Stochastic progression of new states in psychotherapy

This paper presents a study of the stochastic behaviour of the entry of a 5-dimensional, 2400 state, system into new states. Each state consists of five measurements on variables that are sensitive measures of the degree of activated unconscious communication. The data come from six psychotherapy consultation sessions. The study sought to investigate the manner in which patient and therapist progress into new configurations (new states) of the five variables, and the way tension accumulates to the point of overcoming the inertia in the patient/therapist system. We show that the time series that consists of the time intervals between successive new states is a generalized Poisson process. That is, the process is inherently governed by a Poisson parameter that is a continuous piecewise linear function of time. It increases rapidly in the opening moments of an interview and decreases slowly thereafter. This non-linear, non-stationary Markov model fits the data well according to the Kolmogorov-Smirnoff criterion. We discuss the model's implications for the nature of psychotherapy and its participants.