An experimental foundation for the study of cognition as a dynamic, self-organizing system

Date of Completion

January 2009


Psychology, Psychometrics|Psychology, Cognitive




Dynamical approaches to the study and understanding of cognition show promise and have been guiding a growing area of research. To date, however, there are very few tools for studying cognition as a dynamical system in a dynamical manner. This dissertation provides a foundation for how dynamical data can be obtained and analyzed in experiments that still embrace standard cognitive manipulations. The acquisition method relies on participants making continuous movements, both to pace the presentation of items and to make responses using a stylus and digitizing tablet. The data analysis uses newly developed measurement techniques that rest upon capturing topological properties of the so-called distance matrix that is generated in recurrence plot analysis. Four experiments examined the utility of this methodology to study the dynamics of cognitive behavior in a variety of decision tasks. Experiment 1 revealed that dynamics are best uncovered when the response interface is configured as a congruent dual task and not a concurrent dual task. Subsequent experiments verified that dynamical measures do, indeed, reflect cognitive task manipulations. Moreover, the information revealed by these measures is non-redundant with response time and accuracy. In particular, entropy appears to be the best indicator of task difficulty as indexed by the quality of participant responses. It also appears to be indicative of the overall quality of the coupling between system internal-degrees of freedom and environmental external-degrees of freedom. ^