2026-10-12T17:25:49Z https://api.figshare.com/v2/oai

oai:figshare.com:article/34036519 2026-09-30T17:52:18Z category_15 category_24 category_45 category_734 category_64 portal_5 item_type_6 month_year_09_2026

Full hierarchical moderation model: main effects and interaction terms. Sergej A. E. Golowin (25145509) Niall W. Duncan (9177024) Faizan Shaikh (25145512) Christoph W. Korn (11657787) Neuroscience Evolutionary Biology Sociology Biological Sciences not elsewhere classified Cancer without requiring assumptions enabled formal comparison choices showed evidence sequential foraging task integrated value signals approach &# 8211 g ., reward humans adjust decision avoidance &# 8211 threat probabilities ), higher reward probabilities optimal action values individual decision features higher predation risk predation risk action values value structure sequential decision optimal policy relevant features decision features reward probability probabilistic reward xlink "> strongly reflected shared model remains unclear reduced reliance policies based parsimonious account optimization principles increased alignment immediate rewards future risks findings suggest feature policy enhanced integration creating contexts commonly framed avoidance contexts Posterior mean estimates (points) and 95% credible intervals (horizontal lines) are shown for all main effects and interaction terms from the hierarchical logistic regression model. Estimates are reported on the logit scale; positive values indicate an increased probability of choosing the foraging option, whereas negative values indicate a decreased probability. The vertical dashed line denotes the null effect (0). (A) Model including raw . (B) Model including residualized with respect to the variables of the multi-feature model. Main effects reflect baseline contributions of decision variables, including reward probability (), threat probability (), model-derived optimal values (), and state-dependent features (binary energy, wait when safe). Interaction terms capture how these contributions are modulated by experimental condition (approach vs. avoidance) and block number. Across both specifications, results indicate selective feature reweighting rather than global shifts (no condition main effect). In avoidance contexts, the influence of is reduced, whereas alignment with increases, while sensitivity to remains comparatively stable. Block-related interactions indicate gradual changes in feature weighting over time, characterized by increasing alignment with and boundary states (binary energy state and wait when safe). Crucially, the residualized model in (B) confirms that the condition-dependent increase in persists beyond variance shared with other predictors, indicating a unique contribution of integrated value representations. Higher-order interactions, including and condition, were not credibly different from zero in either model, suggesting that explicit feature interactions do not explain behavior beyond integrated value signals captured by . Continuous predictors were standardized prior to model fitting. Condition was coded as 0 = approach and 1 = avoidance. (PDF) 2026-09-30T18:00:31Z Text Journal contribution 10.1371/journal.pcbi.1014793.s002 https://figshare.com/articles/journal_contribution/_p_Full_hierarchical_moderation_model_main_effects_and_interaction_terms_p_/34036519 CC BY 4.0