Reliability shapes coordinate frame selection in spatial working memory

The format of spatial representation is a fundamental issue in cognitive research. Previous studies have shown that blank and grid environments selectively induce polar and Cartesian coordinate patterns, respectively. However, these findings are largely based on reliable blocked designs, leaving it unclear whether such representational formats remain stable under environmental uncertainty. Using a spatial location reproduction paradigm, we applied error correlation analysis to infer the underlying coordinate signatures, while introducing contextual reliability (grid/blank) during either the response phase (Experiment 1) or the encoding phase (Experiment 2). In Experiment 1, when uncertainty was introduced at retrieval, polar-consistent error patterns remained relatively preserved under blank encoding. Experiment 2 revealed that encoding-phase uncertainty significantly reduced the difference between polar and Cartesian error patterns across conditions. Meanwhile, encoding-phase uncertainty led to a preferential reliance on Cartesian coordinate features. Within the error-correlation framework, these findings delineate what contextual uncertainty changes (Cartesian signature expression; frame selection) and what it leaves unchanged (a robust polar signature under encoding reliability). OPEN PRACTICES STATEMENT: The data and analysis code for both experiments are publicly available at the Open Science Framework ( https://osf.io/znrb6/ ). Neither experiment was preregistered.