2026 ECNP Congress

PS01-1107

Within-person changes in positive bias predict subsequent depressive symptoms: a longitudinal mixed-effects study in primary care

Ragnhild Bø — H.S. Hagen, O.A. Andreassen, S.E. Blackwell, G.R. Dawson, C.J. Harmer, N.I. Landrø, J.I. Røssberg · University of Oslo, Department of Pschology, Oslo, Norway

2026 ECNP Congress

Background:

Cognitive models of depression propose that biases in information processing contribute to symptom maintenance [1, 2]. However, longitudinal evidence for within-person temporal effects remains limited, and mostly reliant on non-clinical groups (e.g., [3]). Based on a secondary analysis of a randomized clinical trial conducted in primary care (NCT05503966) comparing Attention bias modification (ABM) to an active control condition [4], we investigate whether changes in positive bias predict subsequent depressive symptoms over time, and whether this differs between intervention groups.

Methods:

Patients with Major Depressive Disorder were prescribed antidepressants by their general practitioner and randomized to two weeks of two daily sessions of remotely delivered ABM (n=21) or an active control condition (n=24). The ABM procedure aim to direct attention towards more positive stimuli. Patients were assessed remotely at baseline, 2, 5 and 8 weeks, which was the primary endpoint of the trial. Self-reported depressive symptoms (Beck’s depression inventory: baseline M=33.1 (8.1)) and positive bias (facial expression recognition task, % accurate perception of happy faces: baseline M=40.7 (15.5)) were assessed at all timepoints. Linear mixed-effects models with intercept on level of the individual were estimated including group, gender, nonlinear time (linear and quadratic terms), and Group x Time-interaction. To test temporal precedence, lagged within-person deviations in positive bias were used to predict subsequent depressive symptoms, controlling for prior symptom levels (autoregressive effect).

Results:

The model showed adequate fit (marginal R2=.0335; conditional R2 = .636), indicating that fixed effects explained 33.5% of variance in depressive symptoms, and 63.6% when including random effects. The ICC (0.453) suggested substantial between-person variability. Depressive symptoms showed a significant non-linear trajectory over time (linear: b = −8.01, p = .036; quadratic: b = 3.02, p = .035), indicating initial improvement followed by attenuation. The ABM group exhibited higher overall symptom levels compared to control (b = 9.08, p = .006), but no significant Group × Time interaction was observed (p = .085) indicating no differential change between groups over time. Prior depressive symptoms predicted subsequent symptoms (b = 0.285, p = .007, β = .29). Within-person increases in positive bias predicted higher subsequent depressive symptoms (b = 0.201, 95% CI [0.008, 0.394], p = .042, β = .33).

Conclusion:

Within-person fluctuations in positive bias were temporally associated with subsequent depressive symptoms, independent of prior symptom levels. The direction of this moderate-sized effect did not reflect standard cognitive models, which typically assumes that increasing positive bias would be followed by symptom reduction. Hence, the results suggest a more complex dynamic relationship between cognitive processing and symptom change than initially assumed. A fully dynamic reciprocal process was not modelled, limiting causal inferences. The interventions did not show differential effects on symptom trajectories. Future research conducted in larger clinical samples on dynamic relations between cognitive processing and symptoms is required.

References

  1. Beck, A.T., (1967). Depression: Clinical, experimental, and theoretical aspects. 1967, New York: Harper & Row.
  2. Gotlib, I.H. and J. Joormann (2010). Cognition and depression: current status and future directions. Annu Rev Clin Psychol, 6, 285–312. https://doi.org/ 10.1146/annurev.clinpsy.121208.131305
  3. Wisco, B.E. and D.R. Harp (2021). Rumination as a mechanism of the association between interpretation bias and depression symptoms: A longitudinal investigation. Journal of Experimental Psychopathology, 12(2), 20438087211015233. https://doi.org/10.1177/204380872110152
  4. Landrø, N.I., et al., (2023). Combining antidepressants and attention bias modification in primary health care (DEPTREAT): Protocol for a pragmatic randomized controlled trial. Contemp Clin Trials, 133, 107326. https://doi.org/10.1016/j.cct.2023.107326