Does the brain change with psychotherapy?
Why do we care about identifying the brain mediators of symptom change
Agenda: 2 Longitudinal Neuroimaging Studies of Symptom Change following Psychotherapy
Every year, 700,000 children experience substantiated abuse http://www.traumaline1.com/node/74
Diagnostic and Statistical Manual Version V (DSM-V): Criteria for Diagnosing PTSD DSM-V: Trauma and Stressor-Related Disorders A. Trauma exposure B. Re-experiencing C. Avoidance D. Negative mood and cognitions E. Altered arousal and reactivity F. Duration > 1 month G. Distress/impairment Preschool Subtype Dissociative Subtype Symptoms of Dissociative Subtype Depersonalization in which the person feels like an outside observer or detached from oneself and is typically accompanied by an attenuation of the emotional experience. Derealization in which the world seems unreal, distant or distorted and often associated with the experience of decreased emotional intensity.
Dissociation as avoidance coping
Why should we study the neural correlates of dissociation in pediatric PTSD? To understand how dissociation differs from and interacts with other PTSD symptoms in determining brain abnormalities Implications for modifying therapeutic approaches for patients with prominent dissociation Information for patients
Study Design MRI Scan 1 MRI Scan 2 adolescents with symptoms of PTSD Measure symptoms of dissociation and PTSD 20 weeks of Traumafocused CBT Measure symptoms of dissociation and PTSD
Trauma-focused cognitive behavioral therapy (TF-CBT) helps kids who have PTSD
Facial Expression Task = Implicit Emotion Perception Is the person in the picture a girl or a boy? Press 1 if girl; Press 2 if boy 8 happy 8 Neutral 8 Angry 8 Scrambled 24s 24s 24s 24s 24s 432s
Neural response to Trauma-focused cognitive behavioral therapy (TF-CBT) for pediatric PTSD *Reasons for Quitting before Scan: - N= 3 transportation issues - N= 1 decided to start medication - N=1 moved far away - N=1 decided against therapy Patients Screened N= 74 Enrolled N=34 Scanned and assigned to treatment N=21 Completed N= 20 Screen Fail N= 40 Dropped after 1 or more sessions N= 1 Pre-pubertal No trauma Taking medications Quit before scan N= 6* Refused MRI N= 2 Not eligible N= 5 Transportation issues could not tolerate MRI -not PTSD -Current substance dependence
ADOLESCENT DISSOCIATIVE EXPERIENCES SCALE SCORES DISSOCIATIVE SYMPTOMS AT BASELINE 6 5 4 3.7= significant dissociation 3 2 1 0 PATIENTS CONTROLS mean score: 2.8 mean score: 0.57 t= 5.46, p= 0.000003
Baseline: Dissociation increases as PTSD increases N=20; Pearson s r=.45, p=.046 PTSD symptoms (UCLA PTSD R.I.) At Baseline Dissociative Symptoms (ADES) at Baseline
40 30 Longitudinal change in PTSD symptoms from baseline to end of treatment 20 10 0-10 -20 0 1 2 3 4 5 6 Dissociative Symptoms (ADES) at Baseline
Changes in Dissociative Symptoms from Baseline to End of Treatment 6 5 Change in Adolescent Dissociative Experiences Scale (ADES) 4 3 2 1 0 BASELINE Mean=2.8 END OF TREATMENT Mean=1.8 Repeated measures t-test t= 4.91, p=.0005
Changes in PTSD symptom severity from baseline to end of treatment 60 UCLA PTSD Reaction Index Scores (average of parent and child ratings) 50 40 30 20 10 0 Baseline End of Treatment Mean=39 mean=23 Repeated measures t-test t=6.10, p=.000007
Longitudinal change in PTSD symptoms from baseline to end of treatment Longitudinal change in Dissociative Symptoms
Brain model of fear conditioning with relevance to PTSD Brain model of dissociation as emotional over-modulation
Pediatric PTSD: GREATER activation in amygdala, insula LESS activation in the anterior dorsolateral prefrontal cortex (Brodmann s Area 10)
Amygdala hippocampus fusiform medial frontal Fear /emotion memory/context face perception fear extinction
Baseline: Correlation with Dissociative symptoms (Angry faces, p=.001 voxel level threshold) More dissociation is correlated with.. Less amygdala less posterior cingulate less visual cortex Emotion/fear autobiographical memory basic visual processing More dorsal prefrontal activation executive control and selection of response
Greater dissociation at baseline is correlated with lower connectivity between amygdala and insula (p<.001, voxel wise) Insula: interoceptive cortex
Baseline: correlation with PTSD symptom severity (Angry faces, p=.001 voxel level threshold) Subgenual cingulate (self-referential thought) Parahippocampal gyrus (input to memory system) Anterior superior frontal (BA 10) Left inferior parietal gyrus (attention)
Improvement in dissociation Longitudinal increase in Amygdala hippocampus medial frontal orbital frontal fusiform gyrus (Emotion) (memory) (fear extinction) (value judgement) (face processing) Longitudinal decrease in Inferior parietal (Attention)
Conclusions Symptoms of Dissociation can improve after TF-CBT (unmodified) Dissociation not so mysterious- looks like emotion modulation in the brain, demonstrating that these patients are able to modulate As Dissociation improves, emotion brain regions respond MORE, frontal regions respond more, and PTSD symptoms improve Share this information with your dissociative patients?
P.I.= Kiki Chang, MD P. I. =David Miklowitz, Ph.D.
Can an early intervention alter the course of mood symptoms in kids at high-risk for bipolar disorder? How are longitudinal changes in mood symptoms reflected in brain function?
Criteria for high-risk for bipolar disorder
Miklowitz DJ (2010). Bipolar Disorder: A Family-Focused Treatment Approach., 2 nd Ed. NY: Guilford Press.
Results of Nine Randomized Trials of FFT plus Medications
Baseline and End of Treatment Clinical Measures Task performed in the scanner
AGAIN, Why do we care about identifying the brain mediators of symptom change
From Fox et al., 2007
Traveling Subjects Calibration Study The travelers=3 volunteers scanned at both UCLA and Stanford At each site, each person scanned on 2 days (Day 1 and Day 2) fmri analysis of finger tapping task showed that location of activation and mean percent activation for Day 1 versus Day 2 was comparable across sites
Total of N=47 subjects have both baseline (BL) and end of treatment (EOT) scans (combined across site (Stanford and UCLA) and treatment type (FFT and EC))
Variable FFT (N=55)* EC (N=51)* Age, yrs, mean ± (SD) 12.9 ± 2.7 13.0 ± 2.4 Female n (%) 30 (54.6%) 36 (70.6%) Non-white, n (%) 13 (23.6%) 9 (17.7%) CDRS entry, mean ± (SD) 45.8 ± 12.4 49.0 ± 16.1 YMRS, entry, mean ± (SD) 12.4 ± 6.7 13.8 ± 7.3 Primary Diagnosis, n (%) Major Depression 33 (60.0%) 26 (50.1%) BP NOS 22 (40.0%) 25 (49.0%) Comorbid disorders, n (%) 45 (81.8%) 39 (76.5%) Anxiety 26 (47.3%) 23 (45.1%) ADHD 20 (36.4%) 15 (29.4%) ODD 13 (23.6%) 11 (21.6%)
Research Questions for studying brain networks associated with symptom change Where in the brain was change in activation correlated with symptom improvement? seed region What networks are linked to this brain region? Are there differences between treatments?
Correlations between longitudinal changes in activation and changes in symptoms of depression (CDRS); p=.005 k=20 Fearful Faces Happy Faces Left hippocampus Less active as symptoms improve Left hippocampus Less active as symptoms improve Left insula: more active as symptoms improve Dorsal ACC Less active as Symptoms improve
Meta-analysis of brain structure 1728 MDD and 7199 HC from 15 samples worldwide Significantly smaller hippocampus in MDD, driven by recurrent MDD Earlier age of onset associated with smaller hippocampus
Correlations between longitudinal changes in activation and longitudinal changes in symptoms of mania (YMRS) (p=.005, k=20) Fear faces MORE ACTIVE AS SYMPTOMS IMPROVE: Right DLPFC Pregenual cingulate Happy faces LESS ACTIVE AS SYMPTOMS IMPROVE: Left hippocampus Left amygdala
Improvement in symptoms of mania is correlated with longitudinal increase in activation in right DLPFC Brodmann s areas 9 and 46 and 45 significant cluster (P,.001, k=92) ; Happy Faces, N=47
longitudinal improvement-ymrs-longitudinal worsening
Our pilot study found that increasing right DLPFC activation is correlated with improvement in symptoms of mania significant correlation between increasing activation in the dorsolateral prefrontal cortex (DLPFC) and improvement in symptoms of mania from pre- to posttreatment (Spearman's rho =.72, p =.008). Changes in brain activation following psychotherapy for youth with mood dysregulation at familial risk for bipolar disorder PROGRESS IN NEURO- PSYCHOPHARMACOLOGY & BIOLOGICAL PSYCHIATRY Garrett, A. S., Miklowitz, D. J., Howe, M. E., Singh, M. K., Acquaye, T. K., Hawkey, C. G., Glover, G. H., Reiss, A. L., Chang, K. D. 2015; 56: 215-220
Longitudinal Increase in Right DLPFC connectivity from Pre to Post Treatment (Repeated Measures T test, P=.005, k=20) Left hippocampus Left DLPFC Right orbitofrontal Stress/memory executive control/working memory emotion regulation
FFT versus EC treatments Comparison
FFT and EC treatments: similar mechanisms to improve symptoms? Symptoms of Mania (YMRS) (correlation within treatment group, p=.005, k=20) Family Focused Therapy N=23 Enhanced Care N=24 Dorsolateral Prefrontal Cortex BA 46: More active as symptoms improve Hippocampus: Less active as symptoms improve
FFT and EC treatments: different mechanisms to improve symptoms? Symptoms of Depression (CDRS) (correlation within treatment group, p=.005, k=20) Family Focused Therapy Enhanced Care Medial Prefrontal (BA 10): Decreases as symptoms improve Left hippocampus: Decreases as symptoms improve
Treatment Effects: Group x Time p=.005, k=20 FFT > EC : medial prefrontal & Pregenual cingulate EC > FFT: Left insula 0.08 0.06 0.04 0.02 0-0.02-0.04-0.06-0.08-0.1-0.12 ECcing1 ECcing2 FFTcing1 FFTcing2 0-0.02-0.04-0.06-0.08-0.1-0.12-0.14
Conclusions The right dorsolateral prefrontal cortex is consistently implicated in symptom improvement From before to after treatment, DLPFC becomes functionally connected to many other regions (emotion, memory, self-referential processing) Left hippocampus also changes with treatment Both types of psychotherapy appear to lead to symptom change via the DLPFC and hippocampus However, change in the medial prefrontal cortex has opposite effect in FFT versus EC
PTSD Research Team Mentors: Stewart Agras, M.D. Judith Cohen, M.D. Allan Reiss, M.D. Victor Carrion, M.D. Kiki Chang, M.D. Therapists: Sanno Zack, Ph.D. Miriam Revne Rowan Jessica Delman Jessica Nagel John Rettger Kristin LaCross Research Assistants Alexandra Ishak Maisi Mayo Cynthia Medina, M.A. Bipolar Disorders Research Team P.I. s David Miklowitz, PhD (UCLA) and Kiki Chang M.D. (SU) Patty Walshaw Ph.D (Co-Investigator) Research Assistants: Jessica Hernandez, B.A., Daniella DeGeorge B.A., Casey Armstrong B.A. Consultants: Allan Reiss and the Center for Interdisciplinary Brain Sciences Research at Stanford Gary Glover and the Lucas Center for Biomedical Imaging at Stanford; Susan Bookheimer at UCLA Dept of Psychiatry