Investigating Neural Response Variability as a Single-patient Predictor of Successful CBT in Clinical Psychiatry

Many psychiatric patients are not sufficiently improved by current interventions. Functional magnetic imaging brain imaging (fMRI) has proven to be a promising method for predicting treatment outcomes in psychiatric treatment. Individuals moment-to-moment variability have not yet been evaluated as a predictor of treatment of three common forms of mental illness: depression, insomnia and health anxiety. The goal is to investigate whether objective measurements of brain function contribute to a better prediction of a patient's success in treatment than experiences and self-reports, e.g., treatment credibility and patients expectations about the treatment. The prediction model will be tested on internet-delivered CBT (iCBT) for depression, insomnia and social anxiety. Patients in each diagnostic group are asked for participation before treatment. The total number of participants in this study will amount to 225 participants. The goal is that 35% consists of healthy controls and that the remaining part is equally distributed between the three diagnostic patient groups. Being able to better predict how well a psychiatric treatment will work for an individual has great value from both an economic and a treatment perspective. The findings from this study may contribute to increased knowledge about neurobiological complications in mental illness. In the longer term, it can lead to improved routines and help in clinical decision-making when patients should be recommended treatment.

Background: There is extensive evidence that cognitive behavioral therapy (CBT) is an effective method of treating common psychiatric disorders such as depression, social anxiety and insomnia. However, access to CBT is very limited. Furthermore, evidence suggest that internet-delivered CBT (iCBT) is as effective as traditional CBT. However, many psychiatric patients are not sufficiently improved by current interventions. Functional brain imaging has proven to be a promising method for predicting treatment outcomes in psychiatric treatment. Calculations on brain signal variability based on the blood-oxygen-level-dependent, BOLD signal (BOLD-fMRI) is a relatively new technique, shown to accurately predict chronological age and cognitive performance. Preliminary data from the investigators' lab suggest that pre-treatment BOLD-fMRI variability is predictive of CBT outcome in patients with social anxiety disorder. Objectives: The objective of the study is to investigate whether objective measurements of brain function, in comparison with subjective experiences and self-reports (e.g., rating on treatment credibility and patients expectations about the treatment), contribute to better prediction of treatment outcome. The prediction model will be tested on iCBT for three common forms of mental disorders (depression, insomnia and social anxiety). Furthermore, participants will be compared with healthy controls to better understand neurobiological factors that may contribute to mental illness. Preliminary data from the investigators' lab suggest that BOLD-fMRI variability differs between social anxiety disorder participants and healthy individuals. Method: A sample of outpatients scheduled for iCBT treatment for depression, insomnia or social anxiety at the Internet Psychiatry Unit at Karolinska University Hospital in Huddinge, Stockholm, will be invited to participate in the study. In addition to clinical participants, a healthy control group will be recruited via advertisement in social media. The length of the iCBT intervention is 12 weeks, during which participants engage in web-based treatment modules in a sequential manner, guided though a digital messaging system by a licensed psychologist who provides support and feedback on progress and assignments. Measuring instruments: Brain imaging is performed before the patients initiate psychiatric treatment. The brain will be examined with structural and functional magnetic resonance imaging (MRI) using an EPI sequence to capture the BOLD signal (Philips 3-Tesla, 32 channel head-coil). During the online self-referral process and during clinical intake interview, data on a number of potential predictor variables are collected: * Within the social-demographic domain: age; gender; relationship status (dichotomized as being single or not); level of education rated on a 7-point scale (1 = less than 7-9 years in school; 2 = 7-9 years in school; 3 = incomplete vocational or secondary school; 4 = vocational school; 5 = secondary school; 6 = university, started but not completed studies; 7 = completed university studies); employment status dichotomized as working full-time or not; having children, living situation (alone, with kids, with partner, other), if they do physical activity weekly or not, if they smoke or not. * Within the clinical characteristics domain: duration of illness (years since onset), history of psychotropic medication, if currently on psychotropic medication, history of inpatient psychiatric care, history of depression and attempted suicide, presence of a neuropsychiatric diagnosis (e.g. ADHD or autism spectrum disorder), current or previous psychological treatment and current or previous periods of sick leave. * Presence (yes/no) of any comorbid illness is assessed using the the MINI-International Neuropsychiatric Interview (MINI), including bipolar disorder, panic disorder, agoraphobia, social anxiety disorder, mild depressive episode, moderate depressive episode, severe depressive episode, recurrent mild depressive episode, recurrent moderate depressive episode, recurrent severe depressive episode, recurrent depression without current symptoms and dysthymia, generalized anxiety disorder, antisocial personality disorder, anorexia nervosa, bulimia, binge eating disorder and ADHD. * Having a family history of mental illness is assessed prior to treatment, screening for having a family history of dependence / substance abuse, bipolar disorder, depression, minor depression, neuropsychiatric condition, anxiety, panic disorder, psychosis, social anxiety disorder, social anxiety disorder-like symptoms, suicide attempts or suicide completed. * Treatment process factors including perceived treatment credibility measured during the second week in treatment, measured with the Credibility/Expectancy Questionnaire (CEQ) where patients' attitudes to the credibility of the treatment and expectancy regarding treatment effectiveness are rated. Second, treatment adherence is measured at post-treatment, reflecting the degree of use of the ICBT program operationalized as the total number of activated treatment modules. Measurements only administered at screening/pre-treatment: * Patient Health Questionnare 9 (PHQ-9; all participants) * Insomnia Severity Index (ISI; all participants) * Liebowitz Social Anxiety Scale - (LSAS-SR; all participants) * Panic Disorder Severity Scale - (PDSS-SR; all participants) * Social Phobia Inventory (SPIN; all patients) * The Alcohol Use Disorders Identification Test (AUDIT; all participants) * The Drug Use Disorders Identification Test (DUDIT; all participants) * The Adult Self Report Scale (ASRS-V1; all patients) Behavioural measurements before/while undergoing MRI (all participants): * Positive and Negative Affect Schedule (PANAS) * Affect (valence and arousal). "How much pleasure or discomfort are you feeling right now? (1-9)" and "How calm or upset do you feel right now? (1-9)" * Karolinska Sleepiness Scale (KSS) * Subjective unit of discomfort (SUD): fear and distress (0-100) * Self-rated health (SRH-7) * Social Network Index (SNI) * Short Grit Scale (SGI) Predictors of treatment outcome 1. Behavioral predictor(s) of treatment outcome: * Questions about desired and expected treatment improvement. These include a) treatment credibility and b) modified PHQ-9, LSAS-SR and ISI questionnaires where the participants' expectation after completed treatment is self-rated for each item as well as participants' desire of improvement on a 0-100 scale for each item. * Ecological momentary assessments (EMA) with repeated sampling of participants' current behaviors and experiences in real time (4 times/day, during approximately 5 days prior to MRI procedure), including sleep problems, sleepiness, social anxiety, depressive symptoms, fear and discomfort. The purpose is to measure different and varied experiences / feelings over time. Variability in experiences will be compared to the participant's variability in brain function to answer the question of whether emotion variability can predict treatment outcomes. The specific EMA questions include the following: PHQ-2, Karolinska Sleepiness Scale (KSS), item 1 from ISI, item 11 and 15 from LSAS-SR and the subjective units of discomfort (fear and distress). * Emotion categorical ambiguity task 2. Brain-signal variability predictor(s), as derived from the following BOLD-fMRI experiments (all participants): * Face localizer task * Repetition suppression: faces (neutral, sad, angry, fear), scenes, verbal condition (á 60-90 seconds) * Faces * Scenes * Mooney faces * Retinotopy, i.e., traveling-wave method