Research

Original Investigation

Predicting Treatment Response to Cognitive Behavioral

Therapy in Panic Disorder With Agoraphobia by
Integrating Local Neural Information
Tim Hahn, PhD; Tilo Kircher, MD; Benjamin Straube, PhD; Hans-Ulrich Wittchen,
PhD ; Carsten Konrad, MD; Andreas Ströhle, MD; André Wittmann, PhD; Bettina
Pfleiderer, MD, PhD; Andreas Reif, MD; Volker Arolt, MD; Ulrike Lueken, PhD

Supplemental content
IMPORTANCE Although neuroimaging research has made substantial progress in at jamapsychiatry.com
identifying the large-scale neural substrate of anxiety disorders, its value for clinical
application lags behind expectations. Machine-learning approaches have predictive
potential for individual-patient prognostic purposes and might thus aid translational efforts
in psychiatric research.

OBJECTIVE To predict treatment response to cognitive behavioral therapy (CBT) on an

individual-patient level based on functional magnetic resonance imaging data in patients
with panic disorder with agoraphobia (PD/AG).

DESIGN, SETTING, AND PARTICIPANTS We included 49 patients free of medication for at

least 4 weeks and with a primary diagnosis of PD/AG in a longitudinal study performed at
8 clinical research institutes and outpatient centers across Germany. The functional
magnetic resonance imaging study was conducted between July 2007 and March 2010.

INTERVENTIONS Twelve CBT sessions conducted 2 times a week focusing on

behavioral exposure.

MAIN OUTCOMES AND MEASURES Treatment response was defined as exceeding a 50%
reduction in Hamilton Anxiety Rating Scale scores. Blood oxygenation level–dependent
signal was measured during a differential fear-conditioning task. Regional and whole-brain
gaussian process classifiers using a nested leave-one-out cross-validation were used to
predict the treatment response from data acquired before CBT.

RESULTS Although no single brain region was predictive of treatment response, integrating
regional classifiers based on data from the acquisition and the extinction phases of the
fear-conditioning task for the whole brain yielded good predictive performance (accuracy,
82%; sensitivity, 92%; specificity, 72%; P < .001). Data from the acquisition phase enabled
73% correct individual-patient classifications (sensitivity, 80%; specificity, 67%; P < .001),
whereas data from the extinction phase led to an accuracy of 74% (sensitivity, 64%;
specificity, 83%; P < .001). Conservative reanalyses under consideration of potential
confounders yielded nominally lower but comparable accuracy rates (acquisition phase,
70%; extinction phase, 71%; combined, 79%).

CONCLUSIONS AND RELEVANCE Predicting treatment response to CBT based on

functional neuroimaging data in PD/AG is possible with high accuracy on an individual-
patient level. This novel machine-learning approach brings personalized medicine within
Author Affiliations: Author
reach, directly supporting clinical decisions for the selection of treatment options, thus
affiliations are listed at the end of
helping to improve response rates. this article.
Corresponding Author: Ulrike
Lueken, PhD, Institute of Clinical
Psychology and Psychotherapy,
Department of Psychology,
Technische Universität Dresden,
Chemnitzer Straße 46, D-01187
JAMA Psychiatry. 2015;72(1):68-74. doi:10.1001/jamapsychiatry.2014.1741
Dresden, Germany
Published online November 19, 2014. (ulrike.lueken@tu-dresden.de).

68 jamapsychiatry.com

Copyright 2015 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 07/17/2019

 Response to CBT in Panic Disorder With Agoraphobia
A lthough neuroimaging research has made substantial
1,2
progress
in identifying the neural substrate of anxiety disorders, its value for
clinical translational research lags behind expectations. One reason is that
most studies are
based on conventional group analyses. In contrast, a useful bio-marker
has to provide sufficient sensitivity and specificity to predict a given
patient’s status on the individual level. 3,4 Machine-learning algorithms,
such as Support Vector Machines or gaussian process classifiers
(GPC), have shown predictive po-tential for single-subject diagnostic
purposes.5-8
Although results of such proof-of-concept studies appear highly
encouraging, predictive approaches need to facilitate clinical
decisions. To date, however, most studies that use ma-chine learning
merely predict a diagnostic label (a review and notable exceptions
appear in Klöppel et al9 and Sundermann et al10). Among such relevant
clinical decisions, the prediction of treatment outcome is of utmost
importance because iden-tifying individuals who will not benefit from
a therapeutic in-tervention enables early treatment modification and
may of-fer better outcomes for otherwise nonresponsive patients.
Evidence from the field of mood disorders suggests that
neuroimaging biomarkers may aid in classifying disease status 6 and
11
predicting response to pharmacological treatment or cog-nitive
behavioral therapy (CBT).12,13 However, despite the large number of
neuroimaging studies on anxiety disorders, analy-ses on response
markers in this patient group are surprisingly scarce. Application of
regression methods to neuroimaging data from social anxiety
disorder14 has resulted in improved rates of explained variance
compared with clinical outcome para-meters. These improved rates
have been reported when add-ing neuroimaging data (12% vs 41%)
that were tested in a cross-validation procedure using machine-
learning algorithms.
As one of the most disabling anxiety disorders, panic dis-order
(PD) is associated with high individual and societal burdens.15,16
Although CBT has been proven efficient and is rec-ommended as a
first-line treatment for PD,17 response rates are far from satisfying.
Recent data from the national research ini-tiative PANIC-NET
originating from the multicenter trial Mechanism of Action in CBT on
patients with PD with agora-phobia (PD/AG) who were treated with
exposure-based CBT in-dicated positive response rates in
approximately half of the sample.18 Comparable effect sizes have been
reported for other randomized clinical trials as shown in a meta-
analysis by Sánchez-Meca et al. 19 Identifying patients with a higher
risk for nonresponse before treatment could allow for individual-ized
treatment decisions and thus improve response rates.
Fear conditioning is a core process for the development and
maintenance of PD/AG.20 The nature of the learning deficit re-mains
poorly described, but studies point toward the rel-evance of
discriminatory learning, overgeneralization of fear, and attenuated
extinction learning.21-23 On a neural level, brain systems mediating
threat in response to stimuli that signal safety have been reported as a
24
pathophysiological correlate of PD/AG. In line with this correlate,
18,25,26
results in a subset of PD/AG patients from the previous study
indicate that, before treatment, nonresponse was characterized by
enhanced safety-signal processing in brain systems associated with the
pro-cessing of threat (eg, pregenual anterior cingulate cortex, right
jamapsychiatry.com
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ on 07/17/2019
Original Investigation Research
hippocampus, and right amygdala) and altered anterior cin-gulate
cortex–amygdala coupling.26
Although these findings are informative about phenotype
characteristics related to treatment response, they are not ap-plicable to
the individual patient. Herein we aim to predict in-dividual responses
to CBT in the previously described PD/AG sample, 25 thereby
providing information of high clinical rel-evance based on functional
magnetic resonance imaging (fMRI) data. We developed a novel
multivariate pattern classification approach with regional patterns of
neural responses while al-lowing for the integration of local
predictions in the whole brain, thus considering local and whole-brain
neural information to facilitate the prediction of treatment response in
individual patients.
Methods
Multicenter Mechanism of Action in CBT Study
This work is part of the German multicenter trial Mechanism of Action
in CBT.27 After a complete description of the study protocol, written
informed consent was obtained from the par-ticipants, and the protocol
was approved by the local ethics committees in each fMRI center
according to the Declaration of Helsinki. The randomized clinical trial
(isrctn.org identi-fier: ISRCTN80046034) was approved by the ethics
commit-tee of the Medical Faculty of the Technische Universität Dres-
den (agreement EK 16 4082006). The neuroimaging components were
approved by the ethics committee of the Medical Faculty of the
RWTH Aachen University, Aachen (agreement EK 073/07) and at all
local sites. The experimen-tal pharmacology study was approved by
the ethics commit-tee of the state of Berlin (EudraCT 2006-00-4860-
29).
Within the German psychotherapy research network PANIC-NET,
a multicenter randomized clinical trial of CBT for PD/AG patients was
conducted in 8 centers across Germany. Three hundred sixty-nine
patients free of medication for at least 4 weeks who met DSM-IV-TR
criteria for PD/AG were treated with a manualized treatment protocol
or assigned to a wait-ing list that consisted of 12 sessions of CBT 2
times a week fo-cusing on behavioral exposure in situ. Two procedural
vari-ants of CBT (identical in content and dose) were compared that
differed only with regard to therapist guidance during expo-sure
sessions. In the therapist-guided condition, the thera-pist accompanied
the patient during exposure, whereas pa-tients were instructed by the
therapist but performed the exposure on their own in the nonguided
condition. Patients were randomly assigned to 1 of the 2 CBT arms;
both groups exhibited significant symptom reduction after CBT 18;
there-fore, they were combined in the fMRI responder analyses 26
(Table 1). Response was defined as a reduction in Hamilton Anxiety
Rating Scale scores32 exceeding 50% from baseline to posttreatment
assessment.26 Details on the study design and immediate and long-
term treatment effects have been re-ported previously 18,33 (eAppendix
in the Supplement). Four of the 8 centers participated in an fMRI add-
on study, yielding 49 quality-controlled fMRI data sets at baseline. The
fMRI study was conducted between July 2007 and March 2010.
Patient
JAMA PsychiatryJanuary 2015 Volume 72, Number 1 69
 Research Original Investigation Response to CBT in Panic Disorder With Agoraphobia

Table 1. Demographic and Clinical Characteristics of Responders and Nonresponders a

Patient Groupb
Responder Nonresponder Statistic χ2 or t P Value
Characteristic (n = 25) (n = 24) Abbreviations: ASI, Anxiety
Demographic Sensitivity Index; BDI II, Beck
Depression Inventory II;
Female sex, No. (%) 17(68) 16 (67) χ21 = 0.01 .92
CBT, cognitive behavioral
Educational level, No. (%) therapy; CGI, Clinical Global
8y 1(4) 3 (13) Impression Scale; PAS, Panic
10 y 11(44) 11 (46) and Agoraphobia Scale; SIGH-A,
Hamilton Anxiety Rating Scale.
12-13 y 13(52) 9 (38) a
Adopted from Lueken et al.26
No formal degree 0 1 (4) χ23 = 2.71 .44 b
Unless otherwise indicated, data
Age, y 32.80(11.15) 37.83 (9.15) t47 = 1.72 .09 are expressed as mean (SD).
Clinical at baseline The t tests are 2-tailed.
c
From Guy.29
Therapist-guided CBT arm, No. (%) 12(48) 16 (67) χ21 = 1.74 .19
d
From Shear et al.32
No. of diagnoses 2.40(1.29) 2.79 (1.41) t47 = 1.01 .32
e
From Bandelow. 28 Assessed
CGIc 5.32(0.69) 5.46 (0.66) t47 = 0.72 .48
directly before treatment onset,
Total scores whereas the other measures were
obtained at study inclusion. By
SIGH-Ad 24.16(5.39) 25.04 (5.23) t47 = 0.58 .56
definition, the first assessment
e 21.93(7.73) 31.78 (6.54) <.001
PAS t47 = 0.22 served as the baseline value.
ASIf 32.56(8.28) 30.00 (11.56) t47 = −0.89 .38 f
From Reiss et al.30
BDI IIg 17.44(9.93) 17.25 (7.39) t44 = −0.08 .94 g
From Beck et al.31

characteristics are described in Table 1 (a detailed description of
measures of quality control and the fMRI patient flowchart appears
elsewhere25).
fMRI Assessment
The task, fMRI data acquisition, and preprocessing pathways have
25,26,34
been described in detail. We applied a differential fear-
conditioning task. Colored geometric stimuli served as con-ditioned
stimuli (CS) (presentation time, 2000 milliseconds, with a variable
intertrial interval of 4785-7250 milliseconds). An aver-sive auditory
tone (white noise, 100 milliseconds) was used as the unconditioned
stimulus (US) that was pseudorandomly paired with a CS (CS+)
during the acquisition phase, whereas the unpaired CS (CS−) was
never followed by the US (counter-balanced among patients;
reinforcement rate, 50%). During ac-quisition, only those trials in
which no US was delivered (CS+ unpaired) were analyzed. Image
acquisition and analysis path-ways are described in the eAppendix in
the Supplement. For the GPC analyses, we used contrast maps from
individual patients reflecting the main contrasts of interest in a
differential fear-conditioning task (acquisition and extinction phases).
Gaussian Process Classification
Regional GPCs
Using the contrast images for the acquisition and the extinc-tion
phases, whole-brain data from 55 regions drawn from the Harvard-
Oxford Brain Atlas as described in Carter et al 35 were extracted for
each patient. These data were analyzed as de-scribed by Marquand et
al36 using GPCs37 (details are given in the eAppendix in the
Supplement). We predicted a patient’s probability to be a responder
independently for each region based on all voxels within the respective
region using leave-one-out cross-validation.
We evaluated the performance of the 55 regional classifi-ers by
converting the predictive probabilities to categorical pre-
dictions. We applied a threshold that categorized a patient as a
responder or a nonresponder if his or her probability of re-sponse was
larger than 0.5 or smaller than 0.5, respectively. Ac-curacies were
calculated as the ratio of correct predictions to the number of patients
for each GPC. Because the numbers of responders and nonresponders
were not the same, classifica-tion accuracies were calibrated. 38 To
establish whether these regional accuracies are statistically significant,
we ran each clas-sifier 1000 times with randomly permuted labels and
counted the number of permutations that achieved higher accuracy
than the one observed with the true labels. We calculated the P value
by dividing this number by 1000. We corrected for multiple com-
parisons (for 55 regions) using the false discovery rate.
Integration of Regional Predictive Probabilities
We applied a second classifier with the same specifications used for
the local classifiers to integrate the predictive classification
probabilities obtained from the local leave-one-out classifiers. Those
probabilities were used as predictors for the algorithm in order to
determine the classification of a participant. To cal-culate the overall
prediction accuracy of this approach, a nested leave-one-out procedure
was implemented (details are given in the eAppendix in the
Supplement). This procedure ensures complete independence of the
training and the test data set.
The significance of the whole-brain classifier’s predic-tion
accuracy was tested in analogy to the regional classifi-ers. To compare
the best local classifier with the whole-brain classifier, we used a
permutation-based procedure. We first cal-culated the difference
between the best local classifier and the whole-brain classifier. Then,
we approximated the null distri-bution of this difference by computing
the difference be-tween the 2 respective classifiers based on 1000
accuracies under permutation for the best local classifier and the
whole-brain classifier. Finally, we deemed 2 accuracies significantly
different if less than 5% of the absolute accuracy differences

70 JAMA Psychiatry January 2015 Volume 72, Number 1 jamapsychiatry.com

Copyright 2015 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 07/17/2019

 Response to CBT in Panic Disorder With Agoraphobia
Figure. Multivariate Gaussian Process Classifier (GPC) Weight Maps
Acquisition phase
AU
The maps display a region’s contribution to overall classification accuracy
combining the acquisition (left) and the extinction (right) phases. Only the
highest 10% of the weights are displayed. Colors represent the absolute
under permutation were larger than the difference obtained with the
true labels.
Regional and Multivariate Mapping
To determine those brain regions that contributed most to
classification, we derived weight maps from the GPC mod-els as
described in Marquand et al36 and calculated the means for all cross-
validation folds. Although this proce-dure provides a multivariate
estimate of the contribution of each region to classifier performance,
one should be aware that the maps describe a nonlinear multivariate
pattern. Importance scores for each of the regions should be inter-
preted in the context of the entire multivariate pattern. Against this
background, we also present a more readily interpretable, univariate
mapping procedure in which we computed classification accuracies for
each of the 55 regions separately as described above. A region is
shown if the accu-racy estimate for this region exceeded chance level
(P < .05). As described above, we corrected for multiple comparisons
(for 55 regions) using the false discovery rate.
Results
Regional GPCs
Independent regional GPCs from each of the 55 brain regions revealed
significant accuracies for conditioned responses dur-ing acquisition
(CS+ unpaired > CS−) for the inferior frontal gy-rus (pars
triangularis), which was greater than chance level (ac-curacy, 69%; P
= .006) but did not survive multiple comparison correction. For the
contrast reflecting conditioned responses during extinction (CS+ >
CS−), 2 regions were identified: the anterior division of the middle
temporal gyrus (accuracy, 65%; P = .03) and the postcentral gyrus
(accuracy, 67%; P = .03). Again, none of these displayed accuracy
greater than chance after correction for multiple comparisons.
jamapsychiatry.com
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ on 07/17/2019
Original Investigation Research
Extinction phase
whole-brain GPC weights; numbers, arbitrary units (AU) to indicate the
relative contribution of a region to the prediction of treatment response.
Integration of Regional Predictive Probabilities
Integrating the descriptive probabilities from all regional clas-sifiers
for the acquisition and the extinction phases of the fear-conditioning
task yielded high predictive performance (accu-racy, 82%; sensitivity,
92%; specificity, 72%; P < .001). This finding constituted an
improvement in accuracy of 13% (P = .04) compared with the single
best of all local classifiers. Data from the acquisition phase alone
enabled 73% correct in-dividual-patient classifications (sensitivity,
80%; specificity, 67%; P < .001). For the contrast extinction (CS+ >
CS−), inte-grating the descriptive probabilities from all regional
classifi-ers led to an accuracy of 74% (sensitivity, 64%; specificity,
83%; P < .001).
When we considered the patterns learned by the whole-brain
classifier for both contrasts of interest, integrating re-gional
predictions using a GPC algorithm substantially boosted classification
accuracy by integrating GPC-predictive prob-abilities (Figure and
Table 2). These regions did not have the highest single GPC
accuracies but represent the most infor-mative regions in a
multivariate, whole-brain framework for both contrasts combined.
We also conducted a tougher test by taking into account a
nonsignificant trend in age differences and a second assessment just
before the start of the therapy (Panic and Agoraphobia Scale values 28).
Because the acquisition of Panic and Agoraphobia Scale values thus
technically took place when the treatment had already begun, this
analysis might be too conservative. Com-bining both data sources as
described above yielded an accu-racy of 79% (sensitivity, 76%;
specificity, 82%; P < .001), with an accuracy of 70% (sensitivity, 64%;
specificity, 76%; P < .001) for the acquisition phase and 71%
(sensitivity, 74%; specificity, 68%; P < .001) for the extinction phase.
Although nominally lower than the accuracies reported in the first
analyses, the ac-curacies obtained when taking into account age and
Panic and Agoraphobia Scale values did not significantly differ from
those obtained in the first analyses (P > .40).
JAMA PsychiatryJanuary 2015 Volume 72, Number 1 71
 Research Original Investigation Response to CBT in Panic Disorder With Agoraphobia

ficity of 86%, leading to an accuracy of 78% for predicting re-sponse

Table 2. Top 10% Whole-Brain GPC Weights for the
in a small sample of 16 patients with major depres-sion. In a similar
Combined Acquisition and Extinction Contrasts
vein, Doehrmann et al14 demonstrated the potential of fMRI data in
Absolute
Whole- predicting treatment response to CBT in social anxiety disorder.
Brain GPC
Present findings with an accuracy rate of 82% are comparable to the
Region of Interest a Contrastb Weightsc
Gyrus
highest obtained in this field. After controlling for potential
Precentral Acquisition: CS+ 3.19
confounders, the accuracy rate was still in the range of previous
unpaired > CS− reports. Because the Panic and Agoraphobia Scale test was taken
Occipital fusiform Acquisition: CS+ 3.04 within a second assess-ment just before CBT onset and not after study
unpaired > CS−
inclusion, these differences may not reflect true baseline differences,
Frontal orbital cortex Acquisition: CS+ 2.79
unpaired > CS− render-ing the reanalysis possibly too conservative.
Middle temporal gyrus Extinction: CS+ > CS− 2.78
temporo-occipital part Because our study used a different functional paradigm and aimed
Putamen Extinction: CS+ > CS− 2.68 at predicting treatment response in patients with PD/AG and not major
Paracingulate gyrus Extinction: CS+ > CS− 2.60 depression, comparison of regional contribu-tions to overall prediction
Supramarginal gyrus Extinction: CS+ > CS− 2.47 is of limited utility. Costafreda et al13 applied an emotional face
anterior division
perception task in patients with ma-jor depression, whereas the present
Pole
study investigated neural responses during fear conditioning as a
Frontal Extinction: CS+ > CS− 2.23
Occipital Extinction: CS+ > CS− 2.15
pathomechanism in PD/AG.20 Fear conditioning has been reported to
Inferior frontal gyrus pars Acquisition: CS+ 2.15
activate a wide-spread network consisting of subcortical structures,
triangularis unpaired > CS− such as the thalamus, amygdala, and hippocampus, but also the
Postcentral gyrus Acquisition: CS+ 2.03 anterior cingulate cortex, prefrontal and orbitofrontal cortices, and
unpaired > CS−
tem-poral regions.42 In previous reports on this sample,25,43 PD/AG
Abbreviations: CS, conditioned stimulus; GPC, gaussian process classifier.
patients exhibited increased inferior frontal gyrus activity dur-ing
a
Derived from the Harvard-Oxford Brain Atlas described by Carter
differential conditioning, which possibly indicates aber-rant cognitive
et al.35 The atlas contains 55 distinct brain regions.
processing or behavioral inhibition that was sen-sitive to treatment.
b
CS+ indicates the conditioned stimulus that was paired with the
unconditioned stimulus during the acquisition phase (50% of trials), The regions contributing to the predictive performance in the present
whereas the CS− was never followed by the unconditioned stimulus. analysis (combined whole-brain ac-quisition and extinction data)
cValues indicate the relative contribution of this region to the partly overlap with fear network circuits associated with fear
prediction of treatment response.
conditioning, including the orbi-tofrontal cortices and the inferior
frontal gyrus.

From a health economics viewpoint, identifying potential

Discussion
Although considerable improvements in understanding the neural
substrates of anxiety disorders have been achieved, the utility of this
knowledge regarding clinical application is limited. As a proof-of-
concept study in PD/AG, this report demonstrates that fMRI data
acquired before CBT allow for response prediction on an individual-
patient level. When we integrated regional predictive probabilities
from whole-brain fMRI data, patients could be classified correctly as
responders or nonresponders with an accuracy of 82%. Con-trolling
for potential confounders still yielded an accuracy of 79%.
Machine learning has been successfully applied to clas-sify
disease state or to predict the onset of disease in neuro-logical or
psychiatric patient groups.8 For psychiatric disor-ders, classification
accuracies based on fMRI data have been reported in the range of
84%39 to 92%40 for schizophrenia or from 67% to 86% in major
depression.5,6,13,41 Compared with the relatively large number of
studies on disease classifica-tion, surprisingly few exist on prognostic
markers, almost all of which focus on major depression. With a single
exception,13 all of these studies5,11,12 investigated response to
pharmaco-therapy. In the only available study on treatment response to
13
CBT, Costafreda et al observed a sensitivity of 71% and a speci-
nonresponders before an intervention is highly desirable. Pa-tients who
are not likely to respond could be provided with ad-ditional treatment
options. Experimental augmentation strat-egies, such as repetitive
transcranial magnetic stimulation,44 or cognitive enhancers, such as D-
cycloserine,45,46 are under-going evaluation for anxiety disorders that
include PD. A prog-nostic marker with high sensitivity would be
desirable to stratify patients to available add-on therapies and therefore
speed up treatment time and increase response rates. In that sense, a
sen-sitivity of 92% represents a near-maximum detection rate. Pa-
tients predicted not to benefit might respond favorably from intensified
psychotherapy or neurotherapy/pharmacotherapy, which has yet to be
established. Nevertheless, such approaches might aid to govern
treatment regimens in the sense of person-alized medicine approaches.
Future research should apply prog-nostic markers in an independent
sample and evaluate the increment in treatment success and cost-
effectiveness. We ac-knowledge that neuroimaging is not yet a routine
diagnostic tool in psychiatry. Thus, health economic analyses are
needed to bal-ance the costs of additional diagnostic procedures
against the potential benefits of improved outcomes.
The present results are based on a medication-free patient sample
treated solely with CBT, which limits their generaliza-tion to other
populations with PD/AG or anxiety disorders and to alternative (eg,
pharmacological) treatment. To conform to

72 JAMA Psychiatry January 2015 Volume 72, Number 1 jamapsychiatry.com

Copyright 2015 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 07/17/2019

 Response to CBT in Panic Disorder With Agoraphobia Original Investigation Research

previous reports on this sample,18,26,33 we dichotomized treat-
ment response according to the primary outcome. However,
treatment effects are continuous, and information might get lost
using dichotomization. Not having true baseline fMRI (eg, rest-
ing state) data precludes us from interpreting the specificity of
findings for the selected contrasts. The discriminatory pattern
could represent state, task-dependent effects as a function of
response (eg, altered fear conditioning and safety-signal
processing26) or trait effects. Although we used a nested leave-
one-out cross-validation procedure and showed good general-
ization performance on a comparatively large sample of patients
in this study, investigations on heterogeneous real-life samples
are needed to determine whether our model will perform equally
well in other patient samples. However, including 4 separate sites
with 4 different scanners represents a strength suggesting a cer-
tain degree of generalizability. Results do not inform about po-
tential mechanisms underlying processes of change. In that
sense, this approach is complementary to group comparisons
that help identify phenotype characteristics.25,26,43,47
Conclusions
We applied a novel machine-learning approach to investigate
the potential of fMRI data for CBT response prediction on the
individual level in a large sample of PD/AG patients. Findings
showed an accuracy of 82% with a high sensitivity of 92%. The
present work can be understood as a proof-of-concept study
showing the potential utility of our approach for individual-
patient response prediction. This multivariate, individual-
patient classification approach substantially facilitates current
translational efforts aimed at personalizing treatment and sup-
porting clinical decisions when selecting among treatment op-
tions, thus helping to improve response rates.

ARTICLE INFORMATION
Submitted for Publication: January 31, 2014; final
revision received July 17, 2014; accepted
July 21, 2014.
Published Online: November 19, 2014.
doi:10.1001/jamapsychiatry.2014.1741.
Author Affiliations: Department of Cognitive
Psychology II, Goethe University Frankfurt am
Main, Frankfurt am Main, Germany (Hahn);
Department of Psychiatry and Psychotherapy,
Philipps University of Marburg, Marburg, Germany
(Kircher, Straube, Konrad); Institute of Clinical
Psychology and Psychotherapy, Department of
Psychology, Technische Universität Dresden,
Dresden, Germany (Wittchen, Lueken);
Neuroimaging Center, Department of Psychology,
Technische Universität Dresden, Dresden,
Germany (Wittchen, Lueken); Department of
Psychiatry and Psychotherapy, Campus Charité
Mitte, Charité–University Medicine Berlin, Berlin,
Germany (Ströhle, Wittmann); Department of
Clinical Radiology, University Hospital Münster,
Münster, Germany (Pfleiderer); Department of
Psychiatry, Psychosomatics, and Psychotherapy,
University Hospital of Würzburg, Würzburg,
Germany (Reif); Department of Psychiatry and
Psychotherapy, University Hospital Münster,
Münster, Germany (Arolt).
Author Contributions: Drs Hahn and Lueken
had full access to all the data in the study and
take responsibility for the integrity of the data
and the accuracy of the data analysis.
Study concept and design: Kircher, Wittchen,
Konrad, Ströhle, Reif, Arolt, Lueken.
Acquisition, analysis, or interpretation of
data: Hahn, Straube, Wittchen, Konrad,
Ströhle, Wittmann, Pfleiderer, Arolt, Lueken.
Drafting of the manuscript: Hahn, Kircher,
Arolt, Lueken.
Critical revision of the manuscript for
important intellectual content: All authors.
Statistical analysis: Hahn, Wittchen,
Lueken. Obtained funding: Kircher,
Wittchen, Konrad, Ströhle, Arolt.
Administrative, technical, or material support:
Hahn, Kircher, Straube, Wittchen, Konrad,
Ströhle, Wittmann, Reif, Arolt.
Study supervision: Kircher, Straube, Wittchen,
Konrad, Ströhle, Pfleiderer, Lueken.
Conflict of Interest Disclosures: Dr Kircher
received fees for educational programs from Astra-
Zeneca, Bristol-Myers Squibb, Eli Lilly and
Company, Janssen-Cilag, Lundbeck, Pfizer, and
Servier; received travel support/sponsorship for
congresses from Servier; received speaker
honoraria from Janssen-Cilag; and received
research grants from Lundbeck and Pfizer.
Dr Wittchen has served as a general consultant (not
related to the product) for Essex Pharma, Organon,
Pfizer, and Servier and has received grant funding for
his institution from Essex Pharma, Lundbeck, Novartis,
Pfizer, sanofi-aventis, Servier, and Whyet. Dr Konrad
received fees for an educational program from
Esparma GmbH/Aristo Pharma GmbH,
Lilly Deutschland GmbH, MagVenture GmbH, and
Servier Deutschland GmbH. Dr Ströhle received
research funding from the European Commission
(FP6), the German Federal Ministry of Education and
Research, and Lundbeck and speaker honoraria from
AstraZeneca, Boehringer Ingelheim, Eli Lilly and
Company, Lundbeck, Pfizer, UCB, and Wyeth. Dr Reif
participated in 2 noninterventional trials sponsored by
AstraZeneca. Dr Arolt is member
of advisory boards and/or gave presentations for
AstraZeneca, Eli Lilly and Company, Janssen-
Organon, Lundbeck, Pfizer, Servier, and Wyeth
and received research grants from AstraZeneca,
Lundbeck, and Servier. Dr Ströhle has received
educational grants given by the Stifterverband für
die Deutsche Wissenschaft, the Berlin
Brandenburgische Akademie der Wissenschaften,
the Boehringer Ingelheim Fonds, and the Eli Lilly
International Foundation. No other disclosures
were reported.
Funding/Support: This work is part of the
German multicenter trial Mechanism of Action in
CBT (MAC), which is supported by projects
01GV0615 and 01GV0611 (neuroimaging study)
by the German Federal Ministry of Education and
Research (BMBF) as part of the BMBF
Psychotherapy Research Funding Initiative.
Role of the Funder/Sponsor: The funding
source had no role in the design and conduct of
the study; collection, management, analysis, and
interpretation of the data; preparation, review, or
approval of the manuscript; and decision to submit
the manuscript for publication.
Previous Presentation: This paper was
presented as a poster at the 20th Annual Meeting
of the Organization for Human Brain Mapping;
June 9, 2014; Hamburg, Germany. The paper
was presented at the 27th European Colleague in
Neuropsychopharmacology Congress; October
19, 2014; Berlin, Germany.
Additional Information: The following principal
investigators (PIs) participated in the MAC trial (areas
of responsibility): V. Arolt (overall MAC program
coordination), Münster, Germany; H.-U. Wittchen (PI
for the randomized clinical trial [RCT] and manual 4
development), Dresden, Germany; A. Hamm (PI for
psychophysiology), Greifswald, Germany; A. L.
Gerlach (PI for psychophysiology and panic subtypes),
Münster; A. Ströhle (PI for experimental
pharmacology), Berlin, Germany; T. Kircher (PI for
functional neuroimaging), Marburg, Germany; and J.
Deckert (PI for genetics), Würzburg, Germany.
Additional site directors in the RCT component of the
program are G. W. Alpers, Würzburg; T. Fydrich and L.
Fehm, Berlin-Adlershof, Germany; and T. Lang,
Bremen, Germany. The following staff members
participated by site:
C. Melzig, J. Richter, S. Richter, and M. von
Rad, Greifswald (coordinating site for
psychophysiology); H. Bruhn, A. Siegmund, M.
Stoy, and A. Wittmann, Berlin-Charité,
Germany (coordinating center for experimental
pharmacology); I. Schulz, Berlin-Adlershof;
A. Behnken, K. Domschke, A. Ewert, C. Konrad,
B. Pfleiderer, C. Uhlmann, and P. Zwanzger, Münster
(overall MAC program coordination, genetics and
functional neuroimaging); J. Eidecker, S. Koller,
F. Rist, and A. Vossbeck-Elsebusch, Münster
(coordinating site for psychophysiology and
subtyping); B. Drüke, S. Eskens, T. Forkmann, S.
Gauggel, S. Gruber, A. Jansen, T. Kellermann, I.
Reinhardt, and N. Vercamer-Fabri, Marburg/
Aachen, Germany (coordinating center for
functional neuroimaging); F. Einsle, C. Froehlich, A.
T. Gloster, C. Hauke, S. Heinze, M. Hoefler,
U. Lueken, P. Neudeck, S. Preiß, and D. Westphal,
Dresden (coordinating site for data collection,
analysis, and the RCT); A. Reif and C. Gagel,
Würzburg Psychiatry Department (coordinating
center for genetics); J. Duerner, H. Eisenbarth,
A. B. M. Gerdes, H. Krebs, P. Pauli, S. Schad, and
N. Steinhäuser, Würzburg Psychology Department;
V. Bamann, S. Helbig-Lang, A. Kordt, P. Ley,

jamapsychiatry.com JAMA PsychiatryJanuary 2015 Volume 72, Number 1 73

Copyright 2015 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 07/17/2019

 Research Original Investigation
F. Petermann, and E.-M. Schroeder, Bremen.
Additional support was provided by X.
Graehlert and M. Käppler, KKS Dresden
(coordinating center for clinical studies).
REFERENCES
1. Etkin A, Wager TD. Functional neuroimaging of
anxiety: a meta-analysis of emotional processing in
PTSD, social anxiety disorder, and specific phobia.
Am J Psychiatry. 2007;164(10):1476-1488.
2. Sylvester CM, Corbetta M, Raichle ME, et al.
Functional network dysfunction in anxiety and
anxiety disorders. Trends Neurosci.
2012;35(9):527-535.
3. Fu CHY, Costafreda SG. Neuroimaging-based
biomarkers in psychiatry. Can J Psychiatry. 2013;58
(9):499-508.
4. Savitz JB, Rauch SL, Drevets WC. Clinical
application of brain imaging for the diagnosis
of mood disorders: the current state of play.
Mol Psychiatry. 2013;18(5):528-539.
5. Fu CHY, Mourao-Miranda J, Costafreda SG,
et al. Pattern classification of sad facial
processing: toward the development of
neurobiological markers in depression. Biol
Psychiatry. 2008;63(7): 656-662.
6. Hahn T, Marquand AF, Ehlis AC, et al.
Integrating neurobiological markers of depression.
Arch Gen Psychiatry. 2011;68(4):361-368.
7. Mourão-Miranda J, Almeida JR, Hassel S,
et al. Pattern recognition analyses of brain
activation elicited by happy and neutral faces
in unipolar and bipolar depression. Bipolar
Disord. 2012;14(4):451-460.
8. Orrù G, Pettersson-Yeo W, Marquand AF, Sartori G,
Mechelli A. Using Support Vector Machine to identify
imaging biomarkers of neurological and psychiatric
disease: a critical review. Neurosci Biobehav Rev.
2012;36(4):1140-1152.
9. Klöppel S, Abdulkadir A, Jack CR Jr,
Koutsouleris N, Mourão-Miranda J, Vemuri P.
Diagnostic neuroimaging across diseases.
Neuroimage. 2012; 61(2):457-463.
10. Sundermann B, Herr D, Schwindt W,
Pfleiderer B. Multivariate classification of blood
oxygen level–dependent fMRI data with
diagnostic intention: a clinical perspective.
AJNR Am J Neuroradiol. 2014;35(5):848-855.
11. Gong Q, Wu Q, Scarpazza C, et al. Prognostic
prediction of therapeutic response in depression
using high-field MR imaging. Neuroimage. 2011;55
(4):1497-1503.
12. Costafreda SG, Chu C, Ashburner J, Fu CHY.
Prognostic and diagnostic potential of the structural
neuroanatomy of depression. PLoS One. 2009;4
(7):e6353. doi:10.1371/journal.pone.0006353.
13. Costafreda SG, Khanna A, Mourao-Miranda J,
Fu CHY. Neural correlates of sad faces predict
clinical remission to cognitive behavioural therapy
in depression. Neuroreport. 2009;20(7):637-641.
14. Doehrmann O, Ghosh SS, Polli FE, et al.
Predicting treatment response in social anxiety
disorder from functional magnetic resonance
imaging. JAMA Psychiatry. 2013;70(1):87-97.
15. Wittchen HU, Jacobi F, Rehm J, et al. The
size and burden of mental disorders and other
disorders of the brain in Europe 2010. Eur
Neuropsychopharmacol. 2011;21(9):655-679.
74 JAMA Psychiatry January 2015 Volume 72, Number 1
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ on 07/17/2019
16. Nutt DJ. The full cost and burden of disorders
of the brain in Europe exposed for the first time.
Eur Neuropsychopharmacol. 2011;21(10):715-717.
17. McHugh RK, Smits JA, Otto MW. Empirically
supported treatments for panic disorder.
Psychiatr Clin North Am. 2009;32(3):593-610.
18. Gloster AT, Wittchen HU, Einsle F, et al.
Psychological treatment for panic disorder
with agoraphobia: a randomized controlled
trial to examine the role of therapist-guided
exposure in situ in CBT. J Consult Clin
Psychol. 2011;79(3):406-420.
19. Sánchez-Meca J, Rosa-Alcázar AI, Marín-
Martínez F, Gómez-Conesa A. Psychological
treatment of panic disorder with or without
agoraphobia: a meta-analysis. Clin Psychol
Rev. 2010;30(1):37-50.
20. Bouton ME, Mineka S, Barlow DH. A modern
learning theory perspective on the etiology of panic
disorder. Psychol Rev. 2001;108(1):4-32.
21. Lissek S, Powers AS, McClure EB, et al. Classical
fear conditioning in the anxiety disorders:
a meta-analysis. Behav Res Ther.
2005;43(11):1391-1424.
22. Lissek S, Rabin S, Heller RE, et al.
Overgeneralization of conditioned fear as a
pathogenic marker of panic disorder. Am J
Psychiatry. 2010;167(1):47-55.
23. Michael T, Blechert J, Vriends N, Margraf J,
Wilhelm FH. Fear conditioning in panic
disorder: enhanced resistance to extinction. J
Abnorm Psychol. 2007;116(3):612-617.
24. Tuescher O, Protopopescu X, Pan H, et al.
Differential activity of subgenual cingulate and
brainstem in panic disorder and PTSD. J
Anxiety Disord. 2011;25(2):251-257.
25. Kircher T, Arolt V, Jansen A, et al. Effect of
cognitive-behavioral therapy on neural
correlates of fear conditioning in panic disorder.
Biol Psychiatry. 2013;73(1):93-101.
26. Lueken U, Straube B, Konrad C, et al. Neural
substrates of treatment response to cognitive-
behavioral therapy in panic disorder with agoraphobia.
Am J Psychiatry. 2013;170(11):1345-1355.
27. Gloster AT, Wittchen HU, Einsle F, et al.
Mechanism of Action in CBT (MAC): methods
of a multi-center randomized controlled trial in
369 patients with panic disorder and
agoraphobia. Eur Arch Psychiatry Clin
Neurosci. 2009;259(suppl 2): S155-S166.
28. Bandelow B. Panic and Agoraphobia
Scale (PAS). Ashland, OH: Hogrefe & Huber
Publishers; 1999.
29. Guy W. ECDEU Assessment Manual for
Psychopharmacology Review. Rockville, MD: US
National Institute of Health,
Psychopharmacology Research Branch; 1976.
30. Reiss S, Peterson RA, Gursky DM, McNally RJ.
Anxiety sensitivity, anxiety frequency and the
prediction of fearfulness. Behav Res Ther. 1986;24
(1):1-8.
31. Beck AT, Steer RA, Brown GK. Beck
Depression Inventory. 2nd ed. San Antonio,
TX: Psychological Corp; 1996.
32. Shear MK, Vander Bilt J, Rucci P, et al.
Reliability and validity of a structured interview
guide for the Hamilton Anxiety Rating Scale
(SIGH-A). Depress Anxiety. 2001;13(4):166-178.
Response to CBT in Panic Disorder With Agoraphobia
33. Gloster AT, Hauke C, Höfler M, et al. Long-
term stability of cognitive behavioral therapy
effects for panic disorder with agoraphobia: a
two-year follow-up study. Behav Res Ther.
2013;51(12):830-839.
34. Reinhardt I, Jansen A, Kellermann T, et al. Neural
correlates of aversive conditioning. Eur Arch
Psychiatry Clin Neurosci. 2010;260(6):443-453.
35. Carter RM, Bowling DL, Reeck C, Huettel
SA. A distinct role of the temporal-parietal
junction in predicting socially guided
decisions. Science. 2012; 337(6090):109-111.
36. Marquand A, Howard M, Brammer M, Chu C,
Coen S, Mourão-Miranda J. Quantitative
prediction of subjective pain intensity from whole-
brain fMRI data using gaussian processes.
Neuroimage. 2010; 49(3):2178-2189.
37. Rasmussen CE, Williams CKI. Gaussian
Processes for Machine Learning.
Cambridge, MA: MIT Press; 2006.
38. Hahn T, Marquand AF, Plichta MM, et al.
A novel approach to probabilistic biomarker-based classification
using functional near-infrared spectroscopy. Hum Brain Mapp.
2013;34(5):1102-1114.
39. Shen H, Wang L, Liu Y, Hu D. Discriminative
analysis of resting-state functional connectivity
patterns of schizophrenia using low dimensional
embedding of fMRI. Neuroimage.
2010;49(4):3110-3121.
40. Costafreda SG, Fu CHY, Picchioni M, et al.
Pattern of neural responses to verbal fluency
shows diagnostic specificity for schizophrenia and
bipolar disorder. BMC Psychiatry. 2011;11(18):18.
doi:10.1186 /1471-244X-11-18.
41. Marquand AF, Mourão-Miranda J, Brammer
MJ, Cleare AJ, Fu CHY. Neuroanatomy of verbal
working memory as a diagnostic biomarker for
depression. Neuroreport. 2008;19(15):1507-1511.
42. Sehlmeyer C, Schöning S, Zwitserlood P,
et al. Human fear conditioning and extinction
in neuroimaging: a systematic review. PLoS
One. 2009;4(6):e5865.
doi:10.1371/journal.pone .0005865.
43. Lueken U, Straube B, Reinhardt I, et al.
Altered top-down and bottom-up processing of
fear conditioning in panic disorder with
agoraphobia. Psychol Med. 2014;44(2):381-394.
44. Mantovani A, Aly M, Dagan Y, Allart A, Lisanby
SH. Randomized sham controlled trial of
repetitive transcranial magnetic stimulation to the
dorsolateral prefrontal cortex for the treatment of
panic disorder with comorbid major depression. J
Affect Disord. 2013;144(1-2):153-159.
45. Otto MW, Tolin DF, Simon NM, et al.
Efficacy of D-cycloserine for enhancing response
to cognitive-behavior therapy for panic disorder.
Biol Psychiatry. 2010;67(4):365-370.
46. Siegmund A, Golfels F, Finck C, et al. D-
Cycloserine does not improve but might slightly speed
up the outcome of in-vivo exposure therapy in
patients with severe agoraphobia and panic disorder
in a randomized double blind clinical trial. J Psychiatr
Res. 2011;45(8):1042-1047.
47. Reif A, Richter J, Straube B, et al. MAOA
and mechanisms of panic disorder revisited.
Mol Psychiatry. 2014;19(1):122-128.
jamapsychiatry.com