Predicting long-term outcome of Internet-delivered cognitive behavior therapy for social anxiety disorder using fMRI and support vector machine learning

Abstract

Cognitive behavior therapy (CBT) is an effective treatment for social anxiety disorder (SAD), but many patients do not respond sufficiently and a substantial proportion relapse after treatment has ended. Predicting an individual's long-term clinical response therefore remains an important challenge. This study aimed at assessing neural predictors of long-term treatment outcome in participants with SAD 1 year after completion of Internet-delivered CBT (iCBT). Twenty-six participants diagnosed with SAD underwent iCBT including attention bias modification for a total of 13 weeks. Support vector machines (SVMs), a supervised pattern recognition method allowing predictions at the individual level, were trained to separate long-term treatment responders from nonresponders based on blood oxygen level-dependent (BOLD) responses to self-referential criticism. The Clinical Global Impression-Improvement scale was the main instrument to determine treatment response at the 1-year follow-up. Results showed that the proportion of long-term responders was 52% (12/23). From multivariate BOLD responses in the dorsal anterior cingulate cortex (dACC) together with the amygdala, we were able to predict long-term response rate of iCBT with an accuracy of 92% (confidence interval 95% 73.2–97.6). This activation pattern was, however, not predictive of improvement in the continuous Liebowitz Social Anxiety Scale—Self-report version. Follow-up psychophysiological interaction analyses revealed that lower dACC–amygdala coupling was associated with better long-term treatment response. Thus, BOLD response patterns in the fear-expressing dACC–amygdala regions were highly predictive of long-term treatment outcome of iCBT, and the initial coupling between these regions differentiated long-term responders from nonresponders. The SVM-neuroimaging approach could be of particular clinical value as it allows for accurate prediction of treatment outcome at the level of the individual.

Hjärnan formas av psykologisk behandling

“…In the psychiatric literature there are no robust predictors of response to CBT for SAD (Eskildsen et al, 2010), and consistent with this, Study IV did not observe any significant clinical or demographic predictors of long-term treatment outcome (Månsson et al, 2015a). In contrast, the current thesis put forward that a pattern of neural response to selfreferential criticism in the dACC could predict an individual's clinical response to treatment, as opposed to predictors at the level of the group.…”

“…Study IV employed a SVM classification approach on neural response to self-referential criticism to evaluate single-subject predictors of long-term response one year after CBT (Månsson et al, 2015a). …”

Computational neuroimaging strategies for single patient predictions

“…This has shown some promising recent results, indicating that it may become possible to predict individual trajectories of patients with schizophrenia (Anticevic et al, 2015) or mood disorders (Lythe et al, 2015;Schmaal et al, 2015) from neuroimaging data, or forecast individual treatment responses to psychotherapy (Mansson et al, 2015), antidepressants (DeBattista et al, 2011;McGrath et al, 2013;Miller et al, 2013) and antipsychotics (Hadley et al, 2014;Nejad et al, 2013).…”

A Hitchhiker's Guide to Functional Magnetic Resonance Imaging

“…Another difference between the approaches relies on the fact that while t -tests model the complete set of time points, a classification trains on a subset of data (Coutanche, 2013). MVPA analyses are typically implemented using a “decoding” approach, which is based on the use of classifiers, such as neural networks (Polyn et al, 2005; Nickl-Jockschat et al, 2015), support vector machines (Meier et al, 2012; Månsson et al, 2015), and linear discriminant analysis (Cox and Savoy, 2003; Mandelkow et al, 2016), as a mean to differentiate between different classes or groups of individuals. Despite its popularity in the neuroimaging field, the “decoding” approach has some limitations, particularly related with the different results obtained with different parameters and/or algorithms.…”