(PS2-C61) A Network Intervention Analysis of Differential Symptom Change in Cognitive-behavioral Therapy and Light Therapy for Seasonal Affective Disorder

First-line treatments for winter seasonal affective disorder (SAD) include light therapy (LT) and group cognitive-behavioral therapy for SAD (CBT-SAD). Although CBT-SAD and LT are comparably efficacious in treating an acute SAD episode (Rohan et al., 2015), prior work has not examined the relative impacts of CBT-SAD and LT on individual SAD symptoms and how those symptoms then influence each other over time in treatment. LT has an assumed chronobiological target and should more directly influence the vegetative symptoms, whereas CBT-SAD has an assumed cognitive target and should more directly influence the cognitive/affective symptoms. We conducted an exploratory network intervention analysis (NIA) using data from a randomized clinical trial comparing LT and CBT-SAD to inform how these treatments effect symptoms.

The study included the 159/177 randomized participants who completed the 6-weeks treatment phase (LT: 88, CBT-SAD: 75). The modified Beck Depression Inventory-Second Edition (BDI-II), including one additional item for weight gain, was administered at pre-, mid-, and post-treatment. This NIA examined scores on individual items and on four previously identified symptom clusters using this sample at baseline (Smetter et al., 2021): Loss of Vitality, Dysregulation, Increased Appetite and Weight, and Negative Cognitions.

There were no LT vs. CBT-SAD differential treatment effects on individual symptoms or symptom clusters at mid- or post-treatment. However, there were network associations among symptoms and among symptom clusters that differed at the two timepoints. For individual symptoms, there were fewer associations at post-treatment (21) compared to mid-treatment (31), with no discernible patterns in the difference. For symptom clusters, at both time points there was an association between Loss of Vitality, Dysregulation, and Negative Cognition. However while Increased Appetite and Weight was unrelated to the other clusters at mid-treatment, at post-treatment Increased Appetite and Weight was associated with Loss of Vitality.

The lack of treatment differences indicates that these two mechanistically different treatments did not differentially effect severity of symptoms or symptom clusters at mid-treatment or at post-treatment. The change in associations among individual symptoms at mid- and post-treatment may be due to a weakening of the symptom network through global change (i.e., lower overall depressive severity). The change in associations among symptom clusters may be due to later treatment influence on appetite and weight changes at the cluster level. Future work should examine whether any treatment differences emerge with greater time resolution (e.g., weekly intervals). Clinically, this study provides evidence CBT-SAD may be just as effective as LT in terms of targeting vegetative symptoms of SAD, which are presumed to be rooted in circadian dysregulation, even after three weeks of treatment.