Hostname: page-component-77f85d65b8-hzqq2 Total loading time: 0 Render date: 2026-04-19T19:32:31.157Z Has data issue: false hasContentIssue false
  • English
  • Français

Neural signatures of emotional biases predict clinical outcomes in difficult-to-treat depression

Published online by Cambridge University Press:  01 October 2024

A response to the following question: Will new brain circuit focused methods (EEG, fMRI etc) lead to more personalized care options?

Diede Fennema Open the ORCID record for Diede Fennema [Opens in a new window] ,
Gareth J. Barker Open the ORCID record for Gareth J. Barker [Opens in a new window] ,
Owen O’Daly ,
Beata R. Godlewska Open the ORCID record for Beata R. Godlewska [Opens in a new window] ,
Ewan Carr Open the ORCID record for Ewan Carr [Opens in a new window] ,
Kimberley Goldsmith Open the ORCID record for Kimberley Goldsmith [Opens in a new window] ,
Allan H. Young Open the ORCID record for Allan H. Young [Opens in a new window] ,
Jorge Moll Open the ORCID record for Jorge Moll [Opens in a new window]  and
Roland Zahn Open the ORCID record for Roland Zahn [Opens in a new window]
Diede Fennema
Affiliation:
Centre of Affective Disorders, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK
Gareth J. Barker
Affiliation:
Department of Neuroimaging, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK
Owen O’Daly
Affiliation:
Department of Neuroimaging, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK
Beata R. Godlewska
Affiliation:
Psychopharmacology Research Unit, University Department of Psychiatry, University of Oxford, Oxford, UK Oxford Health NHS Foundation Trust, Warneford Hospital, Oxford, UK
Ewan Carr
Affiliation:
Department of Biostatistics and Health Informatics, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK
Kimberley Goldsmith
Affiliation:
Department of Biostatistics and Health Informatics, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK
Allan H. Young
Affiliation:
Centre of Affective Disorders, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK National Service for Affective Disorders, South London and Maudsley NHS Foundation Trust, London, UK
Jorge Moll
Affiliation:
Cognitive and Behavioral Neuroscience Unit, D’Or Institute for Research and Education (IDOR), Pioneer Science Program, Rio de Janeiro, Brazil
Roland Zahn*
Affiliation:
Centre of Affective Disorders, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK National Service for Affective Disorders, South London and Maudsley NHS Foundation Trust, London, UK Cognitive and Behavioral Neuroscience Unit, D’Or Institute for Research and Education (IDOR), Pioneer Science Program, Rio de Janeiro, Brazil
*
Corresponding author: Roland Zahn; Email: roland.zahn@kcl.ac.uk
Rights & Permissions [Opens in a new window]

Abstract

Background:

Neural predictors underlying variability in depression outcomes are poorly understood. Functional MRI measures of subgenual cortex connectivity, self-blaming and negative perceptual biases have shown prognostic potential in treatment-naïve, medication-free and fully remitting forms of major depressive disorder (MDD). However, their role in more chronic, difficult-to-treat forms of MDD is unknown.

Methods:

Forty-five participants (n = 38 meeting minimum data quality thresholds) fulfilled criteria for difficult-to-treat MDD. Clinical outcome was determined by computing percentage change at follow-up from baseline (four months) on the self-reported Quick Inventory of Depressive Symptomatology (16-item). Baseline measures included self-blame-selective connectivity of the right superior anterior temporal lobe with an a priori Brodmann Area 25 region-of-interest, blood-oxygen-level-dependent a priori bilateral amygdala activation for subliminal sad vs happy faces, and resting-state connectivity of the subgenual cortex with an a priori defined ventrolateral prefrontal cortex/insula region-of-interest.

Findings:

A linear regression model showed that baseline severity of depressive symptoms explained 3% of the variance in outcomes at follow-up (F[3,34] = .33, p = .81). In contrast, our three pre-registered neural measures combined, explained 32% of the variance in clinical outcomes (F[4,33] = 3.86, p = .01).

Conclusion:

These findings corroborate the pathophysiological relevance of neural signatures of emotional biases and their potential as predictors of outcomes in difficult-to-treat depression.

Keywords

depressionpredictionfMRIprognosisdifficult-to-treat

Information

Type
Results
Information
Research Directions: Depression , Volume 1 , 2024 , e21
Check for updates
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s) 2024. Published by Cambridge University Press

Introduction

Currently, treatment of major depressive disorder (MDD) is based on a trial-and-error approach, with only half of patients responding to their initial treatment (Rush et al. Reference Rush, Trivedi, Wisniewski, Nierenberg, Stewart, Warden, Niederehe, Thase, Lavori, Lebowitz, McGrath, Rosenbaum, Sackeim, Kupfer, Luther and Fava2006). There is a clear need for improving treatment in patients with depression, which could be informed by standard clinical variables and biomarkers (Dunlop and Mayberg Reference Dunlop and Mayberg2014; Fonseka et al. Reference Fonseka, MacQueen and Kennedy2018; Perlman et al. Reference Perlman, Benrimoh, Israel, Rollins, Brown, Tunteng, You, You, Tanguay-Sela, Snook, Miresco and Berlim2019). The field has started to identify various biomarkers showing promise, such as genetic markers (Breitenstein et al. Reference Breitenstein, Scheuer and Holsboer2014; Laje et al. Reference Laje, Perlis, Rush and McMahon2009), behavioural and cognitive markers (Groves et al. Reference Groves, Douglas and Porter2018; Park et al. Reference Park, Pan, Brietzke, Subramaniapillai, Rosenblat, Zuckerman, Lee, Fus and McIntyre2018; Perna et al. Reference Perna, Alciati, Dacco, Grassi and Caldirola2020), metabolic and inflammatory markers (Lopresti et al. Reference Lopresti, Maker, Hood and Drummond2014; Schmidt et al. Reference Schmidt, Shelton and Duman2011) and neuroimaging markers (Breitenstein et al. Reference Breitenstein, Scheuer and Holsboer2014; Dichter et al. Reference Dichter, Gibbs and Smoski2015; Dunlop and Mayberg Reference Dunlop and Mayberg2014; Fonseka et al. Reference Fonseka, MacQueen and Kennedy2018; Fu et al. Reference Fu, Steiner and Costafreda2013).

Such biomarkers are thought to represent underlying biological substrates of depression, which can be used to predict general prognosis regardless of treatment, better outcome with any treatment, or differential treatment response (Simon and Perlis Reference Simon and Perlis2010). For example, baseline metabolic profile was found to differentiate between responders and non-responders to sertraline or placebo (Kaddurah-Daouk et al. Reference Kaddurah-Daouk, Boyle, Matson, Sharma, Matson, Zhu, Bogdanov, Churchill, Krishnan, Rush, Pickering and Delnomdedieu2011), baseline C-reactive protein differentially predicted response to escitalopram or nortriptyline (Uher et al. Reference Uher, Tansey, Dew, Maier, Mors, Hauser, Zvezdana Dernovsek, Henigsberg, Souery, Farmer and McGuffin2014), and baseline resting-state functional connectivity with the subgenual cortex differentially predicted response to antidepressant treatment or cognitive behavioural therapy (Dunlop et al. Reference Dunlop, Rajendra, Craighead, Kelley, McGrath, Choi, Kinkead, Nemeroff and Mayberg2017).

It is important to note, however, that MDD is a multifaceted disorder associated with a wide range of cognitive, behavioural, emotional and physiological symptoms (Disner et al. Reference Disner, Beevers, Haigh and Beck2011). As such, it is unlikely that a single clinical or biological marker can predict treatment outcome (Patel et al. Reference Patel, Khalaf and Aizenstein2016; Phillips et al. Reference Phillips, Chase, Sheline, Etkin, Almeida, Deckersbach and Trivedi2015). In fact, Lee et al. (Reference Lee, Ragguett, Mansur, Boutilier, Rosenblat, Trevizol, Brietzke, Lin, Pan, Subramaniapillai, Chan, Fus, Park, Musial, Zuckerman, Chen, Ho, Rong and McIntyre2018) showed that models informed by multiple data types, such as a composite of clinical features, neuroimaging, or genetic measures, were more accurate at predicting outcome than less complex models. Nonetheless, current clinical practice is mostly based on questionnaire- and interview-based assessments, which represent a wealth of clinical data which can be used to predict treatment outcome (Rost et al. Reference Rost, Binder and Bruckl2022).

In recent years, machine-learning methods have been increasingly employed to examine which clinical variables are most predictive of response or remission, allowing identification of patterns of information at an individual patient level (Chekroud et al. Reference Chekroud, Bondar, Delgadillo, Doherty, Wasil, Fokkema, Cohen, Belgrave, DeRubeis, Iniesta, Dwyer and Choi2021; Jankowsky et al. Reference Jankowsky, Krakay, Schroeders, Zwerenz and Beutel2024). Various studies have consistently identified baseline symptom severity, number of depressive episodes and co-morbid anxiety disorders as predictors of treatment outcome (Balestri et al. Reference Balestri, Calati, Souery, Kautzky, Kasper, Montgomery, Zohar, Mendlewicz and Serretti2016; Bartova et al. Reference Bartova, Dold, Kautzky, Fabbri, Spies, Serretti, Souery, Mendlewicz, Zohar, Montgomery, Schosser and Kasper2019; Chekroud et al. Reference Chekroud, Zotti, Shehzad, Gueorguieva, Johnson, Trivedi, Cannon, Krystal and Corlett2016; Iniesta et al. Reference Iniesta, Malki, Maier, Rietschel, Mors, Hauser, Henigsberg, Dernovsek, Souery, Stahl, Dobson, Aitchison, Farmer, Lewis, McGuffin and Uher2016; Kautzky et al. Reference Kautzky, Dold, Bartova, Spies, Vanicek, Souery, Montgomery, Mendlewicz, Zohar, Fabbri, Serretti, Lanzenberger and Kasper2018; Perlis Reference Perlis2013). However, standard clinical variables alone capture a limited amount of variance in clinical outcome, with estimates in the region of 5–10% (Iniesta et al. Reference Iniesta, Malki, Maier, Rietschel, Mors, Hauser, Henigsberg, Dernovsek, Souery, Stahl, Dobson, Aitchison, Farmer, Lewis, McGuffin and Uher2016; Perlis Reference Perlis2013), and they tend to perform worse than neuroimaging measures (Dunlop Reference Dunlop2015; Jollans and Whelan Reference Jollans and Whelan2016; Lee et al. Reference Lee, Ragguett, Mansur, Boutilier, Rosenblat, Trevizol, Brietzke, Lin, Pan, Subramaniapillai, Chan, Fus, Park, Musial, Zuckerman, Chen, Ho, Rong and McIntyre2018; Poirot et al. Reference Poirot, Ruhe, Mutsaerts, Maximov, Groote, Bjørnerud, Marquering, Reneman and Caan2024; Schmaal et al. Reference Schmaal, Marquand, Rhebergen, van Tol, Ruhe, van der Wee, Veltman and Penninx2015).

Neuroimaging measures may be of particular interest, as dysfunctional neural processes are core to the development and maintenance of depressive symptoms (Godlewska Reference Godlewska2020). They capture emotional biases associated with depression, such as the tendency to focus more on negative facial expressions than positive ones (Bourke et al. Reference Bourke, Douglas and Porter2010; Krause et al. Reference Krause, Linardatos, Fresco and Moore2021), proneness to experience excessive self-blaming emotions, such as overgeneralised guilt and disgust/contempt towards oneself (Duan et al. Reference Duan, Lawrence, Valmaggia, Moll and Zahn2021; Duan et al. Reference Duan, Valmaggia, Fennema, Moll and Zahn2023; Green et al. Reference Green, Moll, Deakin, Hulleman and Zahn2013; Weiner Reference Weiner1985; Zahn et al. Reference Zahn, Lythe, Gethin, Green, Deakin, Young and Moll2015), as well as rumination, i.e., a tendency to engage in recursive, automatic thoughts often linked to self-critical thinking (Berman et al. Reference Berman, Misic, Buschkuehl, Kross, Deldin, Peltier, Churchill, Jaeggi, Vakorin, McIntosh and Jonides2014; Hamilton et al. Reference Hamilton, Farmer, Fogelman and Gotlib2015; Nolen-Hoeksema et al. Reference Nolen-Hoeksema, Wisco and Lyubomirsky2008).

Leading neuroanatomical models of MDD propose that impaired function within prefrontal-limbic neural circuits, particularly the subgenual cingulate cortex and amygdala, explains disruptions of emotional processing and regulation associated with depression (Price and Drevets Reference Price and Drevets2010; Ressler and Mayberg Reference Ressler and Mayberg2007). Neuroimaging biomarkers capturing the aforementioned – often implicit – emotional biases associated with depression have shown promise in predicting prognosis in MDD at an individual level, notably amygdala activation in response to emotional faces (Williams et al. Reference Williams, Korgaonkar, Song, Paton, Eagles, Goldstein-Piekarski, Grieve, Harris, Usherwood and Etkin2015) and resting-state posterior subgenual cortex connectivity (Dunlop et al. Reference Dunlop, Rajendra, Craighead, Kelley, McGrath, Choi, Kinkead, Nemeroff and Mayberg2017) in current MDD, and self-blame-selective anterior temporal-subgenual connectivity in remitted MDD (Lawrence et al. Reference Lawrence, Stahl, Duan, Fennema, Jaeckle, Young, Dazzan, Moll and Zahn2022). Despite these promising findings, studies tend to focus on treatment-naïve and treatment-free samples of MDD, and it is unclear whether these neural signatures generalise to pragmatic samples of patients encountered in clinical settings. Moreover, it is important to establish whether imaging measures provide added value in predicting clinical outcomes compared to standard baseline clinical variables.

Here, we probed the potential of these neural signatures of emotional biases in predicting clinical outcomes in a pragmatic sample of difficult-to-treat MDD after four months of primary care. These pre-registered (NCT04342299) neural signatures were selected based on their potential to predict response to treatment at an individual level and cover complementary neurocognitive aspects of MDD, i.e., self-blaming biases, negative perceptual biases and dysfunction of task-independent subgenual networks.

Methods

The functional MRI (fMRI) dataset reported here was collected as part of an observational sub-study within a feasibility trial, the Antidepressant Advisor Study (NCT03628027) (Harrison et al. Reference Harrison, Carr, Goldsmith, Young, Ashworth, Fennema, Barrett and Zahn2020; Harrison et al. Reference Harrison, Carr, Goldsmith, Young, Ashworth, Fennema, Duan, Barrett and Zahn2022). We have published tasked-based functional imaging (Fennema et al. Reference Fennema, Barker, O’Daly, Duan, Carr, Goldsmith, Young, Moll and Zahn2023; Fennema, Barker, O’Daly, Duan, Godlewska, et al. Reference Fennema, Barker, O’Daly, Duan, Godlewska, Goldsmith, Young, Moll and Zahn2024) and resting-state fMRI results (Fennema, Barker, O’Daly, Duan, Carr, et al. Reference Fennema, Barker, O’Daly, Duan, Carr, Goldsmith, Young, Moll and Zahn2024) from the same cohort previously, but here, we report on the prediction model for the first time.

Participants

Forty-five participants fulfilled criteria for current MDD according to the Diagnostic and Statistical Manual of Mental Health Disorders, Fifth Edition (DSM-5) (First et al. Reference First, Williams, Karg and Spitzer2015) and had not responded to at least two serotonergic antidepressants. Participants were encouraged to book an appointment with their general practitioner (GP) to review their medication and followed up after receiving four months of standard care. For more information about inclusion/exclusion criteria, recruitment and assessment, please see Supplementary Methods.

Prior to their medication review, participants attended an fMRI session, consisting of three paradigms: the moral sentiment task (assessing self-blame-related biases), the subliminal faces task (assessing bias in emotional processing), and a resting-state fMRI scan. As part of the moral sentiment task, participants viewed self- and other-blaming emotion-evoking statements. Participants were shown written statements describing actions counter to socio-moral values described by social concepts (e.g., impatient, dishonest) in which the agent was either the participant (self-agency) or their best friend (other-agency) (Fennema et al. Reference Fennema, Barker, O’Daly, Duan, Carr, Goldsmith, Young, Moll and Zahn2023). As part of the subliminal faces task, participants were presented with a series of faces. The faces were shown in pairs, briefly displaying a “target” face (expressing sad, happy or neutral emotion) followed by another “mask” face of neutral expression (Fennema, Barker, O’Daly, Duan, Godlewska, et al., Reference Fennema, Barker, O’Daly, Duan, Godlewska, Goldsmith, Young, Moll and Zahn2024). As part of the resting-state fMRI scan, participants were instructed to keep their eyes open and let their mind wander while focusing on a cross (Fennema, Barker, O’Daly, Duan, Carr, et al., Reference Fennema, Barker, O’Daly, Duan, Carr, Goldsmith, Young, Moll and Zahn2024). For more details on the fMRI paradigms, please see Supplementary Materials.

Main outcome

As stated in our pre-registered protocol (NCT04342299), we used a continuous measure of clinical outcome rather than categorising participants into responders and non-responders using the standard definition of a 50% reduction (Nierenberg and DeCecco Reference Nierenberg and DeCecco2001) in self-reported Quick Inventory of Depressive Symptomatology (16-item; QIDS-SR16) (Rush et al. Reference Rush, Trivedi, Ibrahim, Carmody, Arnow, Klein, Markowitz, Ninan, Kornstein, Manber, Thase, Kocsis and Keller2003) scores, due to an unbalanced split between the resulting groups (responders n = 8; non-responders n = 30). The outcome was defined as the percentage change at follow-up from baseline on our pre-registered primary outcome measure, QIDS-SR16, where negative scores corresponded to a reduction in depressive symptoms.

fMRI measures

Statistical Parametric Mapping 12 was used for blood-oxygen-level-dependent (BOLD) effect analysis and psychophysiological interaction analysis, while Data Processing Assistant for Resting-State fMRI was used for resting-state analysis (please see Supplementary Methods for more details). Regression coefficient averages (moral sentiment task and subliminal faces task) and cluster mean z-score (resting-state scan) over our pre-registered regions-of-interest (ROIs) were extracted for individual participants using the MarsBaR toolbox (Rorden and Brett Reference Rorden and Brett2000), i.e., self-blame-selective connectivity between the right superior anterior temporal lobe (RSATL) and posterior subgenual cortex (Brodmann Area [BA] 25), bilateral amygdala BOLD activation for subliminal sad vs happy faces, and resting-state functional connectivity between the bilateral posterior subgenual cortex and left ventrolateral prefrontal cortex (BA47; ventrolateral prefrontal cortex [VLPFC])/insula. For more details, please see Supplementary Materials.

Statistical analysis

Multiple linear regression was used to assess potential predictors of QIDS-SR16 percentage change, as well as an exploratory logistic regression to determine likelihood of response vs. non-response. The aim of the study was to estimate the effect size of using our pre-registered imaging measures as predictors of clinical outcomes, rather than tease out the importance of each predictor given the limitations of our sample size. As such, we ran our main “fMRI” multivariable model which assessed the contribution of our three pre-registered fMRI measures as outlined above, with baseline Maudsley Modified Patient Health Questionnaire, 9 items (MM-PHQ-9; measure of severity of depressive symptoms) (Harrison et al. Reference Harrison, Walton, Fennema, Duan, Jaeckle, Goldsmith, Carr, Ashworth, Young and Zahn2021) as a covariate.

In addition, we ran a supplementary “clinical” multivariable model to compare the contribution of standard clinical measures (baseline MM-PHQ-9, baseline Generalised Anxiety Disorder, 7-items (GAD-7; measure of severity of anxiety symptoms) (Spitzer et al. Reference Spitzer, Kroenke, Williams and Lowe2006), and Maudsley Staging Method total score (proxy of treatment-resistance based on duration, severity and treatment failures) (Fekadu et al. Reference Fekadu, Donocik and Cleare2018); please see Supplementary Methods for more details on the clinical measures). Another supplementary “high-quality fMRI” multivariable model assessed the impact of suboptimal fMRI quality, i.e., signal drop-out and/or more motion, on the predictive value of the fMRI measures, including only participants with high-quality fMRI data for all three scans (n=30). Other supplementary models considered the individual contribution of the pre-registered fMRI measures (please see Supplementary Methods and Results).

Please note that our pre-registered imaging measures also included additional regions of interest: functional resting-state subgenual cortex connectivity with the left ventromedial prefrontal cortex (BA10) and with the dorsal midbrain (Fennema, Barker, O’Daly, Duan, Carr, et al. Reference Fennema, Barker, O’Daly, Duan, Godlewska, Goldsmith, Young, Moll and Zahn2024), as well as pregenual anterior cingulate cortex BOLD activation for subliminal sad vs happy faces (Fennema, Barker, O’Daly, Duan, Godlewska, et al. Reference Fennema, Barker, O’Daly, Duan, Godlewska, Goldsmith, Young, Moll and Zahn2024). However, as our sample size only allowed us to model a limited number of variables without risk of overfitting, for our primary prediction model, we solely included variables showing univariate prediction effects in our previous analyses (Fennema, Barker, O’Daly, Duan, Carr, et al. Reference Fennema, Barker, O’Daly, Duan, Carr, Goldsmith, Young, Moll and Zahn2024; Fennema, Barker, O’Daly, Duan, Godlewska, et al., Reference Fennema, Barker, O’Daly, Duan, Godlewska, Goldsmith, Young, Moll and Zahn2024). For more details on the exploratory “pre-registration” model, please see Supplementary Methods.

All variables were Fisher Z-transformed to derive beta coefficients and corresponding standard error. Correlation analysis (Spearman’s rho) was used to investigate the association between the pre-registered neural signatures. To test whether there is any link between treatment change and symptom change, a one-way analysis of variance was conducted (please see Supplementary Methods for a description of treatment change). All tests were carried out using IBM SPSS Statistics 27, using a significance threshold of p = .05, two-tailed.

Results

Subgroup characteristics

Table 1 presents participant characteristics at baseline, split by responders and non-responders. Of 45 included participants, 38 had usable fMRI data (31 [82%] female, mean [SD] age = 41.8 [14.8] years). Most participants fulfilled the DSM-5 anxious distress specifier criteria (82%) and met criteria for a life-time axis I co-morbidity (87%). Average baseline depression severity was severe according to MM-PHQ-9 (mean [SD] = 18.7 [4.7]) and QIDS-SR16 (mean [SD] = 17.3 [3.5]), and 82% of the participants were taking a selective serotonin-reuptake inhibitor. There were no significant differences between responders and non-responders at baseline in terms of demographic and clinical characteristics (t < 1.31 and p > .20), except for current major depressive episode duration (responders mean [SD] = 6.3 [5.3]; non-responders mean [SD] = 32.8 [50.2]; t[31.2] = −2.85, p = .01).

Table 1. Baseline demographic and clinical characteristics by responders and non-responders (n = 38)

Percentages may not add up to 100 due to rounding. MDD = major depressive disorder; DSM-5 = Diagnostic and Statistical Manual for Mental Disorders 5th edition; MDE = major depressive episode; SD = standard deviation; MM-PHQ-9 = Maudsley Modified Patient Health Questionnaire, 9 items; QIDS-SR16 = Quick Inventory Depressive Symptomatology, self-rated, 16 items; MADRS = Montgomery-Åsberg Depression Rating Scale; SOFAS = Social and Occupational Functioning Scale; SSRI = selective serotonin reuptake inhibitor; SNRI = selective norepinephrine reuptake inhibitor; GAD-7 = Generalised Anxiety Disorder, 7 items.

a Missing data for one participant. Ethnicity categories have been combined: “White” includes White: British, Other and Polish; “Asian” includes Asian or Asian British: Indian, Chinese and Other Asian; “Black” includes Black or Black British: Caribbean.

As part of the study, participants were encouraged to book an appointment with their GP to review their antidepressant medication. Even though UK care guidelines would recommend changing antidepressant medications in non-responders, unexpectedly, more than half (55%) did not change their medication and some even stopped their medication (16%; Supplementary Table 1). Despite little change in treatment, on average, participants showed a significant reduction in depressive symptoms from baseline to follow-up in QIDS-SR16 scores (mean [95% CI] = −4.1 [−5.8, −2.4]). This was also the case for other self- and observer-rated scores (Supplementary Table 2).

There was a mean percentage change [SD] of −23.1 [30.0] in QIDS-SR16: those with a relevant change showed the most improvement in QIDS-SR16 (mean percentage change [SD] = −43.8 [20.3]), followed by participants with a minimal change (mean percentage change [SD] = −32.1 [32.4]) and participants with no change (mean percentage change [SD] = −17.6 [29.6]). However, there was no significant difference between the groups (F[2,37] = 1.78, p = .18).

Prediction models

The “fMRI” model using the pre-registered fMRI measures with baseline MM-PHQ-9 as a covariate explained 32% of the variance of QIDS-SR16 percentage change (F[4,33] = 3.86, p = .01, R 2 = .32, R 2 adjusted = .24; Table 2). When including all previously pre-registered regions, the overall prediction effect for the “pre-registration” model was comparable (R 2 = 33%, please see Supplementary Results). When limiting to “high-quality fMRI,” the model explained 43% of the variance of QIDS-SR16 percentage change (F[4,25] = 4.67, p = .01, R 2 = .43, R 2 adjusted = .34; Supplementary Table 3). In contrast, the “clinical” model using standard clinical measures at baseline, i.e. MM-PHQ-9, GAD-7 and Maudsley Staging Method, explained only 3% of the variance of QIDS-SR16 percentage change (F[3,34] = .33, p = .81, R 2 = .03, R 2 adjusted = −.06; Table 2).

Table 2. Prediction models of clinical outcomes in depression (n = 38)

* Significant at p < .05 threshold, two-tailed. SE = standard error; MM-PHQ-9 = Maudsley Modified Patient Health Questionnaire, 9 items; GAD-7 = Generalised Anxiety Disorder, 7 items; RSATL = right superior anterior temporal lobe; BA = Brodmann Area; BOLD = blood-oxygen level-dependent; VLPFC = ventrolateral prefrontal cortex.

Bilateral amygdala BOLD activation positively contributed to the variance in QIDS-SR16 percentage change (partial β = 11.11, t[33] = 2.21), while partial effects of resting-state functional connectivity between the posterior subgenual cortex and left VLPFC/insula as well as self-blame-selective RSATL-BA25 connectivity contributed negatively (resting-state: partial β = −8.15, t[33] = −1.95; RSATL-BA25: partial β = −7.28, t[33] = −1.71; Figure 1). Please see Supplementary Results and Supplementary Table 3 for exploratory separate prediction models for each fMRI paradigm showing a maximum of 18% variance in clinical outcomes explained, when using the bilateral amygdala BOLD signature.

Figure 1. Neural signatures of emotional biases associated with clinical outcomes in difficult-to-treat MDD. Three neural signatures of emotional biases were associated with clinical outcomes in UK primary care. More specifically, it shows cropped sections of voxel-based analyses illustrating the respective pre-registered a priori regions-of-interest, i.e. self-blame-selective right superior anterior temporal lobe-posterior subgenual cortex (BA25) connectivity, resting-state functional connectivity between the subgenual cortex and ventrolateral prefrontal cortex/insula, and bilateral amygdala blood-oxygen-level-dependent activation in response to subliminal sad vs happy faces. These cropped sections are displayed using MRIcron at an uncorrected voxel-level threshold of p=.005, with no cluster-size threshold (the colour bar represents t values) and adapted from figures previously published (Fennema et al. Reference Fennema, Barker, O’Daly, Duan, Carr, Goldsmith, Young, Moll and Zahn2023; Fennema, Barker, O’Daly, Duan, Carr, et al. Reference Fennema, Barker, O’Daly, Duan, Carr, Goldsmith, Young, Moll and Zahn2024; Fennema, Barker, O’Daly, Duan, Godlewska, et al. Reference Fennema, Barker, O’Daly, Duan, Godlewska, Goldsmith, Young, Moll and Zahn2024). A linear model using the pre-registered fMRI measures with baseline Maudsley Modified Patient Health Questionnaire (9 items) as a covariate explained 32% of the variance of QIDS-SR16 percentage change. The red and green lines display the partial effects of the fMRI measures on the variance of QIDS-SR16 percentage change after four months of standard primary care. MDD = major depressive disorder; BA = Brodmann Area; RSATL = right superior anterior temporal lobe; VLPFC = ventrolateral prefrontal cortex; BOLD = blood-oxygen level-dependent; QIDS-SR16 = Quick Inventory of Depressive Symptomatology, self-rated (16 items).

Notably, there were no bivariate associations between the three pre-registered fMRI measures (self-blame-selective RSATL-BA25 connectivity and bilateral amygdala BOLD activation: r s [38] = −.06, p = .71; self-blame-selective RSATL-BA25 connectivity and resting-state functional connectivity between posterior subgenual cortex and left VLPFC/insula: r s [38] = .09, p = .61; bilateral amygdala BOLD activation and resting-state functional connectivity between posterior subgenual cortex and left VLPFC/insula: r s [38] = −.09, p = .61).

Exploratory findings responders vs. non-responders

A logistic regression was performed to determine the effects of the pre-registered neural measures and baseline MM-PHQ-9 on the likelihood of response vs. non-response. The logistic regression model was statistically significant, χ 2 (4) = 11.09, p = .03. The model explained 39% (Nagelkerke R 2) of the variance in responders and correctly classified 81.6% of the cases. Increased functional connectivity between the bilateral subgenual cortex and left VLPFC/insula was associated with an increased likelihood of response. For more details, please see Supplementary Results.

Discussion

To our knowledge, this is the first study to combine complementary functional imaging measures of affective circuits in MDD and to probe their role in prospectively predicting clinical outcomes in a pragmatic setting. We show that neuroimaging markers hold promise: the model with the three pre-registered fMRI measures explained more variance in clinical outcomes compared with the clinical model, i.e. 32% vs 3%. The model that only included participants with high-quality fMRI measures explained an even larger amount of variance (43%), highlighting the need to adequately account for signal drop-out and/or motion artefacts. However, it is important to acknowledge that no formal statistical tests were undertaken to compare the regression models as the study was not powered for such comparisons, which limits the interpretability of differences between the models.

Interestingly, the effects of the three pre-registered fMRI measures were uncorrelated, showing that these may capture distinct aspects of MDD pathophysiology, i.e. self-blaming biases (right superior anterior temporal-subgenual connectivity), negative perceptual biases (amygdala) and dysfunction of task-independent subgenual networks. If these neural signatures were to relate to specific subtypes rather than independently predicting the same underlying general pathophysiology, then this would offer the intriguing possibility of stratification for neuromodulation and neurofeedback studies based on distinct neural circuits of interest, by either modulating self-blaming or emotional perception biases in patients non-responsive to standard treatments. The feasibility of such interventions has recently been confirmed, with reports of a training-induced reduction in self-blame-selective connectivity (Jaeckle et al. Reference Jaeckle, Williams, Barker, Basilio, Carr, Goldsmith, Colasanti, Giampietro, Cleare, Young, Moll and Zahn2023) and an enhancement of amygdala responsiveness to positive autobiographical memories (Young et al. Reference Young, Siegle, Zotev, Phillips, Misaki, Yuan, Drevets and Bodurka2019).

However, it is important to first determine whether these neural signatures represent a trait-like feature of a fully remitting subtype of MDD, or whether it is also modulated by depressive state. For example, both self-blame-related and emotional perception-related changes have been identified in remitted MDD (Joormann and Gotlib Reference Joormann and Gotlib2007; Lythe et al. Reference Lythe, Gethin, Workman, Lambon Ralph, Deakin, Moll and Zahn2022; Ruhe et al. Reference Ruhe, Mocking, Figueroa, Seeverens, Ikani, Tyborowska, Browning, Vrijsen, Harmer and Schene2019). It is unclear whether these changes are more pronounced when people develop a recurrent episode or are merely due to underlying vulnerabilities which are not modulated by symptomatic state. This question is key to a deeper pathophysiological understanding of MDD in that little is known about how trait-related changes interact with precipitating biological and psychological trigger events to result in a depressive brain state, and how it affects subsequent response to treatment.

Limitations

Due to our relatively modest sample size, we were unable to use cross-validated and data-driven machine learning algorithms, which may have improved the prediction model performance. Moreover, our sample consisted of chronic MDD patients, often with anxious distress and other co-morbidities. In addition, treatment was not standardised and, unlike previous studies in randomised controlled trials, did not allow us to distinguish spontaneous remission and placebo effects from treatment-related effects. Given the selection biases in randomised controlled trials, however, it was important to investigate a pragmatic sample as we have undertaken in this study.

Clinical utility is complicated by the heterogenous nature of MDD, resulting in patients with a wide variety of symptoms, disease severity and treatment history (Strawbridge et al. Reference Strawbridge, Young and Cleare2017), as well as patient response to treatment (Mayberg and Dunlop Reference Mayberg and Dunlop2023). Further complementary predictive measures, such as novel cognitive markers (Lawrence et al. Reference Lawrence, Stahl, Duan, Fennema, Jaeckle, Young, Dazzan, Moll and Zahn2022), would be useful in addition to imaging markers to achieve clinically relevant levels of individual prediction of response to specific types of treatment.

Moreover, it is important to acknowledge that percentage-based reduction scores to define treatment response have been criticised, as it is biased towards more severe depressive symptoms at baseline (Rost et al. Reference Rost, Binder and Bruckl2022). As a result, it is plausible for a responder to still experience clinically significant distress or impairment when starting with a baseline score in the severe range, while a non-responder may show a clinically significant improvement – which was also observed in the current study.

Conclusions

Taken together, we reproduced clinically relevant neural signatures in an independent, pragmatic sample of difficult-to-treat MDD. The findings confirm the pathophysiological relevance and potential of the proposed candidate neural signatures to make relevant contributions to the prospective prediction of clinical outcomes in more chronic, difficult-to-treat forms of MDD and call for stratified neurofeedback and neuromodulation interventions.

Supplementary material

The supplementary material for this article can be found at https://doi.org/10.1017/dep.2024.6.

Data availability statement

The data that support the findings of this study are available on request from the corresponding author, RZ. We will only be able to share fully anonymised, no pseudonymised data and requests will have to go through a King’s College London repository.

Acknowledgements

We are grateful to Drs Mark Ashworth and Barbara Barrett who contributed to the trial study design, and to Drs Phillippa Harrison and Suqian Duan who collected trial data. We also thank the study participants for their support.

Part of the study has been published in a PhD thesis available on the King’s College London institutional repository, Pure, see Fennema (Reference Fennema2022): https://kclpure.kcl.ac.uk/portal/en/studentTheses/neural-signatures-of-emotional-biases-and-prognosisin-treatment.

Author contribution

Diede Fennema: Conceptualisation, Methodology, Formal analysis, Investigation, Data curation, Writing – original draft, Visualisation, Funding acquisition. Gareth Barker: Conceptualisation, Methodology, Writing – review & editing, Supervision. Owen O’Daly: Conceptualisation, Methodology, Writing – review & editing. Beata Godlewska: Methodology, Resources, Writing – review & editing. Ewan Carr: Conceptualisation, Methodology, Writing – review & editing. Kimberley Goldsmith: Conceptualisation, Methodology, Writing – review & editing. Allan Young: Conceptualisation, Methodology, Writing – review & editing, Supervision, Project administration, Funding acquisition. Jorge Moll: Conceptualisation, Writing – review & editing. Roland Zahn: Conceptualisation, Methodology, Formal analysis, Writing – review & editing, Supervision, Project administration, Funding acquisition.

Financial support

This study represents independent research funded by the National Institute for Health and Care Research (NIHR) under its Research for Patient Benefit (RfPB) Programme (Grant Reference Number PB-PG-0416-20039) and independent research part funded by the National Institute for Health and Care (NIHR) Biomedical Research Centre at South London and Maudsley National Health Service (NHS) Foundation Trust and King’s College London (Profs Zahn, Young, Goldsmith; Dr Carr). Prof Zahn was also partly funded by a Medical Research Council grant (ref. MR/T017538/1), while Prof Goldsmith was also supported by the National Institute for Health and Care Research (NIHR) Applied Research Collaboration South London (NIHR ARC South London) at King’s College Hospital NHS Foundation Trust. This study was also supported by the Rosetrees Trust (M816) awarded to Prof Zahn. Dr Fennema was funded by a Medical Research Council Doctoral Training Partnership Studentship (ref. 2064430) and partly supported by a King’s College London/D’Or Institute for Research and Education (KCL/IDOR) Pioneer Science Fellowship, funded by Scients Institute and the IDOR Pioneer Science Initiative. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care. Additional support was provided to the study by the South London Clinical Research Network and sponsorship by Lambeth CCG.

Competing interests

Prof Zahn is a private psychiatrist service provider at The London Depression Institute and co-investigator on a Livanova-funded observational study of Vagus Nerve Stimulation (VNS) for Depression. Prof Zahn has received honoraria for talks at medical symposia sponsored by Lundbeck as well as Janssen. Prof Zahn has collaborated with EMOTRA, EMIS PLC and Depsee Ltd. Prof Zahn is affiliated with the D’Or Institute of Research and Education, Rio de Janeiro and advises the Scients Institute, USA. Prof Barker receives honoraria for teaching from GE Healthcare. Prof Young is employed by King’s College London as an honorary consultant in the South London and Maudsley Trust (NHS UK) and is a consultant to Johnson & Johnson and Livanova. Prof Young has given paid lectures and sat on advisory open access boards for the following companies with drugs used in affective and related disorders: Astrazenaca, Eli Lilly, Lundbeck, Sunovion, Servier, Livanova, Janssen, Allegan, Bionomics, Sumitomo Dainippon Pharma. Prof Young has received honoraria for attending advisory boards and presenting talks at meetings organised by LivaNova. Prof Young is the Principal Investigator of the following studies: Restore-Life VNS registry study funded by LivaNova, ESKETINTRD3004: ‘An Open-label, Long-term, Safety and Efficacy Study of Intranasal Esketamine in Treatment-resistant Depression’, ‘The Effects of Psilocybin on Cognitive Function in Healthy Participants’ and ‘The Safety and Efficacy of Psilocybin in Participants with Treatment-Resistant Depression (P-TRD)’. Prof Young has received grant funding (past and present) from the following: National Institute of Mental Health (USA); Canadian Institutes for Health Research (Canada); National Alliance for Research on Schizophrenia and Depression (USA); Stanley Medical Research Institute (USA); Medical Research Council (UK); Wellcome Trust (UK); Royal College of Physicians (Edin); British Medical Assocation (UK); University of British Columbia-Vancouver General Hospital Foundation (Canada); Wisconsin Economic Development Corporation (Canada); Coast Capital Savings Depression Research Fund (Canada); Michael Smith Foundation for Health Research (Canada); NIHR (UK); Janssen (UK). Prof Young has no shareholdings in pharmaceutical companies. Prof Goldsmith reports grants from NIHR, Stroke association, National Institutes of Health (US) and Juvenile Diabetes Research Foundation (US) during the conduct of the study. None of the other authors reports biomedical financial interests or potential conflicts of interest related to the subject of this paper.

Ethical standards

Ethical approval was obtained from the NHS Health Research Authority and National Research Ethics Service London – Camberwell St Giles Committee (REC reference: 17/LO/2074). Written informed consent was obtained from all participants.

References

Connections references

Hickie, I and Williams, L (2023) Will new brain circuit focused methods (EEG, fMRI etc) lead to more personalised care options? Research Directions: Depression 1(E12). https://doi.org/10.1017/dep.2023.22.Google Scholar

References

Balestri, M, Calati, R, Souery, D, Kautzky, A, Kasper, S, Montgomery, S, Zohar, J, Mendlewicz, J and Serretti, A (2016) Socio-demographic and clinical predictors of treatment resistant depression: a prospective European multicenter study. Journal of Affective Disorders 189, 224–232. https://doi.org/10.1016/j.jad.2015.09.033.CrossRefGoogle ScholarPubMed
Bartova, L, Dold, M, Kautzky, A, Fabbri, C, Spies, M, Serretti, A, Souery, D, Mendlewicz, J, Zohar, J, Montgomery, S, Schosser, A and Kasper, S (2019) Results of the European Group for the Study of Resistant Depression (GSRD) - basis for further research and clinical practice. The World Journal of Biological Psychiatry 20(6), 427–448. https://doi.org/10.1080/15622975.2019.1635270.CrossRefGoogle Scholar
Berman, MG, Misic, B, Buschkuehl, M, Kross, E, Deldin, PJ, Peltier, S, Churchill, NW, Jaeggi, SM, Vakorin, V, McIntosh, AR and Jonides, J (2014) Does resting-state connectivity reflect depressive rumination? A tale of two analyses. Neuroimage 103, 267–279. https://doi.org/10.1016/j.neuroimage.2014.09.027.CrossRefGoogle ScholarPubMed
Bourke, C, Douglas, K and Porter, R (2010) Processing of facial emotion expression in major depression: a review. Australian & New Zealand Journal of Psychiatry 44(8), 681–696. https://doi.org/10.3109/00048674.2010.496359.CrossRefGoogle ScholarPubMed
Breitenstein, B, Scheuer, S and Holsboer, F (2014) Are there meaningful biomarkers of treatment response for depression? Drug Discovery Today 19(5), 539–561. https://doi.org/10.1016/j.drudis.2014.02.002.CrossRefGoogle ScholarPubMed
Chekroud, AM, Bondar, J, Delgadillo, J, Doherty, G, Wasil, A, Fokkema, M, Cohen, Z, Belgrave, D, DeRubeis, R, Iniesta, R, Dwyer, D and Choi, K (2021) The promise of machine learning in predicting treatment outcomes in psychiatry. World Psychiatry 20(2), 154–170. https://doi.org/10.1002/wps.20882.CrossRefGoogle ScholarPubMed
Chekroud, AM, Zotti, RJ, Shehzad, Z, Gueorguieva, R, Johnson, MK, Trivedi, MH, Cannon, TD, Krystal, JH and Corlett, PR (2016) Cross-trial prediction of treatment outcome in depression: a machine learning approach. The Lancet Psychiatry 3(3), 243–250. https://doi.org/10.1016/s2215-0366(15)00471-x.CrossRefGoogle ScholarPubMed
Dichter, GS, Gibbs, D and Smoski, MJ (2015) A systematic review of relations between resting-state functional-MRI and treatment response in major depressive disorder. Journal of Affective Disorders 172, 8–17. https://doi.org/10.1016/j.jad.2014.09.028.CrossRefGoogle ScholarPubMed
Disner, SG, Beevers, CG, Haigh, EA and Beck, AT (2011) Neural mechanisms of the cognitive model of depression. Nature Reviews Neuroscience 12(8), 467–477. https://doi.org/10.1038/nrn3027.CrossRefGoogle ScholarPubMed
Duan, S, Lawrence, AJ, Valmaggia, L, Moll, J and Zahn, R (2021) Maladaptive blame-related action tendencies are associated with vulnerability to major depressive disorder. Journal of Psychiatric Research 145, 70–76. https://doi.org/10.1016/j.jpsychires.2021.11.043.CrossRefGoogle ScholarPubMed
Duan, S, Valmaggia, L, Fennema, D, Moll, J and Zahn, R (2023) Remote virtual reality assessment elucidates self-blame-related action tendencies in depression. Journal of Psychiatric Research 161, 77–83. https://doi.org/10.1016/j.jpsychires.2023.02.031.CrossRefGoogle ScholarPubMed
Dunlop, BW (2015) Prediction of treatment outcomes in major depressive disorder. Expert Review of Clinical Pharmacology 8(6), 669–672. https://doi.org/10.1586/17512433.2015.1075390.CrossRefGoogle ScholarPubMed
Dunlop, BW and Mayberg, HS (2014) Neuroimaging-based biomarkers for treatment selection in major depressive disorder. Dialogues in Clinical Neuroscience 16(4), 479–490. https://doi.org/10.31887/DCNS.2014.16.4/bdunlop.CrossRefGoogle ScholarPubMed
Dunlop, BW, Rajendra, JK, Craighead, WE, Kelley, ME, McGrath, CL, Choi, KS, Kinkead, B, Nemeroff, CB and Mayberg, HS (2017) Functional connectivity of the subcallosal cingulate cortex and differential outcomes to treatment with cognitive-behavioral therapy or antidepressant medication for major depressive disorder. American Journal of Psychiatry 174(6), 533–545. https://doi.org/10.1176/appi.ajp.2016.16050518.CrossRefGoogle ScholarPubMed
Fekadu, A, Donocik, JG and Cleare, AJ (2018) Standardisation framework for the Maudsley staging method for treatment resistance in depression. BMC Psychiatry 18(1), 100. https://doi.org/10.1186/s12888-018-1679-x.CrossRefGoogle ScholarPubMed
Fennema, D (2022) Neural signatures of emotional biases and prognosis in treatment-resistant depression. PhD theses, King’s College London.Google Scholar
Fennema, D, Barker, GJ, O’Daly, O, Duan, S, Carr, E, Goldsmith, K, Young, AH, Moll, J and Zahn, R (2023) Self-blame-selective hyper-connectivity between anterior temporal and subgenual cortices predicts prognosis in major depressive disorder. NeuroImage: Clinical 39, 103453. https://doi.org/10.1016/j.nicl.2023.103453.CrossRefGoogle ScholarPubMed
Fennema, D, Barker, GJ, O’Daly, O, Duan, S, Carr, E, Goldsmith, K, Young, AH, Moll, J and Zahn, R (2024) The role of subgenual resting-state connectivity networks in predicting prognosis in major depressive disorder. Biological Psychiatry Global Open Science 4(3), 100308. https://doi.org/10.1016/j.bpsgos.2024.100308.CrossRefGoogle ScholarPubMed
Fennema, D, Barker, GJ, O’Daly, O, Duan, S, Godlewska, BR, Goldsmith, K, Young, AH, Moll, J and Zahn, R (2024) Neural responses to facial emotions and subsequent clinical outcomes in difficult-to-treat depression. Psychological Medicine. Advance online publication. https://doi.org/10.1017/S0033291724001144.CrossRefGoogle ScholarPubMed
First, MB, Williams, JBW, Karg, RS and Spitzer, RL (2015) Structured Clinical Interview for DSM-5 - Research Version (SCID-5 for DSM-5, Research Version; SCID-5-RV, Version 1.0.0). Arlington, VA: American Psychiatric Association.Google Scholar
Fonseka, TM, MacQueen, GM and Kennedy, SH (2018) Neuroimaging biomarkers as predictors of treatment outcome in major depressive disorder. Journal of Affective Disorders 233, 21–35. https://doi.org/10.1016/j.jad.2017.10.049.CrossRefGoogle ScholarPubMed
Fu, CH, Steiner, H and Costafreda, SG (2013) Predictive neural biomarkers of clinical response in depression: a meta-analysis of functional and structural neuroimaging studies of pharmacological and psychological therapies. Neurobiology of Disease 52, 75–83. https://doi.org/10.1016/j.nbd.2012.05.008.CrossRefGoogle ScholarPubMed
Godlewska, BR (2020) Neuroimaging as a tool for individualized treatment choice in depression: the past, the present and the future. Current Behavioral Neuroscience Reports 7(1), 32–39. https://doi.org/10.1007/s40473-020-00198-2.CrossRefGoogle Scholar
Green, S, Moll, J, Deakin, JF, Hulleman, J and Zahn, R (2013) Proneness to decreased negative emotions in major depressive disorder when blaming others rather than oneself. Psychopathology 46(1), 34–44. https://doi.org/10.1159/000338632.CrossRefGoogle ScholarPubMed
Groves, SJ, Douglas, KM and Porter, RJ (2018) A systematic review of cognitive predictors of treatment outcome in major depression. Frontiers in Psychiatry 9, 382. https://doi.org/10.3389/fpsyt.2018.00382.CrossRefGoogle ScholarPubMed
Hamilton, JP, Farmer, M, Fogelman, P and Gotlib, IH (2015) Depressive rumination, the default-mode network, and the dark matter of clinical neuroscience. Biological Psychiatry 78(4), 224–230. https://doi.org/10.1016/j.biopsych.2015.02.020.CrossRefGoogle ScholarPubMed
Harrison, P, Carr, E, Goldsmith, K, Young, AH, Ashworth, M, Fennema, D, Barrett, B and Zahn, R (2020) Study protocol for the antidepressant advisor (ADeSS): a decision support system for antidepressant treatment for depression in UK primary care: a feasibility study. BMJ Open 10(5), e035905. https://doi.org/10.1136/bmjopen-2019-035905.CrossRefGoogle ScholarPubMed
Harrison, P, Carr, E, Goldsmith, K, Young, AH, Ashworth, M, Fennema, D, Duan, S, Barrett, B and Zahn, R (2022) The Antidepressant Advisor (ADeSS): A Decision Support System for Antidepressant Treatment for Depression in UK Primary Care - A Feasibility Study. BMJ Open 13(3), e060516. https://doi.org/10.1136/bmjopen-2021-060516.CrossRefGoogle Scholar
Harrison, P, Walton, S, Fennema, D, Duan, S, Jaeckle, T, Goldsmith, K, Carr, E, Ashworth, M, Young, AH and Zahn, R (2021) Development and validation of the Maudsley Modified Patient Health Questionnaire (MM-PHQ-9). BJPsych Open 7(4), e123. https://doi.org/10.1192/bjo.2021.953.CrossRefGoogle ScholarPubMed
Iniesta, R, Malki, K, Maier, W, Rietschel, M, Mors, O, Hauser, J, Henigsberg, N, Dernovsek, MZ, Souery, D, Stahl, D, Dobson, R, Aitchison, KJ, Farmer, A, Lewis, CM, McGuffin, P and Uher, R (2016) Combining clinical variables to optimize prediction of antidepressant treatment outcomes. Journal of Psychiatric Research 78, 94–102. https://doi.org/10.1016/j.jpsychires.2016.03.016.CrossRefGoogle ScholarPubMed
Jaeckle, T, Williams, SCR, Barker, GJ, Basilio, R, Carr, E, Goldsmith, K, Colasanti, A, Giampietro, V, Cleare, A, Young, AH, Moll, J and Zahn, R (2023) Self-blame in major depression: a randomised pilot trial comparing fMRI neurofeedback with self-guided psychological strategies. Psychological Medicine 53(7), 2831–2841. https://doi.org/10.1017/S0033291721004797.CrossRefGoogle ScholarPubMed
Jankowsky, K, Krakay, L, Schroeders, U, Zwerenz, R and Beutel, ME (2024) Predicting treatment response using machine learning: A registered report. British Journal of Clinical Psychology 63, 137–155. https://doi.org/10.1111/bjc.12452.CrossRefGoogle ScholarPubMed
Jollans, L and Whelan, R (2016) The clinical added value of imaging: a perspective from outcome prediction. Biological Psychiatry: Cognitive Neuroscience and Neuroimaging 1(5), 423–432. https://doi.org/10.1016/j.bpsc.2016.04.005.Google ScholarPubMed
Joormann, J and Gotlib, IH (2007) Selective attention to emotional faces following recovery from depression. Journal of Abnormal Psychology 116(1), 80–85. https://doi.org/10.1037/0021-843X.116.1.80.CrossRefGoogle ScholarPubMed
Kaddurah-Daouk, R, Boyle, SH, Matson, W, Sharma, S, Matson, S, Zhu, H, Bogdanov, MB, Churchill, E, Krishnan, RR, Rush, AJ, Pickering, E and Delnomdedieu, M (2011) Pretreatment metabotype as a predictor of response to sertraline or placebo in depressed outpatients: a proof of concept. Translational Psychiatry 1(7), e26. https://doi.org/10.1038/tp.2011.22.CrossRefGoogle ScholarPubMed
Kautzky, A, Dold, M, Bartova, L, Spies, M, Vanicek, T, Souery, D, Montgomery, S, Mendlewicz, J, Zohar, J, Fabbri, C, Serretti, A, Lanzenberger, R and Kasper, S (2018) Refining prediction in treatment-resistant depression: results of machine learning analyses in the TRD III sample. Journal of Clinical Psychiatry 79(1), 16m11385. https://doi.org/10.4088/JCP.16m11385.CrossRefGoogle ScholarPubMed
Krause, FC, Linardatos, E, Fresco, DM and Moore, MT (2021) Facial emotion recognition in major depressive disorder: A meta-analytic review. Journal of Affective Disorders 293, 320–328. https://doi.org/10.1016/j.jad.2021.06.053.CrossRefGoogle ScholarPubMed
Laje, G, Perlis, RH, Rush, AJ and McMahon, FJ (2009) Pharmacogenetics studies in STAR*D: strengths, limitations, and results. Psychiatric Services 60(11), 1446–1457. https://doi.org/10.1176/appi.ps.60.11.1446.CrossRefGoogle Scholar
Lawrence, AJ, Stahl, D, Duan, S, Fennema, D, Jaeckle, T, Young, AH, Dazzan, P, Moll, J and Zahn, R (2022) Neurocognitive measures of self-blame and risk prediction models of recurrence in major depressive disorder. Biological Psychiatry: Cognitive Neuroscience and Neuroimaging 7(3), 256–264. https://doi.org/10.1016/j.bpsc.2021.06.010.Google ScholarPubMed
Lee, Y, Ragguett, RM, Mansur, RB, Boutilier, JJ, Rosenblat, JD, Trevizol, A, Brietzke, E, Lin, K, Pan, Z, Subramaniapillai, M, Chan, TCY, Fus, D, Park, C, Musial, N, Zuckerman, H, Chen, VC, Ho, R, Rong, C and McIntyre, RS (2018) Applications of machine learning algorithms to predict therapeutic outcomes in depression: a meta-analysis and systematic review. Journal of Affective Disorders 241, 519–532. https://doi.org/10.1016/j.jad.2018.08.073.CrossRefGoogle ScholarPubMed
Lopresti, AL, Maker, GL, Hood, SD and Drummond, PD (2014) A review of peripheral biomarkers in major depression: the potential of inflammatory and oxidative stress biomarkers. Progress in Neuro-Psychopharmacology and Biological Psychiatry 48, 102–111. https://doi.org/10.1016/j.pnpbp.2013.09.017.CrossRefGoogle ScholarPubMed
Lythe, KE, Gethin, JA, Workman, CI, Lambon Ralph, MA, Deakin, JFW, Moll, J and Zahn, R (2022) Subgenual activation and the finger of blame: individual differences and depression vulnerability. Psychological Medicine 52(8), 1560–1568. https://doi.org/10.1017/S0033291720003372.CrossRefGoogle ScholarPubMed
Mayberg, H and Dunlop, BW (2023) Balancing the beautiful and the good in pursuit of biomarkers for depression. World Psychiatry 22(2), 265–267. https://doi.org/10.1002/wps.21081.CrossRefGoogle ScholarPubMed
Nierenberg, AA and DeCecco, L (2001) Definitions of antidepressant treatment response, remission, nonresponse, partial response, and other relevant treatment outcomes: a focus on treatment-resistant depression. Journal of Clinical Psychiatry 62, 5–9.Google ScholarPubMed
Nolen-Hoeksema, S, Wisco, BE and Lyubomirsky, S (2008) Rethinking rumination. Perspectives on Psychological Science 3(5), 400–424. https://doi.org/10.1111/j.1745-6924.2008.00088.x.CrossRefGoogle ScholarPubMed
Park, C, Pan, Z, Brietzke, E, Subramaniapillai, M, Rosenblat, JD, Zuckerman, H, Lee, Y, Fus, D and McIntyre, RS (2018) Predicting antidepressant response using early changes in cognition: a systematic review. Behavioural Brain Research 353, 154–160. https://doi.org/10.1016/j.bbr.2018.07.011.CrossRefGoogle ScholarPubMed
Patel, MJ, Khalaf, A and Aizenstein, HJ (2016) Studying depression using imaging and machine learning methods. NeuroImage: Clinical 10, 115–123. https://doi.org/10.1016/j.nicl.2015.11.003.CrossRefGoogle ScholarPubMed
Perlis, RH (2013) A clinical risk stratification tool for predicting treatment resistance in major depressive disorder. Biological Psychiatry 74(1), 7–14. https://doi.org/10.1016/j.biopsych.2012.12.007.CrossRefGoogle ScholarPubMed
Perlman, K, Benrimoh, D, Israel, S, Rollins, C, Brown, E, Tunteng, JF, You, R, You, E, Tanguay-Sela, M, Snook, E, Miresco, M and Berlim, MT (2019) A systematic meta-review of predictors of antidepressant treatment outcome in major depressive disorder. Journal of Affective Disorders 243, 503–515. https://doi.org/10.1016/j.jad.2018.09.067.CrossRefGoogle ScholarPubMed
Perna, G, Alciati, A, Dacco, S, Grassi, M and Caldirola, D (2020) Personalized psychiatry and depression: the role of sociodemographic and clinical variables. Psychiatry Investigation 17(3), 193–206. https://doi.org/10.30773/pi.2019.0289.CrossRefGoogle ScholarPubMed
Phillips, ML, Chase, HW, Sheline, YI, Etkin, A, Almeida, JR, Deckersbach, T and Trivedi, MH (2015) Identifying predictors, moderators, and mediators of antidepressant response in major depressive disorder: neuroimaging approaches. American Journal of Psychiatry 172(2), 124–138. https://doi.org/10.1176/appi.ajp.2014.14010076.CrossRefGoogle ScholarPubMed
Poirot, MG, Ruhe, HG, Mutsaerts, HMM, Maximov, II, Groote, IR, Bjørnerud, A, Marquering, HA, Reneman, L and Caan, MWA (2024) Treatment response prediction in major depressive disorder using multimodal MRI and clinical data: Secondary analysis of a randomized clinical trial. American Journal of Psychiatry 181(3), 223–233. https://doi.org/10.1176/appi.ajp.20230206.CrossRefGoogle ScholarPubMed
Price, JL and Drevets, WC (2010) Neurocircuitry of mood disorders. Neuropsychopharmacology 35(1), 192–216. https://doi.org/10.1038/npp.2009.104.CrossRefGoogle ScholarPubMed
Ressler, KJ and Mayberg, HS (2007) Targeting abnormal neural circuits in mood and anxiety disorders: from laboratory to the clinic. Nature Neuroscience 10(9), 1116–1124.CrossRefGoogle ScholarPubMed
Rorden, C and Brett, M (2000) Stereotaxic display of brain lesions. Behavioural Neurology 12(4), 191–200. https://doi.org/10.1155/2000/421719.CrossRefGoogle ScholarPubMed
Rost, N, Binder, EB and Bruckl, TM (2022) Predicting treatment outcome in depression: an introduction into current concepts and challenges. European Archives of Psychiatry and Clinical Neuroscience 273(1), 113–127. https://doi.org/10.1007/s00406-022-01418-4.CrossRefGoogle Scholar
Ruhe, HG, Mocking, RJT, Figueroa, CA, Seeverens, PWJ, Ikani, N, Tyborowska, A, Browning, M, Vrijsen, JN, Harmer, CJ and Schene, AH (2019) Emotional biases and recurrence in major depressive disorder. Results of 2.5 years follow-up of drug-free cohort vulnerable for recurrence. Frontiers in Psychiatry 10, 145. https://doi.org/10.3389/fpsyt.2019.00145.CrossRefGoogle ScholarPubMed
Rush, AJ, Trivedi, MH, Ibrahim, HM, Carmody, TJ, Arnow, B, Klein, DN, Markowitz, JC, Ninan, PT, Kornstein, S, Manber, R, Thase, ME, Kocsis, JH and Keller, MB (2003) The 16-item Quick Inventory of Depressive Symptomatology (QIDS), clinician rating (QIDS-C), and self-report (QIDS-SR): a psychometric evaluation in patients with chronic major depression. Biological Psychiatry 54(5), 573–583. https://doi.org/10.1016/s0006-3223(02)01866-8.CrossRefGoogle ScholarPubMed
Rush, AJ, Trivedi, MH, Wisniewski, SR, Nierenberg, AA, Stewart, JW, Warden, D, Niederehe, G, Thase, ME, Lavori, PW, Lebowitz, BD, McGrath, PJ, Rosenbaum, JF, Sackeim, HA, Kupfer, DJ, Luther, J and Fava, M (2006) Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. American Journal of Psychiatry 163(11), 1905–1917. https://doi.org/10.1176/ajp.2006.163.11.1905.CrossRefGoogle ScholarPubMed
Schmaal, L, Marquand, AF, Rhebergen, D, van Tol, MJ, Ruhe, HG, van der Wee, NJ, Veltman, DJ and Penninx, BW (2015) Predicting the naturalistic course of major depressive disorder using clinical and multimodal neuroimaging information: a multivariate pattern recognition study. Biological Psychiatry 78(4), 278–286. https://doi.org/10.1016/j.biopsych.2014.11.018.CrossRefGoogle ScholarPubMed
Schmidt, HD, Shelton, RC and Duman, RS (2011) Functional biomarkers of depression: diagnosis, treatment, and pathophysiology. Neuropsychopharmacology 36(12), 2375–2394. https://doi.org/10.1038/npp.2011.151.CrossRefGoogle ScholarPubMed
Simon, GE and Perlis, RH (2010) Personalized medicine for depression: can we match patients with treatments? American Journal of Psychiatry 167(12), 1445–1455. https://doi.org/10.1176/appi.ajp.2010.09111680.CrossRefGoogle ScholarPubMed
Spitzer, RL, Kroenke, K, Williams, JBW and Lowe, B (2006) A brief measure for assessing generalised anxiety disorder: the GAD-7. Archives of Internal Medicine 166(10), 1092–1097. https://doi.org/10.1001/archinte.166.10.1092.CrossRefGoogle ScholarPubMed
Strawbridge, R, Young, AH and Cleare, AJ (2017) Biomarkers for depression: recent insights, current challenges and future prospects. Neuropsychiatric Disease and Treatment 13, 1245–1262. https://doi.org/10.2147/NDT.S114542.CrossRefGoogle ScholarPubMed
Uher, R, Tansey, KE, Dew, T, Maier, W, Mors, O, Hauser, J, Zvezdana Dernovsek, M, Henigsberg, N, Souery, D, Farmer, A and McGuffin, P (2014) An inflammatory biomarker as a differential predictor of outcome of depression treatment with escitalopram and nortriptyline. American Journal of Psychiatry 171(12), 1278–1286. https://doi.org/10.1176/appi.ajp.2014.14010094.CrossRefGoogle ScholarPubMed
Weiner, B (1985) An attributional theory of achievement motivation and emotion. Psychological Review 92(4), 548–573.CrossRefGoogle ScholarPubMed
Williams, LM, Korgaonkar, MS, Song, YC, Paton, R, Eagles, S, Goldstein-Piekarski, A, Grieve, SM, Harris, AW, Usherwood, T and Etkin, A (2015) Amygdala reactivity to emotional faces in the prediction of general and medication-specific responses to antidepressant treatment in the randomized iSPOT-D trial. Neuropsychopharmacology 40(10), 2398–2408. https://doi.org/10.1038/npp.2015.89.CrossRefGoogle ScholarPubMed
Young, KD, Siegle, GJ, Zotev, V, Phillips, R, Misaki, M, Yuan, H, Drevets, WC and Bodurka, J (2019) Randomized clinical trial of real-time fMRI amygdala neurofeedback for major depressive disorder: Effects on symptoms and autobiographical memory recall. American Journal of Psychiatry 174, 748–755. https://doi.org/10.1176/appi.ajp.2017.16060637.CrossRefGoogle Scholar
Zahn, R, Lythe, KE, Gethin, JA, Green, S, Deakin, JF, Young, AH and Moll, J (2015) The role of self-blame and worthlessness in the psychopathology of major depressive disorder. Journal of Affective Disorders 186, 337–341. https://doi.org/10.1016/j.jad.2015.08.001.CrossRefGoogle ScholarPubMed
Figure 0

Table 1. Baseline demographic and clinical characteristics by responders and non-responders (n = 38)

Figure 1

Table 2. Prediction models of clinical outcomes in depression (n = 38)

Figure 2

Figure 1. Neural signatures of emotional biases associated with clinical outcomes in difficult-to-treat MDD. Three neural signatures of emotional biases were associated with clinical outcomes in UK primary care. More specifically, it shows cropped sections of voxel-based analyses illustrating the respective pre-registered a priori regions-of-interest, i.e. self-blame-selective right superior anterior temporal lobe-posterior subgenual cortex (BA25) connectivity, resting-state functional connectivity between the subgenual cortex and ventrolateral prefrontal cortex/insula, and bilateral amygdala blood-oxygen-level-dependent activation in response to subliminal sad vs happy faces. These cropped sections are displayed using MRIcron at an uncorrected voxel-level threshold of p=.005, with no cluster-size threshold (the colour bar represents t values) and adapted from figures previously published (Fennema et al. 2023; Fennema, Barker, O’Daly, Duan, Carr, et al. 2024; Fennema, Barker, O’Daly, Duan, Godlewska, et al. 2024). A linear model using the pre-registered fMRI measures with baseline Maudsley Modified Patient Health Questionnaire (9 items) as a covariate explained 32% of the variance of QIDS-SR16 percentage change. The red and green lines display the partial effects of the fMRI measures on the variance of QIDS-SR16 percentage change after four months of standard primary care. MDD = major depressive disorder; BA = Brodmann Area; RSATL = right superior anterior temporal lobe; VLPFC = ventrolateral prefrontal cortex; BOLD = blood-oxygen level-dependent; QIDS-SR16 = Quick Inventory of Depressive Symptomatology, self-rated (16 items).

Supplementary material: File

Fennema et al. supplementary material

Fennema et al. supplementary material
Download Fennema et al. supplementary material(File)
File 114.5 KB

Author comment: Neural signatures of emotional biases predict clinical outcomes in difficult-to-treat depression — R0/PR1

Published online by Cambridge University Press:  01 October 2024
DOI: https://doi.org/10.1017/dep.2024.6.pr1 [Opens in a new window]
Roland Zahn Centre of Affective Disorders, King’s College London Institute of Psychiatry Psychology & Neuroscience, London, United Kingdom of Great Britain and Northern Ireland
Revision round: 0
Role: author

Comments

No accompanying comment.

Review: Neural signatures of emotional biases predict clinical outcomes in difficult-to-treat depression — R0/PR2

Published online by Cambridge University Press:  01 October 2024
DOI: https://doi.org/10.1017/dep.2024.6.pr2 [Opens in a new window]
Date of review:  04 June 2024
Role: reviewer

Review: Neural signatures of emotional biases predict clinical outcomes in difficult-to-treat depression — R0/PR3

Published online by Cambridge University Press:  01 October 2024
DOI: https://doi.org/10.1017/dep.2024.6.pr3 [Opens in a new window]
Alec Jamieson [Opens in a new window] Psychiatry, The University of Melbourne, Carlton, Victoria, Australia
Date of review:  05 August 2024
Revision round: 0
Role: reviewer
Recommendation/decision: major-revision

Comments

The manuscript by Fennema and colleagues investigated the utility of neural markers in predicting treatment outcomes for difficult-to-treat major depressive disorder (MDD). Specifically, they identified that in contrast to baseline symptom severity, which only explained 3% of outcome variance, three specific neural measures across three separate tasks explained 32%. These findings highlight the potential of these neural signatures as predictors of clinical outcomes in chronic, difficult-to-treat MDD. Despite the relatively small sample size and that individually these effects have been previously published, I do believe that there is some novelty in the approach taken in this study.Attached are my concerns in detail, I hope the authors find these comments helpful.

Introduction

The introduction is very short and does not adequately introduce the background for the topic, particularly for a journal which does not specialise in neuroimaging. While the study’s pre-registered regions of interest are a strength, engaging in wider and more recent literature concerning general prognosis or prediction of treatment outcome for SSRIs would be useful.

Line 52 “Despite these promising findings, their reproducibility has not been established and it is unclear whether these neural signatures generalise to pragmatic samples of patients encountered in clinical settings.” I don’t believe that this is the primary issue that this paper is set up to answer (given that apart from the amygdala the individual markers don’t replicate). It appears to be more tailored to examine whether together these parameters provide greater utility than was observed in the individual studies.

Methods

It wasn’t immediately clear from the manuscript that the descriptions of your tasks were located in the supplementary materials. A brief description of the tasks in the main manuscript and reference to the full description in the supplementary materials is warranted.

The pre-registration report mentions a number of additional regions of interest including functional connectivity between SCC - VMPFC and SCC - dorsal midbrain, and pregenual ACC activity for the implicit face processing task as a region of interest which do not appear in this study. Why were these not included in the analysis model? Also, it highlights that a logistic regression will be used to predicted binarised response/non-response. While I think that symptom change is a more useful measure, I believe that examining the binarised outcome as well would also be informative to readers (and consistent with your power analysis).

“baseline Maudsley Modified Patient Health Questionnaire, 9 items 101 (MM-PHQ-9; measure of severity of depressive symptoms) (Harrison et al. 2021) as a covariate”. Why wasn’t baseline QIDS used as the covariate here? Surely this would do a better job of capturing the same depressive symptoms at baseline.

Results

An additional table examining any baseline differences between responders/non-responders at baseline would be useful in identifying whether any factors were confounding the prediction results.

Discussion

The discussion does not adequately articulate this study’s addition to the literature and the author’s interpretation of the findings. Given that they have used measures of activity and connectivity which have been previously published, it is important to emphasise the novelty of this paper in combining these measures. It is also important, given the non-specific nature of the treatment being applied and the general prognostic effects identified, that the hypothesised clinical utility of such findings are explicitly stated and interpreted (e.g. identifying a prognostic marker no matter how good does not eliminate the trail-and-error nature of prescribing antidepressants).

Furthermore, it is important for the authors to highlight how they suggest to improve the explained variance in future studies. While 32% is much better than the amount given by the clinical variables alone, it is unlikely sufficient for translation given the costs associated with running three different scans.

Line 174 “This offers the intriguing possibility of stratification for neuromodulation and neurofeedback studies based on distinct neural circuits of interest, by either modulating self-blaming or emotional perception biases in patients non-responsive to standard treatments”. While I agree with this statement generally,the fact that the measures were uncorrelated isn’t evidence for this point. It is possible that all these features independently predict the same underlying general pathophysiology. Without testing whether these features relate to specific symptoms this point is therefore difficult to disentangle.

Minor notes

Line 36 “to respond more strongly” please be specific to what you mean by strongly here.

Your reference manager appears to have formatted the in-text citations strangely.

Presentation

Overall score 3 out of 5
Is the article written in clear and proper English? (30%)
4 out of 5
Is the data presented in the most useful manner? (40%)
3 out of 5
Does the paper cite relevant and related articles appropriately? (30%)
3 out of 5

Context

Overall score 3 out of 5
Does the title suitably represent the article? (25%)
4 out of 5
Does the abstract correctly embody the content of the article? (25%)
4 out of 5
Does the introduction give appropriate context and indicate the relevance of the results to the question or hypothesis under consideration? (25%)
2 out of 5
Is the objective of the experiment clearly defined? (25%)
2 out of 5

Results

Overall score 3 out of 5
Is sufficient detail provided to allow replication of the study? (50%)
3 out of 5
Are the limitations of the experiment as well as the contributions of the results clearly outlined? (50%)
3 out of 5

Review: Neural signatures of emotional biases predict clinical outcomes in difficult-to-treat depression — R0/PR4

Published online by Cambridge University Press:  01 October 2024
DOI: https://doi.org/10.1017/dep.2024.6.pr4 [Opens in a new window]
Ian Hickie [Opens in a new window] University of Sydney CAR, Glebe, New South Wales, Australia
Date of review:  06 August 2024
Revision round: 0
Role: reviewer
Recommendation/decision: minor-revision

Comments

1. Justify the use of %change in self-reported QIDS as main clinical outcome - %change has known limitations in outcomes research compared with other absolute change measures, and thresholds for absolute improvement; As part of this, additional reporting of the distribution of absolute changes in QIDS scores would be helpful in the actual text and not just supplementary tables.

2. Some further elaboration of the proposed (structural or functional) circuitry delineated here – not just in terms of the propsed ‘cogntive aspects (‘self-blame’, ‘negative perception’) but their actual anatomical or physiological characteristics and the extent to which they related to circuits proposed by others (notably Williams et al) for predicting depression outcomes on medication.

Presentation

Overall score 3 out of 5
Is the article written in clear and proper English? (30%)
4 out of 5
Is the data presented in the most useful manner? (40%)
3 out of 5
Does the paper cite relevant and related articles appropriately? (30%)
3 out of 5

Context

Overall score 3 out of 5
Does the title suitably represent the article? (25%)
4 out of 5
Does the abstract correctly embody the content of the article? (25%)
4 out of 5
Does the introduction give appropriate context and indicate the relevance of the results to the question or hypothesis under consideration? (25%)
2 out of 5
Is the objective of the experiment clearly defined? (25%)
2 out of 5

Results

Overall score 4 out of 5
Is sufficient detail provided to allow replication of the study? (50%)
4 out of 5
Are the limitations of the experiment as well as the contributions of the results clearly outlined? (50%)
4 out of 5

Decision: Neural signatures of emotional biases predict clinical outcomes in difficult-to-treat depression — R0/PR5

Published online by Cambridge University Press:  01 October 2024
DOI: https://doi.org/10.1017/dep.2024.6.pr5 [Opens in a new window]
Ian Hickie [Opens in a new window] University of Sydney CAR, Glebe, New South Wales, Australia
Revision round: 0
Role: Editor in Chief
Recommendation/decision: minor-revision

Comments

No accompanying comment.

Author comment: Neural signatures of emotional biases predict clinical outcomes in difficult-to-treat depression — R1/PR6

Published online by Cambridge University Press:  01 October 2024
DOI: https://doi.org/10.1017/dep.2024.6.pr6 [Opens in a new window]
Roland Zahn Centre of Affective Disorders, King’s College London Institute of Psychiatry Psychology & Neuroscience, London, United Kingdom of Great Britain and Northern Ireland
Revision round: 1
Role: author

Comments

No accompanying comment.

Decision: Neural signatures of emotional biases predict clinical outcomes in difficult-to-treat depression — R1/PR7

Published online by Cambridge University Press:  01 October 2024
DOI: https://doi.org/10.1017/dep.2024.6.pr7 [Opens in a new window]
Ian Hickie [Opens in a new window] University of Sydney CAR, Glebe, New South Wales, Australia
Revision round: 1
Role: Editor in Chief
Recommendation/decision: accept

Comments

The authors have provided a thoughtful revision of the manuscript and have appropriately addressed the reviewers’s comments. I am happy to recommend that this manuscript be Accepted.