Mixing properties in human behavioral style and time dependencies in behavior identification: The modeling and application of a universal dynamical law

Human subjects perform simple, relatively unconstrained, game-like computer tasks. "Meso-level" measures of behavioral complexity and time-dependencies (including entropies, grammatical complexity estimates and run statistics) are derived and computed. Individual behavioral differences in the resulting complexity measures are robust and, in a temporal-forcing paradigm, are statistically significantly related to the same individual's scores on a range of personality and demographic variables. Through an experimental manipulation and the statistical selection of maximally useful predictor sets personality and demographic variables are united in a "macro-level" temperament typology, based on "micro-level" behavioral tendencies. Further, I can compute a parameter value of a one dimensional dynamical system, the symmetric tent map, matched to the symbol sequence "meso-level" parities of the subject. When this parameter is used in the iterated map, it produces sequences that are of the same autocorrelation "category" and share much of the fine structure of the autocorrelograms of the subjects to which the map parameter had been matched.