Psychotherapy Processes in PTSD Treatment: Trajectories of Positive and Negative Valence Systems

Abstract

Despite strong empirical support for the effectiveness of existing exposure-based and pharmacological treatments for trauma-related psychopathology (Cusack et al., 2016; Jeffreys et al., 2012; Sakaluk et al., 2019), far less is known about the processes driving clinical change. Existing research points to change processes involving both positively and negatively valenced affect, cognitions, and processing styles; yet past studies have been limited by overreliance on self-report measures, a dearth of longitudinal studies examining session-to-session change, and failures to integrate cognitive, affective, and processing variables in computational modeling approaches designed to study change. Although several studies have documented the role of decreases in negative processes in promoting recovery from trauma-related psychopathology (e.g., Cooper, Clifton, et al., 2017), inhibitory learning models of extinction processes during exposure therapy increasingly point towards the importance of cultivating increases in positive processes as well (Craske et al., 2016, 2019; Zbozinek & Craske, 2017). In the present study, patients (N = 149) with primary PTSD enrolled in a clinical trial of prolonged exposure (PE) delivered alone or augmented with sertraline completed self-report measures of positive and negative affect (PA and NA) across ten sessions of treatment. Trained coders rated cognitive, affective, and processing psychotherapy change processes during psychotherapy sessions, coding both patient statements and behaviors at the first session and subsequent sessions after imaginal exposure to the trauma memory. Utilizing dynamic structural equation modeling, Study 1 examined temporal patterns of change in self-reported positive and negative affect and PTSD symptoms over treatment, and Study 2 examined patterns of change in positive and negative systems of psychotherapy process change processes and their relation to PTSD symptoms from session to session. In Study 1, positive affect increased moderately (d = 0.51) and NA decreased strongly (d = 0.78) across treatment sessions. Changes in PA and NA were generally reciprocal (PAtà NAt+1: ES = -0.09, 95%CI = -0.15, -0.02; NAtà PAt+1: ES = -0.20, 95%CI = -0.28, -0.13). However, changes in PTSD more strongly predicted next session negative affect (PTSDtà NAt+1: ES = 0.50, 95%CI = 0.38, 0.60) and positive affect (PTSDtà PAt+1: ES = -0.26, 95%CI= -0.34, -0.17) than the reverse. PE augmentation with an SSRI did not moderate temporal associations. In Study 2, positive system activation increased strongly (d = 1.42) and negative system activation decreased strongly (d = 0.98) over treatment sessions. Changes in positive and negative systems were reciprocal, with stronger effects of positive system activation on subsequent negative system activation than vice versa (positive systemt à negative systemt+1: ES = -0.25, 95% CI [-0.38, -0.11]); negative systemt à positive systemt+1: ES = -0.09, 95% CI [-0.16, -0.004]), and the effect of positive system changes on negative system changes was stronger with sertraline augmentation (interaction ES = -0.42, 95% CI [-0.80, -0.10]). These findings were the opposite of the patterns observed for self-reported affect and PTSD. Decreases in PTSD symptoms predicted subsequent increases in positive (PTSDt à positive systemt+1: ES = -0.44, 95% CI [-0.55, -0.37]) and decreases negative system activation (PTSDt à negative systemt+1: ES = 0.38, 95% CI [0.26, 0.52]), respectively, but system changes did not predict subsequent changes in PTSD symptoms. Across studies, prolonged exposure produced substantial improvements in self-reported PA and NA as well as positive and negative emotions, cognitions, and processing styles during imaginal exposure processing. General affective changes may be more a consequence than a driver of PTSD improvement during PE, with improvements in NA and PA potentially linked to the extinction of negative emotional responses to trauma cues and increased engagement with rewarding activities, respectively. Increases in positive psychotherapy processes may be an especially robust indicator of improvement in treatment, and may help patients overcome problems in psychotherapy indexed by increases in negative psychotherapy processes. The impacts of positive psychotherapy processes may be more pronounced with medication augmentation of psychotherapy.