A STRATEGY–STABILIZATION VIEW OF THE RGT: PROBING CHOICE-STRATEGY MODULATION WITH THE PSYCHEDELIC DOI ACROSS ENDOPHENOTYPES AND SEX

We reframe rGT analysis from endpoint preference averages to a strategy-stabilization framework that captures learning dynamics while preserving individual differences. We model daily behavior as a full P1–P4 probability profile and quantify day-to-day similarity to define an objective endpoint: the first occurrence of three consecutive days that meet a stability criterion; we test this endpoint across multiple stringency levels (low/medium/high variability). Across thresholds, males stabilize earlier than females, indicating faster convergence to a consistent decision strategy rather than a simple shift in mean choice. We then apply unsupervised clustering to identify distinct choice-profile endophenotypes (risk-averse, risk-seeking, mixed strategies) and test how endophenotype membership shapes time-to-stability, linking strategy dynamics to interpretable behavioral subtypes. Finally, we use DOI, a serotonergic psychedelic, to probe the causal flexibility of established strategies: we test whether DOI shifts the choice-profile geometry (P1–P4 probabilities), alters stability (e.g., destabilization vs accelerated restabilization), and changes transition patterns between endophenotypes.
Funding: This work was supported by the Spanish Ministry of Science and Innovation (MCIN) and the State Research Agency (AEI) through the Proyectos de Generación de Conocimiento (PGC) programme (MCIN/AEI/10.13039/501100011033; Grant PID2022-139286NB-I00) and co-funded by the European Regional Development Fund (ERDF/FEDER) (“A way of making Europe”). Julián Guarque Chabrera was supported by the Generalitat Valenciana (Conselleria de Educación, Cultura, Universidades y Empleo), APOSTD 2025 postdoctoral contract programme (Exp. CIAPOS/2024/342; 16/10/2025–15/10/2027).