Objective: This study investigates the entire course of an Affect Phobia Therapy (APT; McCullough, 1997) treatment. APT postulates that affective activation, insight into maladaptive patterns and an increased sense of Self are essential for positive treatment outcome. The present study investigates the change of three process variables: Insight (into defensive patterns; D), Activating affects (F[eeling]) and Sense of Self (SoS), both within and across sessions. The thesis studies D and F as process predictors, SoS as intermediate outcome variable, and overall outcome. Further, it also inquires into whether there are sequential relationships between these four variables. We test three hypotheses: I. More than expected F in a session will predict higher SoS in the following session(s). II. More than expected F in a session will predict higher D in (the) subsequent session(s). III. More than expected D in a session will predict higher SoS in the subsequent session(s). Method: Data from a single subject based on the previously published Randomized Control Trial of Brief Cognitive and Dynamic Therapy (Svartberg, Stiles & Seltzer, 2004) were scored with the Achievement of Therapeutic Objectives Scale (ATOS). ATOS ratings were based on video-recorded therapy. The patient met the criteria for a Cluster C personality disorder (dependent personality disorder). Time series modeling of process variables was performed to test our hypotheses. Outcome measures were SCL-90, MCMI-C, BDI and IIP, in addition to qualitative observation and SoS. The entire treatment was transcribed. Analysis: The data were analyzed by a two-way Analysis of Variance design, linear regression, ARIMA and a Cross Lagged Correlation design. Variance, trends and the sequential relationships between the process variables were investigated. Qualitative analysis was done based on verbatim transcripts of the entire treatment. Results: The results indicated a significant clinical change, overall high ATOS-scores and very good/excellent alliance. Between therapy sessions we found three small trends (nonsignificant) of sequential relationships: 1. More than expected F in a session predicted more than expected SoS; 2. More than expected D predicted less than expected F; 3. More than expected D predicted less than expected SoS. Conclusion: The overall results indicated significant positive clinical change, which signals that this was a successful treatment course. Increased SoS and F can to a certain extent predict this outcome, but the vague indication that D negatively predicts SoS and F is puzzling. All results from the lagged crosscorrelations were nonsignificant. This single case study offers small effect size, and results cannot be easily generalized. However, even though results must be interpreted as change mechanisms in this unique case, they also seem to confirm some central tenets in APT. Our findings may also largely be due to diverse methodological weaknesses/challenges, and the properties of the outcome and/or process measures.