2026 · 28th National Clinical Education Symposium · Turkish Journal of Psychiatry · presented
Bekdemir F, Özdin S · doi:10.5080/61upk.ozt461
State anxiety and uncertainty processing in generalized anxiety disorder
Generalized anxiety disorder is often described as a problem with learning. This study asked a narrower question: is it what people learn that changes, or how readily they revise a belief once the world turns out to be unstable?
Eighteen adults with GAD and 18 age- and education-matched controls completed a probabilistic reversal-learning task while belief updating was modelled with the Hierarchical Gaussian Filter. The parameter of interest, tonic volatility (ω₂), is how unstable a person expects their environment to be by default. Participants also completed the State–Trait Anxiety Inventory and the Intolerance of Uncertainty Scale. Approved by the Ondokuz Mayıs University Clinical Research Ethics Committee (2026/43).
The GAD group showed higher ω₂ than controls (p < .001), with less win-stay and more lose-shift behaviour (both p < .001) and a smaller difference in overall accuracy (p = .019). Volatility sensitivity correlated with state anxiety (ρ = 0.42, p = .010), intolerance of uncertainty (ρ = 0.39, p = .020) and, most strongly, lose-shift behaviour (ρ = 0.55, p < .001). The correlation with trait anxiety did not reach significance (ρ = 0.32, p = .059); whether it differs from the state-anxiety correlation was not tested directly, so reading the two as different is provisional.
The pattern is consistent with excessive updating under uncertainty rather than a general learning deficit. The design is cross-sectional, so it does not establish which way the relationship runs, and with 36 participants and no correction for multiple comparisons the findings are exploratory.
This is a separate study from the schizophrenia thesis, not a preliminary stage of it. It is where the modelling approach the thesis depends on was first worked out.