Abstract
<p>Experimental paradigms using continuous-response scales are widely employed across cognitive psychology. Over the past decade, multi-dimensional diffusion decision models (MD-DDMs) have been developed to investigate the cognitive processes underlying continuous-response decisions. Although these models have successfully captured and explained behavioral regularities across a range of cognitive domains, their adoption in applied research remains limited. One major barrier is the lack of accessible and flexible computational tools for model estimation. In this work, we introduce and provide a tutorial on using JEAM, a Python package for evidence accumulation modeling of continuous judgments. JEAM provides a fast and numerically stable framework for likelihood-based estimation of MD-DDMs. Moreover, it supports a broad range of response formats, including one-dimensional bounded and wrapped scales, as well as two-dimensional planar scales. The package also supports diffusion models with time-varying decision thresholds and arbitrary threshold dynamics, enabling the modeling of decision urgency. In addition, JEAM allows users to implement regression-like mapping functions that modulate model parameters as a function of experimental conditions, stimulus properties, or auxiliary variables such as neural activity. The package further supports mixture-process modeling, which is commonly used in judgment studies. Together, JEAM provides a flexible computational framework for modeling decision processes in continuous judgment tasks.</p>