Content uploaded by Michelle Hampson
Author content
All content in this area was uploaded by Michelle Hampson
Content may be subject to copyright.
Orbitofrontal cortex neurofeedback produces lasting
changes in contamination anxiety and resting-state
connectivity
D Scheinost
1
, T Stoica
2
, J Saksa
3
, X Papademetris
1,2
, RT Constable
1,2,4
, C Pittenger
3,5,6
and M Hampson
2
Anxiety is a core human emotion but can become pathologically dysregulated. We used functional magnetic resonance imaging
(fMRI) neurofeedback (NF) to noninvasively alter patterns of brain connectivity, as measured by resting-state fMRI, and to reduce
contamination anxiety. Activity of a region of the orbitofrontal cortex associated with contamination anxiety was measured in real
time and provided to subjects with significant but subclinical anxiety as a NF signal, permitting them to learn tomodulate the target
brain region. NF altered network connectivity of brain regions involved in anxiety regulation: subjects exhibited reduced resting-
state connectivity in limbic circuitry and increased connectivity in the dorsolateral prefrontal cortex. NF has been shown to alter
brain connectivity in other contexts, but it has been unclear whether these changes persist; critically, we observed changes in
connectivity several days after the completion of NF training, demonstrating that such training can lead to lasting modifications of
brain functional architecture. Training also increased subjects’ control over contamination anxiety several days after the
completion of NF training. Changes in resting-state connectivity in the target orbitofrontal region correlated with these
improvements in anxiety. Matched subjects undergoing a sham feedback control task showed neither a reorganization of resting-
state functional connectivity nor an improvement in anxiety. These data suggest that NF can enable enhanced control over anxiety
by persistently reorganizing relevant brain networks and thus support the potential of NF as a clinically useful therapy.
Translational Psychiatry (2013) 3, e250; doi:10.1038/tp.2013.24; published online 30 April 2013
Introduction
Normal and pathological patterns of behavior and thought
correspond to the activity of particular brain circuits. Interven-
tions that alter patterns of behavior and thought therefore
must act on the organization of the underlying circuits; some
clinical interventions, such as deep brain stimulation, do so
explicitly through anatomically targeted manipulations of brain
function.
1
The ability to manipulate targeted brain circuits of
relevance to particular behavior patterns in a non-invasive
manner would be of immense interest and clinical utility.
Poorly controlled anxiety reduces the quality of life of many
healthy individuals and is a key symptom of numerous
neuropsychiatric conditions. Contamination anxiety, in part-
icular, is prevalent in the healthy population and is a common
symptom in obsessive-compulsive disorder (OCD).
2
Pharma-
cological and behavioral interventions are widely used in the
treatment of anxiety and of OCD, but for many individuals these
are of little efficacy or are associated with troublesome side
effects. In extreme cases, invasive anatomically targeted inter-
ventions are sometimes used for OCD and can be effective.
1
Neurofeedback (NF) describes the process of learning to
control neural processes via an explicit feedback signal. Real-
time functional magnetic resonance imaging (rt-fMRI) NF is a
novel approach in which subjects can receive direct feedback
regarding neural activity, as reflected in the BOLD signal, of a
defined brain region. Recent studies have reported success in
training subjects to manipulate activity in specific target
brain regions using this approach.
3–5
Several studies have
also reported that such training can translate into changes in
behavioral measures
6–9
or clinical symptoms.
10,11
Learned
control over a specific region of the brain has been shown to
lead to alterations in brain networks, with documented
alterations in brain function lasting only during
3
or immediately
after the presentation of feedback.
5,12
For NF-induced
changes in brain networks to be of clinical utility, it is essential
that such changes persist beyond the scanning session in
which training occurred.
We asked whether NF can produce functional alterations in
the circuitry associated with contamination anxiety, thereby
reducing it, and whether such changes can persist over the
days following training. Activation in the orbitofrontal cortex
(OFC) has been implicated in contamination anxiety in healthy
individuals
13
and in patients with OCD.
14
We hypothesized
that feedback training that permitted subjects to manipulate
neural activity in the OFC would result in a reorganization
of associated functional brain networks, as measured by
resting-state fMRI, and in a reduction in experienced anxiety.
We investigated this in individuals with significant contamina-
tion anxiety but without any clinical diagnosis of an anxiety
disorder.
1
Department of Biomedical Engineering, Yale University, New Haven, CT, USA;
2
Department of Diagnostic Radiology, Yale School of Medicine, New Haven, CT, USA;
3
Department of Psychiatry, Yale School of Medicine, New Haven, CT, USA;
4
Department of Neurosurgery, Yale School of Medicine, New Haven, CT, USA;
5
Department
of Psychology, Yale University, New Haven, CT, USA and
6
Child Study Center, Yale School of Medicine, New Haven, CT, USA
Correspondence: Dr M Hampson, Magnetic Resonance Research Center (MRRC), Yale School of Medicine, The Anlyan Center, N121, 300 Cedar Street, PO Box
208043, New Haven, CT 06520-8043, USA.
E-mail: michelle.hampson@yale.edu
Received 26 November 2012; revised 24 January 2013; accepted 18 February 2013
Keywords: contamination anxiety; neurofeedback; obsessive-compulsive disorder; real-time fMRI; resting connectivity
Citation: Transl Psychiatry (2013) 3, e250; doi:10.1038/tp.2013.24
&
2013 Macmillan Publishers Limited All rights reserved 2158-3188/13
www.nature.com/tp
Materials and methods
Subject recruitment. Subjects with high contamination-
related anxiety were recruited and consented in accordance
with a protocol reviewed and approved by the Yale University
Human Research Protection Program. Individuals with
a history of neuropsychiatric disorder or currently using
medication (other than antibiotics or birth control) were
excluded from the study. All included subjects had a score of
X8 on the Padua Inventory—Washington State University
Revision
15
—Contamination Obsessions and Washing Compul-
sions subscale. A total of 12 subjects were recruited for the NF
group, and 11 subjects, matched for age and gender, were
recruited for the sham-feedback (SF) control group. One NF
subject was removed from analysis for not following instructions
and was not matched with a SF subject. Further, one NF
subject and the matched SF subject were removed from
analysis due to an error in the localization of the target region
discovered after their data collection (but before assessment of
imaging outcome measures). A total of 10 NF and 10 matched
SF subjects were included in the final analysis (4 females in
each group). Each of the 20 subjects had four separate
scanning sessions over the course of 3 weeks (80 scanning
session in total were analyzed). There was no significant
difference in age or scores on the Padua Contamination
subscale between these two groups.
Study protocol. Our rt-fMRI NF protocol has been
described in detail in a methods publication
16
using the
system described in Scheinost et al.
17
We provide here a
general description but refer the reader there for more
specific information. Briefly, the experimental paradigm
consisted of four MRI scanning sessions, spaced roughly
half a week apart. The first visit began with an imaging
session involving high-resolution anatomical images, two
resting-state functional runs (5 min each) and a functional
localizer that alternately presented neutral and contamina-
tion-related images. The localizer was used to identify the
target region of interest, defined as the 30 OFC voxels,
bilaterally, that were maximally differentially activated by
contamination-related stimuli. After this first imaging session,
subjects met with a clinical psychologist to help them develop
initial strategies for manipulating their brain activity that could
later be refined via NF. The second visit began with an out-of-
magnet assessment of how well subjects could control their
anxiety. In this assessment session, subjects viewed 25
contamination-related images and were instructed to try to
minimize their anxiety in response to each image and to
indicate on a 1–5 scale the anxiety they experienced.
Subjects then participated in a 90-min NF (or SF) imaging
session involving 10 functional runs during which they tried to
control brain activity, as reported to them in a line graph on
the visual display, while viewing contamination-related and
neutral images (as further described below). The first and last
two runs were performed without NF; the middle six were
performed with NF. During the NF runs, subjects were
instructed to try to learn what worked the best for controlling
their OFC activity. During the non-NF runs, subjects were
instructed to do whatever they felt worked best for controlling
OFC activity. The non-NF runs, referred to as control task
runs, were used to assess changes in the OFC in the
absence of feedback. The third visit was identical to the
second. The fourth visit consisted of a final assessment of
anxiety evoked by contamination-associated stimuli and a
final imaging session in which two resting-state runs were
collected to assess changes in functional connectivity. The
precise scheduling of the visits varied based on availability of
scan slots, but the final visit was always scheduled 1–7 days
after the third visit. The average was 3 days, and there was
no significant difference between the sham and real NF
subjects in this variable. As the final visit, in which we
assessed the effects of the NF on both anxiety and brain
functional connectivity, occurred several days after feedback
training, the observed effects do not reflect transient results
of feedback training but rather reflect changes that persisted
for several days. In total, each subject participated in four
MRI sessions during the experiment and 480 scanning
sessions were performed in this experiment.
Subjects never viewed the same stimulus twice. The sets of
contamination and neutral images used at different time
points in the protocol were balanced for induced anxiety, as
described previously.
16
Task during NF sessions. During NF, subjects viewed an
arrow on the left side of the screen that cued them regarding
the current task. A red up-arrow indicated they should try
to increase activity in their OFC, a blue down-arrow
indicated they should try to decrease activity in the region,
and a white forward-arrow indicated they should rest and not
try to control the region. To the right of this arrow, a large
contamination-related image was shown during the increase
and decrease conditions and a neutral image was shown
during the rest condition. The arrows and images changed
every 26 s, cycling through the three conditions. During NF
runs, a line graph was included at the bottom of the screen
indicating the activity in the subject’s OFC. The line was
color-coded to indicate the current task and was updated
after each newly acquired volume was analyzed, approxi-
mately every 2 s. Although the NF group viewed the
actual time course of the OFC, the SF group viewed the
time course of the corresponding run performed previously of
their paired NF subject. As a result, the SF subjects viewed
exactly the same stimuli as their paired NF subjects. An
example of what subjects viewed during the NF runs is
provided in Figure 1.
MR imaging protocol. All imaging was done on a 1.5-T
Siemens Sonata scanner (Siemens Medical Systems, Erlan-
gen, Germany). A sequence designed to optimize signal in
the OFC was used for all functional data collection
(repetition time ¼2000 ms, echo time ¼30 ms, flip angle ¼80,
bandwidth ¼2604, 200 mm field of vi ew for 3.1 mm isotro pic
voxels, 31 axial-oblique slices covering the OFC and brain
above).
Preprocessing. Images were slice time corrected using
sinc interpolation in Matlab (www.mathworks.com) and
motion corrected using SPM5 (http://www.fil.ion.ucl.ac.uk/
spm/software/spm5/). Unless noted, all analyses were
conducted using BioImage Suite.
18
Neurofeedback modulates connectivity and anxiety
D Scheinost et al
2
Translational Psychiatry
Resting-state connectivity analysis. Several covariates of
no interest were regressed from the data, including linear and
quadratic drift, six rigid-body motion parameters and mean
cerebral–spinal fluid, white matter and global signals. The data
were low-pass filtered (approximate cutoff frequency ¼0.12
Hz) and a gray matter mask was applied to the data so that
only voxels in the gray matter were included. The network
measure of degree
19,20
was computed for each voxel.
Comparison of degree maps before and after NF allows
exploration of functional connectivity changes, unbiased by a
priori assumptions regarding regions of interest. First, the time
series for a voxel was correlated with every other voxel in the
gray matter with degree defined as the number of connection
with a correlation greater than r¼0.40. The process was
repeated for each voxel. Degree maps were normalized as
described previously.
21,22
For group comparisons, single
subject results were smoothed (6 mm Gaussian kernel) and
warped to common space through a series of linear and non-
linear registrations as described previously.
19,22
Control task analysis. Control task data collected immedi-
ately before and after NF on the second and third
experimental days were used to assess changes in the target
region of interest. Task regressors for increase and decrease
blocks were convolved with a hemodynamic response
function and included in the design matrix of a general linear
model. Motion parameters and drift terms were added as
regressors of no interest. Activity in the target OFC voxels
during decrease blocks was subtracted from activity during
increase blocks to yield an estimate of control over the region
during control task runs.
Statistical analyses
Resting connectivity data. Group level statistics were per-
formed in a voxel-wise manner to identify regions showing
changes in functional connectivity after rt-fMRI training. For
each group, paired differences in the resting data collected
before and after training were identified using Wilcoxon’s
signed-rank test. Between-group differences were identified
using Wilcoxon’s rank-sum test, implemented in Analysis of
Functional NeuroImages (AFNI) (http://afni.nimh.nih.gov/
afni). To assess the relationship between behavior and
intrinsic connectivity patterns, as previous studies have
done,
23
we performed a voxel-by-voxel Pearson correlation
analysis between the change in behavior and change in
degree. Significance was assessed at a Po0.05 level after
correcting for multiple comparisons across the grey matter
via AFNI’s AlphaSim program. To check for possible
confounds caused by group differences in head motion,
24
we computed the mean frame-to-frame displacement for
each subject across all time points. Wilcoxon’s rank-sum
tests revealed no group difference in motion (P¼0.67,
U¼44) and no difference in motion between the first and last
imaging session (P¼0.57, U¼42).
Behavioral data. The anxiety ratings were averaged across
stimuli in each assessment session. The average in the final
session was subtracted from the average in the first session
to yield an estimate of how much each subject increased
control over their contamination anxiety. Wilcoxon’s signed-
rank tests were used to assess changes within each group,
and a Wilcoxon’s rank-sum test was used to contrast
changes between groups. Because we hypothesized a priori
that feedback subjects would increase control over their
anxiety, and would have a greater increase in control than
the sham subjects, significance was assessed using one-
tailed tests at a Po0.05 level.
Control task data. Within-group changes in OFC (percentage
of signal change at the final time point compared with per-
centage of signal change at the first time point) were tested
with Wilcoxon’s signed-rank tests and between-group differ-
ences were tested with a Wilcoxon’s rank-sum test. Because
we hypothesized a priori that feedback subjects would
increase control over their OFC activity, and would have a
greater increase in control than the sham subjects, signifi-
cance was assessed using one-tailed tests at a Po0.05 level.
Results
Changes in resting state functional connectivity patterns before
and after NF were assessed using a network theory metric,
degree of connectivity, computed on a voxel-wise basis.
20
The
group composite map of whole-brain changes in degree of
connectivity in the feedback group showed significant (Po0.05
corrected) decreases in brain regions associated with emotion
processing, including the insula and adjacent regions, the
hippocampi, parahippocampal and entorhinal cortex, the right
amygdala, the brainstem in the vicinity of the substantia nigra,
the temporal pole, superior temporal sulcus, thalamus and
fusiform gyrus. By contrast, significantly (Po0.05 corrected)
increased degree of connectivity was seen in prefrontal areas
associated with emotion regulation and cognitive control,
25
including right lateral prefrontal cortex and bilateral portions of
Brodmann’s area 8 (Figure 2a). The composite map from the
sham group did not show any loci of significant changes in
connectivity surviving multiple comparisons correction (not
Figure 1 A screen-shot showing the visual display at the end of a
neurofeedback run, which ended with a rest block and a corresponding neutral
image. The images shown during increase and decrease blocks were designed to
induce contamination anxiety. The time course of the orbitofrontal cortex, shown at
the bottom, tended to be higher in the red relative to the blue periods, indicating
some control over their target region.
Neurofeedback modulates connectivity and anxiety
D Scheinost et al
3
Translational Psychiatry
shown). The contrast between the two groups at Po0.05
corrected was similar to the feedback group composite map
(Figure 2b), further suggesting that the changes in connectivity
were due specifically to the feedback and not to non-specific
aspects of the task.
Control over contamination anxiety was assessed by
showing subjects contamination-associated images before
and after NF and instructing them to control their anxiety and
report how much anxiety they experienced. NF subjects
showed a significant reduction in contamination anxiety
several days following the feedback sessions (P¼0.02,
median ¼0.27), based on these self-report measures, while
sham subjects did not (P¼0.45, median ¼0.04). There was a
significant difference between the groups (P¼0.034,
U¼25.5; Figure 3), confirming that the improvement in the
intervention group was due to the feedback rather than to
habituation or some other non-specific aspect of training.
To identify the specific connectivity changes most asso-
ciated with improved control over anxiety, we created a map
for the feedback group of the correlation between changes in
degree of connectivity (Figure 2a) and increased control over
anxiety (Figure 3). Two areas of significant correlation
(Po0.05 corrected) are apparent: the target region of the
OFC, bilaterally, and a right lateral parietal area (Figure 4).
Changes in control over OFC activity were assessed in the
absence of a NF signal at the start and end of each of the two
90-min feedback sessions. Comparison of the first and final
assessments revealed that the NF subjects changed activity
in the target area of their OFC (Po0.01, median ¼0.2689),
while the sham subjects did not (P¼0.43, median ¼
0.0642; Figure 5). The contrast between groups
approached statistical significance (P¼0.07, U¼30).
Discussion
Mitigation of maladaptive behavioral states entails inducing
change in the underlying brain circuitry. We demonstrate that
rt-fMRI NF can modulate intrinsic brain connectivity patterns
associated with anxiety control, thereby enhancing control
over contamination anxiety.
These results extend previous applications of rt-fMRI NF in
two critical ways. First, we show changes in both anxiety
regulation and brain connectivity that persist for several days
following the end of NF training. Previous rt-fMRI studies have
reported success in training subjects using specific target
brain regions
4,26
and in altering relevant connectivity patterns
during task performance
5,27
or at rest.
12
However, these
changes in brain network function have been reported only
during NF or immediately after it and may therefore represent
state changes, rather than lasting alterations in brain
functional architecture. Transient symptomatic improvements
or changes in brain network connectivity would clearly be of
only limited clinical utility and interest. We demonstrate that
rt-fMRI training can induce changes in both resting-state brain
patterns and behavior that last for days after the last NF
session: both the post-feedback resting-state data and the
post-intervention behavioral data were collected several days
after the completion of the NF intervention. Therefore, the
increased control over anxiety produced by NF (Figure 3)
reflects a persistent change, not an acute consequence of the
training. Similarly, the changes in connectivity we observe
Figure 2 (a) Changes in degree of connectivity in the feedback group. Increases are shown in red/yellow and decreases in blue/purple. Decreases in connectivity are seen
in limbic areas, and increases are seen in prefrontal regions. (b) Contrast between the feedback and sham groups. This contrast is similar to the feedback group composite
map, suggesting that the changes in connectivity in the feedback group were a result of the feedback rather than habituation or some other non-specificaspect of training. All
slices are shown with radiological convention (left is on the right) at a whole-brain-corrected Po0.05 threshold.
Figure 3 Change in control over contamination anxiety in the two groups. The
neurofeedback subjects significantly increased their control over anxiety (indicated
by an asterisk) and the sham subjects did not.
Neurofeedback modulates connectivity and anxiety
D Scheinost et al
4
Translational Psychiatry
(Figures 2a, b and 4) reflect persistent alterations produced by
the intervention. These lasting alterations suggest that rt-fMRI
is a promising mechanism to induce clinically meaningful
changes in the brain.
Second, we show that this approach can be used to
modulate a particular form of anxiety. Anxiety is a core
symptom of many psychiatric conditions and a substantial
source of distress in both clinical and non-clinical populations.
Contamination anxiety, in particular, is a cardinal symptom of
OCD and is common in the non-clinical population,
2
such as
the subjects described here. The ability of rt-fMRI NF to
modulate this core emotional state creates clear possibilities
for clinical applications.
The observed alterations in resting-state connectivity in the
feedback group (Figures 2a and b) show that rt-fMRI is
capable of reorganizing the functional brain architecture
associated with emotion processing. The pattern of change
in functional connectivity, with decreased connectivity in
limbic and paralimbic areas and increased connectivity in
lateral and anterior frontal areas associated with cognitive
control, is consistent with a model in which dorsolateral
prefrontal cortical areas have a role in emotional control by
downregulating the function of areas involved in emotion
generation.
25
The specificity of changes in functional con-
nectivity to the feedback group (and absence of these
changes in the sham subjects) indicates that it was not a
result simply of repeated exposure to contamination-related
images, or of practice effects, but was contingent upon the
receipt of accurate, subject-specific NF.
By correlating changes in resting-state connectivity with
changes in anxiety, it is possible to probe the neural basis of
anxiety control. We found a negative correlation between
OFC global connectivity and improved anxiety regulation,
suggesting a critical role for this region in controlling
contamination anxiety and confirming that we have targeted
an appropriate region of the brain with our intervention
(Figure 4). The correlation between increased connectivity
in the right lateral parietal region and improved control over
anxiety was not predicted, but it is interesting to note that this
area has been implicated in OCD.
28
Susceptibility artifacts make the OFC a difficult region to
image. However, using an optimized pulse sequence, we
were able to successfully train individuals to reorganize
their brain patterns so as to better control activity in this
region. This has implications beyond this study and beyond
the management of anxiety, as this critical limbic area has
been implicated in many different disorders involving dis-
rupted emotional processing, including conduct disorder,
bipolar disorder and addiction.
29
Despite a growing literature on rt-fMRI NF, there is, to date,
limited evidence that it can induce persistent changes in brain
function. Most NF studies have not included a follow-up
imaging session. At least two previous studies have reported
persistent alterations in behavioral measures days after
NF,
30,31
supporting the view that rt-fMRI NF can induce a
lasting reorganization of brain function. On the other hand,
one study that identified changes in brain function during the
NF found that those changes did not persist after the
feedback.
3
Two factors may have contributed to the more
persistent changes in resting-state connectivity in the current
study. First, by having two feedback sessions spaced several
days apart, subjects in this study had a chance to consolidate
their learning overnight; sleep is a critical part of some forms of
learning
32
and may have enhanced the persistence of plastic
changes in this case. Second, anxiety-related brain regions
may be more subject to lasting changes than other brain
circuits. More studies are needed examining the persistence
of changes induced by rt-fMRI NF in a variety of contexts to
clarify these issues.
It is possible that the neural basis of contamination anxiety
in patients with OCD differs from that of healthy subjects.
However, the region that we targeted (the OFC) has been
reported to have elevated activation during the experience of
contamination anxiety in both healthy individuals and OCD
patients.
13,14
Thus, it is plausible that a protocol such as this
one, which uses rt-fMRI NF to train subjects to control activity
in their OFC, will translate into improved control over
contamination anxiety in OCD patients. A study in the patient
population is needed to address this issue definitively.
As our NF protocol was limited to two sessions, we are not
able to identify the optimal number of sessions for this
intervention. Our results suggest that two sessions are better
than one: the subjects in our study did not show a significant
increase in control over their OFC after a single session, but
Figure 4 Regions of the brain where changes in degree of connectivity correlated with increased control over contamination anxiety, at a whole-brain-corrected Po0.05
threshold. Positive correlations between increased control over anxiety and increased degree of connectivity are shown in red/yellow and negative correlations are shown in
blue/purple. Increased control over anxiety was associated with decreased connectivity in the orbitofrontal cortex and increased connectivity in a right parietal region.
Neurofeedback modulates connectivity and anxiety
D Scheinost et al
5
Translational Psychiatry
they did improve with two NF sessions (Figure 5). It is possible
that a larger number of NF sessions would be able to produce
still greater control of OFC activity. Future work should include
studies aimed at optimizing the number of NF sessions.
Rt-fMRI NF was used to reorganize the intrinsic functional
brain patterns of subjects so as to allow them to gain greater
control over their contamination anxiety. The alterations
lasted for several days after the completion of NF training.
These results have implications for mechanism, as they imply
brain plasticity rather than just a state change induced in the
short term by the feedback experience. More generally, these
data indicate that rt-fMRI NF can produce persistent changes
in the brain circuitry underlying maladaptive behavioral states
and thereby support its potential a treatment for neuropsy-
chiatric disease.
Conflict of interest
The authors declare no conflict of interest.
Acknowledgements. This work was supported by funding from the
National Institutes of Health (MH090384, EB012969, MH081190) and the Doris
Duke Charitable Foundation. We thank Maolin Qiu, Jitendra Bhawnani and Hedy
Sarofin for technical assistance and the State of Connecticut for its support of the
Ribicoff Research Facilities at the Connecticut Mental Health Center.
1. Greenberg BD, Rauch SL, Haber SN. Invasive circuitry-based neurotherapeutics:
stereotactic ablation and deep brain stimulation for OCD. Neuropsychopharmacology
2010; 35: 317–336.
2. Ruscio AM, Stein DJ, Chiu WT, Kessler RC. The epidemiology of obsessive-compulsive
disorder in the National Comorbidity Survey Replication. Mol Psychiatry 2010; 15: 53–63.
3. Hamilton JP, Glover GH, Hsu JJ, Johnson RF, Gotlib IH. Modulation of subgenual anterior
cingulate cortex activity with real-time neurofeedback. Hum Brain Mapp 2011; 32: 22–31.
4. Yoo S-S, Jolesz FA. Functional MRI for neurofeedback: feasibility study on a hand motor
task. NeuroReport 2002; 13: 1377–1381.
5. Zotev V, Krueger F, Phillips R, Alvarez RP, Simmons WK, Bellgowan P et al. Self-
regulation of amygdala activation using real-time fMRI neurofeedback. PLoS One 2011; 6:
e24522.
6. Caria A, Sitaram R, Veit R, Begliomini C, Birbaumer N. Volitional control of anterior insula
activity modulates the response to aversive stimuli. A real-time functional magnetic
resonance imaging study. Biol Psychiatry 2010; 68: 425–432.
7. Li X, Hartwell KJ, Borckardt J, Prisciandaro JJ, Saladin ME, Morgan PS et al. Volitional
reduction of anterior cingulate cortex activity produces decreased cue craving in smoking
cessation: a preliminary real-time fMRI study. Addict Biol 2012 (e-pub ahead of print).
8. Rota G, Sitaram R, Veit R, Erb M, Weiskopf N, Dogil G et al. Self-regulation of regional
cortical activity using real-time fMRI: the right inferior frontal gyrus and linguistic
processing. Hum Brain Mapp 2009; 30: 1605–1614.
9. Ruiz S, Lee S, Soekadar SR, Caria A, Veit R, Kircher T et al. Acquired self-control of insula
cortex modulates emotion recognition and brain network connectivity in schizophrenia.
Human Brain Mapp 2011; 34: 200–212.
10. deCharms RC, Maeda F, Glover GH, Ludlow D, Pauly JM, Soneji D et al. Control over brain
activation and pain learned by using real-time functional MRI. Proc Natl Acad Sci 2005;
102: 18626–18631.
11. Subramanian L, Hindle JV, Johnston S, Roberts MV, Husain M, Goebel R et al. Real-time
functional magnetic resonance imaging neurofeedback for treatment of Parkinson’s
disease. J Neurosci 2011; 31: 16309–16317.
12. Hampson M, Scheinost D, Qiu M, Bhawnani J, Lacadie C, Leckman JF et al. Biofeedback
from the supplementary motor area reduces functional connectivity to subcortical regions.
Brain Connectivity 2011; 1: 91–98.
13. Mataix-Cols D, Cullen S, Lange K, Zelaya F, Andr ew C, Amaro E et al. Neural correlates of
anxiety associated with obsessive-compulsive symptom dimensions in normal volunteers.
Biol Psychiatry 2003; 53: 482–493.
14. Mataix-Cols D, Wooderson S, Lawren ce N, Brammer MJ, Speckens A, Phillips ML. Distinct
neural correlates of washing, checking, and hoarding symptom dimensions in obsessive-
compulsive disorder. Arch Gen Psychiatry 2004; 61: 564–576.
15. Burns GL, Keortge SG, Formea GM, Sternberger LG. Revision of the Padua Inventory of
obsessive compulsive disorder symptoms: distinctions between worry, obsessions, and
compulsions. Behav Res Ther 1996; 34: 163–173.
16. Hampson M, Stoica T, Saksa J, Scheinost D, Qiu M, Bhawnani J et al. Real-time fMRI
biofeedback targeting the orbitofrontal cortex for contamination anxiety. J Vis Exp 2012 pii
3535.
17. Scheinost D, Hampson M, Qiu M, Bhawnani J, Constable RT, Papademetris XA. Graphics
processing unit accelerated motion correction algorithm and modular system for real-time
fMRI. Neuroinformatics 2013 (e-pub ahead of print).
18. Joshi A, Scheinost D, Okuda H, Belhachemi D, Murphy I, Staib LH et al. Unified framework
for development, deployment and robust testing of neuroimaging algorithms. Neuroinfor-
matics 2011; 9: 69–84.
19. Martuzzi R, Ramani R, Qiu M, Rajeevan N, Constable RT. Functional connectivity and
alterations in baseline brain state in humans. Neuroimage 2010; 49: 823–834.
20. Rubinov M, Sporns O. Complex network measures of brain connectivity: uses and
interpretations. Neuroimage 2010; 52: 1059–1069.
21. Buckner RL, Vincent JL. Unrest at rest: default activity and spontaneous network
correlations. NeuroImage 2007; 37: 1091–1096.
22. Hampson M, Tokoglu F, Shen X, Scheinost D, Papademetris X, Constable RT. Intrinsic
brain connectivity related to age in young and middle aged adults. PloS One 2012; 7:
e44067.
23. Cole MW, Yarkoni T, Repovs G, Anticevic A, Braver TS. Global connectivity of
prefrontal cortex predicts cognitive control and intelligence. J Neurosci 2012; 32:
8988–8999.
24. Van Dijk KR, Sabuncu MR, Buckner RL. The influence of head motion on intrinsic
functional connectivity MRI. Neuroimage 2012; 59: 431–438.
25. Ochsner KN, Gross JJ. Cognitive emotion regulation: Insights from social cognitive and
affective neuroscience. Curr Dir Psychol Sci 2008; 17: 153–158.
26. deCharms RC, Christoff K, Glover GH, Pauly JM, Whitfield S, Gabrieli JDE. Learned
regulation of spatially localized brain activation using real-time fMRI. NeuroImage 2004; 21:
436–443.
27. Rota G, Handjaras G, Sitaram R, Birbaumer N, Dogil G. Reorganization of functional and
effective connectivity during real-time fMRI-BCI modulation of prosody processing. Brain
Lang 2011; 117: 123–132.
28. Menzies L, Chamberlain SR, Laird AR, Thelan SM, Sahakian BJ, Bullmore ET. Integrating
evidence from neuroimaging and neuropsychological studies of obsessive-compulsive
disorder: the orbitofrontal-striatl model revisited. Neurosci Biobehav Rev 2008; 32:
525–549.
29. Jackowski AP, Filho GM, Almeida AG, Araujo CM, Reis M, Nery F et al. The involvement of
the orbitofrontal cortex in psychiatric disorders: an update of neuroimaging findings. Rev
Bras Psiquiatr 2012; 34: 207–212.
30. Shibata K, Watanabe T, Sasaki Y, Kawato M. Perceptual learning incepted by decoded
fMRI neurofeedback without stimulus presentation. Science 2011; 334: 1413–1415.
31. Yoo SS, O’Leary HM, Fairneny T, Chen NK, Panych LP, Park H et al. Increasing cortical
activity in auditory areas through neurofeedback functional magnetic resonance imaging.
Neuroreport 2006; 17: 1273–1278.
32. Gais S, Born J. Low acetylcholine during slow-wave sleep is critical for declarative memory
consolidation. Proc Natl Acad Sci USA 2004; 101: 2140–2144.
Translational Psychiatry is an open-access journal
published by Nature Publishing Group.Thisworkis
licensed under a Creative Commons Attribution-NonCommercial-
ShareAlike 3.0 Unported License. To view a copy of this license, visit
http://creativecommons.org/licenses/by-nc-sa/3.0/
Figure 5 Control over the orbitofrontal cortex target area at the four different
time points when control over the brain area was assessed. The median and quartile
ranges are shown for each of the two groups at each time point. BF, biofeedback;
SF, sham feedback.
Neurofeedback modulates connectivity and anxiety
D Scheinost et al
6
Translational Psychiatry
