, results showed significant effects for instruction type, F(3, 294) = 21.7, p < .001, η2p = .307, with greater amplitudes for distraction compared to other types of instruction; and time, F(6, 294) = 3.89, p < .001, η2p = .074, with greater amplitudes for the 300ms-500ms time segment than the remaining time segments.
However, group effects were not found for instruction type, F(1, 49) = 0.25, p = .61, η2p = .005, nor time, F(6, 294) = 1.36, p = .22, η2p = .027. Also, no significant interaction between instruction type and time segment were found F(6, 882) = 1.37, p = .13, η2p = .027.
Independent t-test used to analyse group differences for each type of instruction across time segments (Table 3) revealed a larger LPP for reappraisal at 500-700ms time segment, t(49) = 2.47, p < .001, for the OB (M = 2.06, SD = 2.30) compared to HC (M = 0.60, SD = 1.91). Similar results were found for distraction also at 500-700ms time segment, t(49) = 1.74, p = .04, for the OB (M = 1.55, SD = 2.71) compared to HC (M = 0.44, SD = 1.80).
Paired t-tests comparing negative-view with neutral-view for each time segment showed greater amplitudes for neutral-view (M = 1.42, SD = 1.57) compared to negative-view (M = 0.89, SD = 1.57) for the 300-500ms time window, t(27) = -1.92, p = .03, for the OB participants. The effect size, as measured by Cohen's d, was d = 0.36, indicating a small effect. No significant differences were found for the remaining time segments (all ps > .05). In regards to the HC participants, no statistically significant differences were found for the LPP amplitudes between negative-watch and neutral-view for all time segments (all ps > .05). For each emotion regulation strategy, results showed that distraction reduced the LPP (M = -0.61, SD = 2.22), compared to negative-view (M = 0.36, SD = 2.31), only at the 1500-1700ms time window for OB participants, t(27) = 1.78, p = .04. The effect size, as measured by Cohen's d, was d = 0.33, indicating a small effect. On the other hand, reappraisal showed greater LPP amplitudes compared to negative-view, for time segments between 300ms and 1300ms. For the HC participants, distraction and reappraisal showed no significant differences compared to negative-view for all time segments (Table 4).
Covariates analysis
For arousal, nor BES or anxiety, changed the pattern of results previously found (all ps < .001). Same results were found for valence ratings.
In regard to mean LPP amplitudes for the entire image duration (300-5000ms), also BES and anxiety did not change the pattern of results found previously (all ps > .001).
Considering time segments for each type of instruction, similar results were found for instruction type (p < .001). However, changes in time results pattern was reported for, BES, F(6, 270) = 1.11, p = .34, η2p = .021, and anxiety, F(6, 270) = 1.30, p = .27, η2p = .022.