Table 5.
Cox proportional hazards models.
| Unit | Exp(B) | 95% CI | p (Wald) | p (model) | |
|---|---|---|---|---|---|
| Emotional Categorization (EmCat) | |||||
| Previous episodes−10 years | n | 1.119 | 1.042–1.203 | 0.002 | |
| HDRS-score residual symptoms | 1 point | 1.014 | 0.880–1.168 | 0.849 | 0.003 |
| Positive Mis. + Negative Mis. | n | 0.970 | 0.871–1.080 | 0.567 | |
| RT Neg. – RT Pos. Acc. | 10 ms | 1.013 | 0.986–1.041 | 0.345 | |
| RT Neg. – RT Pos. Mis.* | 10 ms | 1.003 | 0.995–1.011 | 0.423 | |
| Emotional Memory (EmMem) | |||||
| Previous episodes−10 years | n | 1.118 | 1.040–1.201 | 0.002 | |
| HDRS-score residual symptoms | 1 point | 1.019 | 0.886–1.171 | 0.797 | 0.004 |
| Negative Acc. – Positive Acc. | n | 0.923 | 0.748–1.138 | 0.453 | |
| Negative Mis. – Positive Mis. | n | 0.918 | 0.744–1.132 | 0.422 | |
| Positive Acc. – Positive Mis. | n | 0.924 | 0.763–1.120 | 0.422 | |
| Negative Acc. – Negative Mis. | n | 0.981 | 0.848–1.135 | 0.798 | |
| (Negative Mis. – Positive Mis.)/(Positive Acc. – Negative Acc)§ | % | 0.934 | 0.755–1.155 | 0.528 | |
| Facial Emotion Recognition (FERT) | |||||
| Previous episodes−10 years | n | 1.117 | 1.038–1.202 | 0.003 | |
| HDRS-score residual symptoms | 1 point | 0.983 | 0.848–1.139 | 0.818 | |
| Angry Mis. | n | 1.038 | 1.006–1.070 | 0.019 | 0.001 |
| Sad Acc. | n | 1.044 | 0.948–1.148 | 0.382 | |
| Neutral Acc. | n | 0.923 | 0.660–1.290 | 0.637 | |
| Disgust Mis. | n | 0.968 | 0.887–1.057 | 0.467 | |
| Neutral Mis. | n | 1.062 | 0.895–1.261 | 0.490 | |
| Positive Acc. – Neutral Acc. | n | 1.019 | 0.950–1.092 | 0.603 | |
| Negative Acc. – Neutral Acc. | n | 0.996 | 0.937–1.060 | 0.905 | |
| Negative Mis. – Neutral Mis. | n | 1.028 | 0.966–1.095 | 0.383 | |
| Neutral Mis. – Positive Mis. | n | 0.992 | 0.924–1.065 | 0.819 | |
Outcome-specific composite scores were defined based on significant differences and interactions of outcomes between patients and controls (see Table 2). Variables in italics represent the final models, for which a p-value (χ2) is given. All models contained previous episodes (last 10 years) and HDRS-score (residual symptoms). Selection of additional variables was done by forward stepwise selection from all listed outcome variables for each emotional bias task separately.
As not all subjects misclassified characteristics during emotional categorization, these mean reaction-times are based on less subjects (16 with recurrence and 16 resilient).
Cases with Positive Acc. – Negative Acc = 0 omitted from analyses (7 recurrent/7 resilient).