tDCS indicates transcranial direct current stimulation.
A, Thresholded statistical map depicts clusters in the right and left amygdala (z > 2.3, small volume correction) showing reduced response to fearful face distractors vs neutral face distractors under conditions of low attentional load for active compared with sham tDCS (within subject). Amygdala regions-of-interest were defined with use of the Harvard-Oxford Subcortical Structural Atlas (thresholded at 50% probability). B, Right amygdala percent signal change extracted and plotted from the significant interaction of fearful face distractors (fear) vs neutral face distractors (neutral) under conditions of low vs high attentional load in the sham (ie, baseline) condition. C, Right and left amygdala percent signal change data extracted and plotted for the significant cluster shown in A for sham vs active tDCS and fear vs neutral under conditions of low attentional load. The functional maps are overlaid on an average anatomical image spatially normalized to Montreal Neurological Institute space. Error bars indicate 2 SEMs. L indicates left and R, right.
aP < .05.
A, Thresholded statistical maps (z > 2.3, whole-brain correction) depict significant clusters in cortical regions associated with attentional processing (left superior frontal sulcus, precentral gyrus, frontal eye fields and right supramarginal gyrus, and superior temporal sulcus) for active tDCS compared with sham for the contrast of fearful face distractors (fear) vs neutral face distractors (neutral) under conditions of low attentional load. B, Percent signal change extracted and plotted for the significant clusters shown in A for active vs sham tDCS for the contrast of fear vs neutral under conditions of low attentional load. The functional maps are overlaid on an average anatomical image spatially normalized to Montreal Neurological Institute space. Error bars indicate 2 SEMs. FEF indicates frontal eye fields; L, left; PMd, dorsal premotor cortex; R, right; SMG, supramarginal gyrus; and TPJ, temporoparietal junction.
aP < .007.
bP < .001.
eAppendix 1. Methods and Materials
eAppendix 2. Results
eFigure. tDCS Montage and Corresponding Simulated Electric Field Distribution
Customize your JAMA Network experience by selecting one or more topics from the list below.
Ironside M, Browning M, Ansari TL, et al. Effect of Prefrontal Cortex Stimulation on Regulation of Amygdala Response to Threat in Individuals With Trait Anxiety: A Randomized Clinical Trial. JAMA Psychiatry. 2019;76(1):71–78. doi:10.1001/jamapsychiatry.2018.2172
Does stimulation of prefrontal cortex reduce amygdala threat reactivity in individuals with human trait anxiety?
In this randomized clinical trial of 16 women, placebo-controlled stimulation of the prefrontal cortex vs sham procedure reduced amygdala fear signaling and increased frontoparietal attentional control, reducing attentional capture by threat.
Single-session stimulation can improve prefrontal control of limbic threat reactivity, indicating a candidate mechanism underlying treatment effects of noninvasive brain stimulation in affective disorders.
Transcranial direct current stimulation (tDCS) of the dorsolateral prefrontal cortex (DLPFC) is under clinical investigation as a treatment for major depressive disorder. However, the mechanisms of action are unclear, and there is a lack of neuroimaging evidence, particularly among individuals with affective dysfunction. Furthermore, there is no direct causal evidence among humans that the prefrontal-amygdala circuit functions as described in animal models (ie, that increasing activity in prefrontal cortical control regions inhibits amygdala response to threat).
To determine whether stimulation of the prefrontal cortex reduces amygdala threat reactivity in individuals with trait anxiety.
Design, Setting, and Participants
This community-based randomized clinical trial used a double-blind, within-participants design (2 imaging sessions per participant). Eighteen women with high trait anxiety (age range, 18-42 years) who scored greater than 45 on the trait measure of State-Trait Anxiety Inventory were randomized to receive active or sham tDCS of the DLPFC during the first session and the other intervention during the next session. Each intervention was followed immediately by a functional imaging scan during which participants performed an attentional task requiring them to ignore threatening face distractors. Data were collected from May 7 to October 6, 2015.
Main Outcomes and Measures
Amygdala threat response, measured with functional magnetic resonance imaging.
Data from 16 female participants (mean age, 23 years; range, 18-42 years), with 8 in each group, were analyzed. Compared with sham stimulation, active DLPFC stimulation significantly reduced bilateral amygdala threat reactivity (z = 3.30, P = .04) and simultaneously increased activity in cortical regions associated with attentional control (z = 3.28, P < .001). In confirmatory behavioral analyses, there was a mean improvement in task accuracy of 12.2% (95% CI, 0.30%-24.0%; mean [SD] difference in number of correct answers, 2.2 [4.5]; t15 = 1.94, P = .04) after active DLPFC stimulation.
Conclusions and Relevance
These results reveal a causal role for prefrontal regulation of amygdala function in attentional capture by threat in individuals with high trait anxiety. The finding that prefrontal stimulation acutely increases attentional control signals and reduces amygdala threat reactivity may indicate a neurocognitive mechanism that could contribute to tDCS treatment effects in affective disorders.
isrctn.org Identifier: ISRCTN78638425
The difficulty of treating highly comorbid mood and anxiety disorders1 has led to increased clinical interest in potential alternative treatments, such as noninvasive brain stimulation techniques, including transcranial direct current stimulation (tDCS) of the dorsolateral prefrontal cortex (DLPFC). A recent meta-analysis2 of individual data from 289 patients with major depressive disorder treated with tDCS showed that, compared with placebo-controlled clinical trials of antidepressant drugs, sham-controlled clinical trials of DLPFC tDCS had a similar number needed to treat (7 for response, 9 for remission).
However, as with many antidepressant treatments, the mechanism of action is unclear. The prefrontal cortex is implicated in animal studies, which provide compelling evidence of its importance in regulating responses to threat via direct inhibition of the amygdala complex. In rodent models, the prefrontal cortex has been shown to inhibit aversive associations established in fear conditioning, with prefrontal lesions impeding3 and prefrontal electrical stimulation enhancing4 the extinction of a conditioned response. Furthermore, electrical prestimulation of the prefrontal cortex in rats and cats5 specifically reduced amygdala responses. These preclinical models provide the foundation for theoretical models of emotional dysfunction in human disorders of anxiety and depression, in which deficient prefrontal control is believed to result in overactivity within areas responsible for assigning salience and attention to threatening stimuli, such as the amygdala.
The amygdala is a critical component of the neural circuitry underlying fear processing.6 Consistent with this, human neuroimaging studies have confirmed hyperactive amygdala and/or hypoactive prefrontal activity in patients with anxiety disorders7 and major depressive disorder.8,9 Furthermore, there is evidence that treatment with antidepressant drugs10 or cognitive behavioral therapy11 can reduce amygdala hyperactivity. However, there is no direct causal evidence that the prefrontal-amygdala circuit functions in humans as reported in animal models, that is, that increasing activity in prefrontal cortical control regions inhibits amygdala responses to threat. We combined DLPFC tDCS with functional magnetic resonance imaging (fMRI) of a specific cognitive task that recruits the prefrontal-amygdala circuit to perform a causal test of this hypothesis in a sample population of individuals with high trait anxiety.
Transcranial direct current stimulation can be used to tonically increase or decrease cortical excitability using weak electrical currents.12 Induced changes in tissue excitability can persist over minutes to hours after stimulation, effects that are N-methyl-d-aspartate receptor dependent and presumed to reflect changes in synaptic efficacy and plasticity.13,14 Initially used as a tool to induce changes in motor evoked potentials, tDCS has more recently been used to modulate cognition, such as attentional control15 and working memory.16 Neuroimaging indicates that DLPFC tDCS alters functional activation and connectivity in brain regions that support cognitive function, including regions distal from the stimulating electrodes.17-19 Thus, the mechanism of action of DLPFC tDCS in the treatment of major depressive disorder may arise through the induction of plasticity in distributed cortical-striatal and limbic circuits, a network hypothesis that can only be assessed through combined neurostimulation and neuroimaging research. By use of stimulation to change the electrical state of cortical tissue, it becomes possible to test hypotheses about functional interactions between the cortex and connected subcortical structures20; for example, it is possible to test the causal influence of the prefrontal cortex on regulation of the amygdala response to threat.
It has been shown through a behavioral study that bilateral DLPFC tDCS is associated with a reduction in vigilance to threat in an attentional task validated to predict the clinical response to anxiolytic drug treatment.21 This reduction in attentional bias to threat has been replicated among healthy volunteers22 and among individuals with social anxiety disorder.23 We hypothesize that prefrontal stimulation increases cortical activity, which increases top-down attentional control of connected limbic structures, thus increasing regulation of the amygdala threat response. Davidson’s model24 proposes decreased activation in the left prefrontal cortex among individuals with major depressive disorder and anxiety-specific increased activation of the right prefrontal cortex. This hypothesis is supported with findings that right DLPFC stimulation with repetitive transcranial magnetic stimulation (rTMS) is associated with an increase in right amygdala activation25 and that right DLPFC tDCS26 and rTMS27 are associated with an increase in attentional allocation to threatening stimuli. Excitation of the left DLPFC with tDCS28,29 and rTMS30 decreases attentional bias to threat, whereas inhibition of the left DLPFC with single-pulse TMS increases attentional bias to threat in individuals with anxiety.31 To test this mechanistic hypothesis, we assessed the effect of simultaneous left anodal and right cathodal DLPFC tDCS on neural threat reactivity measured with fMRI among a group of individuals with high trait anxiety during a well-validated, attentional control paradigm that was sensitive to anxiety-related differences in attentional function. Earlier work with this task across a range of trait anxiety has shown that individuals with high anxiety exhibit hypoactive prefrontal response and hyperactive amygdala response to fearful face distractors.32 This was apparent only under conditions of low attentional load. When the task was undemanding and did not fully occupy attentional resources, individuals with high anxiety exhibited impoverished prefrontal recruitment and increased amygdala responsivity to threat-related distractors. However, the causal relationship between this concurrent prefrontal hyporesponsivity and amygdala hyperreactivity has not been determined to date. We sought to address this with use of tDCS to stimulate the DLPFC and fMRI to measure changes in the neural response to threat among a group of individuals with high trait anxiety. We used a single dose of the multisession tDCS protocol used in clinical trials of major depressive disorder. We predicted that DLPFC tDCS would modulate this pattern of activation and behavior. Specifically, we predicted that, under conditions of low attentional load with fearful distractors, receipt of tDCS would have 3 directional effects: an increase in cortical activation associated with attentional control, a decrease in amygdala activation, and an improvement in task accuracy. We stimulated bilateral DLPFC and then assessed changes in brain activity and behavior.
The study was approved by the University of Oxford Central University Research Ethics Committee. Eighteen female participants (all right handed; age range, 18-42 years; mean [SD] age, 23.1 [3.7] years) were recruited from the community. Female participants were chosen to avoid sex-related differences in brain activation during emotional tasks33 and because there is a higher prevalence of anxiety disorders among women.34 Participants were prescreened with an online version of the Spielberger State-Trait Anxiety Inventory (STAI).35 Participants with high trait anxiety, defined as those who scored greater than 45 on the trait questionnaire (STAI-T), were invited for screening at the Warneford Hospital (Oxford, England), where they completed the Structured Clinical Interview for DSM-IV disorders.36 Written informed consent was obtained from all participants, and data were deidentified. Participants who successfully met full screening requirements were invited to take part in 2 tDCS and fMRI scanning sessions at the John Radcliffe Hospital (Oxford, England). Additional details are given in the Trial Protocol in Supplement 1 and in Supplement 2.
This trial used a within-participants double-blind design with 18 participants, each of whom attended 2 separate tDCS and fMRI sessions, randomized to stimulation order (real or sham tDCS followed by sham or real tDCS 1 month later, counterbalanced) (Figure 1). On the day of the study, participants received tDCS while they sat at rest. After the stimulation ended, participants entered the scanner (mean time from tDCS offset to task onset, approximately 7 minutes) and performed the attentional control task.
The attentional load paradigm was adapted from Bishop et al.32 On each trial, a 6-letter string superimposed on a task-irrelevant unfamiliar face was presented for 200 milliseconds. The face stimuli comprised 4 different individuals with fearful or neutral expressions taken from the Pictures of Facial Affect.37
The task was to decide whether the letter string contained an X or an N. In one-half of the blocks (the high attentional load condition), the string comprised a single target letter (N or X) and 5 nontarget letters (H, K, M, W, and Z) arranged in random order. In the other half of the blocks, (the low attentional load condition), the letter string comprised 6 Xs or 6 Ns, which removed attentional search requirements. This manipulation of attentional load was identical to that used in earlier work32,38 and conforms to Lavie’s39 description of cognitive effort.
The key hypothesis-driven condition of interest was low attentional load with fearful distractors. Earlier work showed that amygdala response to threat was observed only under the low attentional load condition in this task.32 Therefore, by examining the effect of tDCS on brain regions selectively activated by this key hypothesis-driven contrast (fearful vs neutral face distractors under conditions of low attentional load), it was possible to test the hypothesis that tDCS reduces vigilance to threat in individuals with trait anxiety by altering frontolimbic activity and, specifically, by reducing amygdala response to fearful distractors.
Stimulation was delivered using a battery-powered device (DC Stimulator Plus; Neuroconn40). The full parameters are given in Supplement 1 and Supplement 2.
Blood oxygenation level–dependent contrast functional images were acquired with echo-planar T2*-weighted imaging using a Siemens 3T Magnetom TrioTim syngo with a head coil gradient set. The full parameters are given in Supplement 2. Data were lost for 2 participants because of server issues, which reduced the analyzed sample to 16 participants (32 scans).
The fMRI data processing was carried out using FEAT (FMRI Expert Analysis Tool), version 6.00, part of FSL (Functional MRI of the Brain Software Library). Registration to high-resolution structural and standard space was carried out using FLIRT (FMRIB's Linear Image Registration Tool).41,42 Registration from high-resolution structural to standard space was then further refined using FNIRT (FMRIB's Nonlinear Image Registration Tool),43 and motion correction was carried out with MCFLIRT, applying rigid-body transformations. Regressors for each condition yoked to trial onset were convolved with the canonical hemodynamic response function. An intermediate analysis was first performed, combining 3 runs into a single data set for each participant for each testing session (2 per participant). A within-subjects analysis was performed, and z (gaussian T) statistic images were thresholded using clusters determined by z > 2.3 and a (corrected) cluster significance threshold of P = .05. One-sample t tests were used to determine baseline effects (eg, fear vs neutral) in the sham condition only (1-tailed), and paired t tests were used to determine differences between the active and sham conditions (1-tailed). Small volume–corrected analyses were performed in bilateral amygdala regions of interest. The amygdala regions of interest were defined using the Harvard-Oxford Subcortical Structural Atlas (using a standard threshold including all voxels with >50% probability of lying within the amygdala). The hypothesis-driven contrasts analyzed are specified in Table 1. Behavioral analysis was carried out using SPSS, version 24.0 (IBM Corp), with repeated measures analysis of variance (2-tailed) and follow up t tests (1-tailed).
Data from 16 female participants (age range, 18-42 years; mean [SD] age, 23.1 [3.7] years) were analyzed. Because an earlier investigation that used this task across a range of trait anxiety levels32 found that the right amygdala responds selectively during trials with fearful (vs neutral) distractors when attentional load is low (vs high), we first tested whether this baseline effect was replicated in the sham condition. As predicted, right amygdala activation occurred selectively during trials with fearful (not neutral) distractors only when attentional load was low (not high) (loadlow-high × emotionfear-neutral, z = 3.06, P = .04, small volume correction) (Figure 2B). This finding confirmed that our emotional task was sufficiently sensitive to detect the expected presence of amygdala threat signaling in this sample of individuals with high trait anxiety.
This baseline amygdala response was altered by tDCS. We tested for the predicted effect of tDCS, which was a reduction of amygdala signal for the contrast of fearful-neutral distractors under the low attentional load condition. Region-of-interest analysis revealed that, under the low attentional load condition only, bilateral DLPFC stimulation significantly reduced right amygdala threat response (tDCSreal-sham × emotionfear-neutral, z = 3.30, P = .04, small volume correction) (Figure 2A and C). A similar reduction was observed in the left amygdala (z = 2.82, P = .04, small volume correction) (Figure 2A and C). Consistent with the absence of a baseline effect of fear in high attentional load trials, tDCS did not modify amygdala activation in response to fearful vs neutral faces under the high attentional load condition (tDCSreal-sham × emotionfear-neutral,z < 2.30, P > .05). Instead, tDCS reduced amygdala response to fearful vs neutral distractors selectively under the low attentional load condition.
Whole-brain analysis revealed that, contrary to the amygdala effect, stimulation significantly increased activation in frontal, temporal, and parietal clusters. The left frontal cluster extended from the frontal eye fields to dorsal premotor cortex (collectively, area 8 in the Sallet atlas; Oxford Centre for fMRI of the Brain Software Library), areas associated with attention and action selection44 (tDCSreal-sham × emotionfear-neutral, z = 3.74, P < .001, whole-brain corrected) (Figure 3A and B). In addition, clusters emerged in the right superior parietal lobule; the angular gyrus and supramarginal gyrus (key nodes of the dorsal attention network) (tDCSreal-sham × emotionfear-neutral, z = 3.47, P < .001, whole-brain corrected) (Figure 3A and B); and the temporoparietal junction (a key node of the ventral attention network45) (tDCSreal-sham × emotionfear-neutral, z = 3.28, P < .001, whole-brain corrected) (Figure 3A and B). Table 2 gives a summary of the results. Consistent with the absence of a baseline effect of fear and no effect of tDCS on high attentional load trials, tDCS did not modify whole-brain activation to fearful vs neutral faces in the high attentional load condition (tDCSreal-sham × emotionfear-neutral, z < 2.30, P > .05). Instead, tDCS increased cortical response to fearful vs neutral distractors selectively under the low attentional load condition.
This task was optimized to drive differential activity in the prefrontal cortex and amygdala rather than to produce behavioral differences between conditions. However, earlier studies have reported marginal effects of anxiety on task behavior. Consistent with these findings, we compared task accuracy after real and sham tDCS. The tDCS × load × emotion interaction (F1, 15 = 3.25, P = .09) showed a trend and thus we performed confirmatory analyses that suggest that, in the key condition in which there was a hypothesized effect of tDCS (low attentional load, fearful distraction trials), the contrast of real vs sham tDCS indicated a mean improvement in accuracy of 12.2% (95% CI, 0.30%-24.0%) after stimulation (mean [SD] difference in number of correct answers, 2.2 [4.5]; t15 = 1.936, 1-tailed P = .04, Cohen d = 0.49). Full behavioral results are given in Supplement 2.
In this sample population of women with high trait anxiety, DLPFC stimulation was associated with frontolimbic governance of attentional control under threat. From a functional presective, stimulation simultaneously reduced amygdala and increased cortical activation in response to fearful face distractors. From a behavioral perspective, this functional response was accompanied by reduced influence of threat distractors on task accuracy. The demonstration that DLPFC stimulation can reduce amygdala threat reactivity and simultaneously increase activity in key nodes of the dorsal and ventral attentional control networks (which are associated with the detection of behaviorally relevant stimuli,45 attentional control,46 and action selection44) provides the first experimental evidence, to our knowledge, for a direct causal inhibitory role of prefrontal cortex on amygdala threat response in individuals with trait anxiety. These findings suggest a mechanism of action that may contribute to the treatment effects of tDCS observed in clinical trials of affective disorders.
The amygdala is one of the key brain regions implicated in the pathophysiology of depression and anxiety disorders. Seminal preclinical research showed that conditioned fear is mediated by projections to the amygdala47 and that inactivation of the amygdala prevents the acquisition of fear conditioning.48 Patients with depression exhibit hyperactive amygdala responses to emotional information,9 as do patients with anxiety disorders.49,50 Treatment with selective serotonin reuptake inhibitors has attenuated amygdala response to fearful faces in healthy control individuals51 and patients with depression52 both before53 and after54 reported treatment response. This indicates that a reduction in amygdala hyperactivity may be an important part of the mechanism of action of selective serotonin reuptake inhibitors.
Animal studies suggest that the amygdala response to threat is reduced by top-down inhibition from the prefrontal cortex.4 Electric field modeling (Supplement 2) has shown that the montage used in this study may evoke the strongest electric field in the medial prefrontal cortex. This parallels preclinical research3-5 that focused on the role of this region in downregulating amygdala threat response. Individuals with depression9 and anxiety7 have decreased prefrontal activation55 in response to cognitive tasks,8 particularly in the context of fearful distractors,56 which is thought to reflect deficient attentional control. However, fMRI is a correlational technique. The addition of a causal intervention, such as tDCS, enabled us to change brain activity and determine the direction of the association between prefrontal and amygdala activity in humans. This study revealed that stimulating the DLPFC can significantly reduce the amygdala response to threat-related distractors.
A growing literature,23,28 including an earlier behavioral investigation of a paradigm reported to predict the clinical response to anxiolytic treatment,21 indicates that DLPFC tDCS has the potential to reduce vigilance to threat. We investigated a neural mechanism that may mediate this effect. Earlier work32 found that participants with high anxiety had increased amygdala and decreased prefrontal activation in response to fearful distractor faces under conditions of low attentional load compared with participants with low anxiety. The sample with high anxiety in our study showed a similar profile of amygdala response to fearful faces under the sham stimulation condition. As hypothesized, we found that DLPFC tDCS reduced this activation, such that, after stimulation, the findings among the group with high anxiety in the present study resembled the findings among the participants with low anxiety in the earlier study, with increased cortical and reduced amygdala response to threat distractors. Furthermore, confirmatory behavioral analyses indicated that this finding was accompanied by reduced attentional capture by threat distractors under low attentional load, which was reflected in increased accuracy after tDCS.
Repetitive transcranial magnetic stimulation of the DLPFC is a US Food and Drug Administration–approved treatment for depression,57 and acute protocols have indicated associations between DLPFC rTMS and threat processing.25 By contrast, the evidence base for clinical efficacy of DLPFC tDCS is still in development. If efficacy is established, tDCS offers several potential advantages over rTMS, including being better tolerated and being cheaper and simpler to administer, and the development of home-use devices broadens potential patient uptake and clinical research. This proof-of-concept study used a single dose of the multisession tDCS protocol used in clinical trials of major depressive disorder and describes a neurocognitive mechanism of action of DLPFC tDCS in patients with trait anxiety that should be investigated in future therapeutic trial designs of clinical efficacy as a function of reducing hyperactive amygdala-dependent threat vigilance.
The purpose of this study was to investigate acute effects of single-session DLPFC tDCS by testing for an induced change in amygdala response to threat. Future work is required to determine whether these effects are extended over time when repeated tDCS interventions are used and to determine whether these acute changes in neurobehavioral markers of threat vigilance are predictive of clinical treatment response. Because of the increased prevalence of high trait anxiety and anxiety disorders among women34 and to increase sample homogeneity, only female participants were recruited for this study. Future work should investigate whether these findings are generalizable to the male population. Finally, because we recruited only participants with high trait anxiety, we did not have a range of trait anxiety to correlate with amygdala response, as earlier studies have shown.
A number of recent studies have indicated that 2 weeks of daily prefrontal DLPFC tDCS may be an effective treatment for depression.58,59 Our results from a single session among a sample of participants with trait anxiety revealed an effect of DLPFC tDCS on a neural biomarker relevant to comorbid clinical depression and anxiety. Taken together with earlier behavioral findings obtained with the same stimulation protocol,21 these findings indicate a potential neurocognitive mechanism that may partially mediate the reported clinical efficacy of DLPFC tDCS. This biomarker may have potential for testing and benchmarking novel stimulation protocols in the development phase to accelerate treatment development for major depressive disorder and anxiety disorders. Confirming findings from preclinical animal studies and correlational evidence from human neuroimaging, this study offers, to our knowledge, the first causal evidence among humans of an association among the prefrontal cortex, attentional control networks, and the amygdala in regulating threat processing in individuals with trait anxiety.
Accepted for Publication: June 25, 2018.
Corresponding Author: Maria Ironside, DPhil, Center for Depression, Anxiety and Stress Research, McLean Hospital, Belmont, MA 02478 (email@example.com).
Published Online: October 17, 2018. doi:10.1001/jamapsychiatry.2018.2172
Author Contributions: Drs Harmer and O’Shea served as co–senior authors and contributed equally to the manuscript. Dr Ironside had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Ironside, Sekyi-Djan, Harmer, O'Shea.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Ironside, Browning, Ansari, Sekyi-Djan, O'Shea.
Critical revision of the manuscript for important intellectual content: Browning, Harvey, Bishop, Harmer, O'Shea.
Statistical analysis: Ironside, Browning, Sekyi-Djan, O'Shea.
Obtained funding: Ironside, Harmer.
Administrative, technical, or material support: Ansari, Harvey, Sekyi-Djan, Bishop.
Supervision: Harmer, O'Shea.
Guidance on task design and analysis: Bishop.
Conflict of Interest Disclosures: Dr Browning reports having received travel expenses from Lundbeck, owning shares in P1vital Products Ltd, and having worked as a consultant for Johnson & Johnson. Dr Harmer reports having received consultancy income from P1vital, Lundbeck, Johnson & Johnson, and Servier. No other disclosures are reported.
Funding/Support: This work was funded by a Medical Research Council studentship (Dr Ironside), Medical Research Council Clinician Scientist Fellowship MR/N008103/1 (Dr Browning), European Research Council Starting grant 260932 and National Institutes of Health grant R01MH091848 (Dr Bishop), and the National Institute of Health Research Oxford Biomedical Research Centre (Drs O’Shea and Harmer).
Role of the Funder/Sponsor: The funders 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.
Disclaimer: The views expressed are those of the authors and not necessarily those of the National Health Service, the National Institute for Health Research, or the US Department of Health. All information and materials in the manuscript are original.
Meeting Presentation: Meeting of the Society of Biological Psychiatry; May 13, 2016; Atlanta, Georgia.