- Email: [email protected]
Ventral attention-network effective connectivity predicts individual differences in adolescent depression
Journal of Affective Disorders 252 (2019) 55–59
Contents lists available at ScienceDirect
Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad
Research paper
Ventral attention-network effective connectivity predicts individual differences in adolescent depression ⁎
Jie Liua,b,c, Pengfei Xua,c, Jingyuan Zhanga, Nengzhi Jiangb,d, Xinying Lib,d, , Yuejia Luoa,e,f,
T ⁎⁎
a
College of Psychology and Sociology, Shenzhen University, Shenzhen, China Key Laboratory of Mental Health, Institute of Psychology, Chinese Academy of Sciences, Beijing, China c Center for Brain Disorders and Cognitive Neuroscience, Shenzhen University, Shenzhen, China d Department of Psychology, University of Chinese Academy of Sciences, Beijing, China e Centre for Emotion and Brain, Shenzhen Institute of Neuroscience, Shenzhen, China f Department of Psychology, Southern Medical University, Guangzhou, China b
A R T I C LE I N FO
A B S T R A C T
Keywords: Ventral attention network Depression Attention bias Resting-state fMRI Stochastic dynamic causal modelling
Background: Stimulus-driven negative attention bias is a central deficit in depression and might play an important role in vulnerability to depression Adolescents are susceptible to depression. Thus, investigating the neural correlates of attention bias in adolescents is a critical step for identifying neural markers of early onset of depression. Previous studies have shown that the ventral attention network (VAN), which includes bilateral ventrolateral prefrontal cortex (VLPFC) and bilateral temporal-parietal junction (TPJ), is the key brain network for stimulus-driven attention. However, the relationship between depression and effective connectivity within the VAN in adolescents is poorly understood. Method: We employed resting-state fMRI to assess the relationship between directional effective connectivity within the VAN and depression scores in 216 healthy adolescents. Results: Using stochastic dynamic modeling, we found that individuals who exhibited higher self-reported depression showed stronger effective connectivity between right VLPFC and left TPJ within the VAN. Limitation: The level of depression in this study was assessed with self-reported questionnaire. This measure might be more influenced by current mood in adolescents than that in adults. Future studies should emplo more objective measures to index levels of depression. Conclusions: Our findings indicate that effective connectivity between right VLPFC and left TPJ could at least partially serve as a biomarker for bottom-up processing of depression in adolescents.
1. Introduction Depression is a prevalent and pervasive mental health disorder that is often not optimally treated (Disner et al., 2011). Adolescents are susceptible to depression, with a prevalence that can reach up to 14% in adolescents (Kessler andWalters, 1998). Adolescent-onset depression is associated with greater severity of symptoms, likelihood of relapse, and suicidal tendencies throughout life than adult-onset depression (Berndt et al., 2000; Zisook et al., 2007; Hollon et al., 2006). A central cognitive impairment that occurs in depression is the inability to allocate attention to appropriate emotional cues (Gotlib et al., 2004; Peckham et al., 2010). Adults with depression showed increased attention for negative stimuli and decreased attention for positive stimuli compared with nondepressed individuals (e.g., Eizenman et al., 2003; Kellough et al.,
⁎
2008). Moreover, although the speed with which adults with depression orient to sad stimuli is comparable to that of non-depressed individuals (Cisler and Koster 2010; Teachman et al., 2012), the amount of time they spend paying attention to them is longer by several seconds (Peckham et al., 2010). Similar attention bias can also be observed in adolescents with depression (e.g., Sylvester et al., 2016). The ventral attention network (VAN) is associated with the orientation of stimulus-driven attention, and its key brain regions are the bilateral ventrolateral prefrontal cortex (VLPFC) and bilateral temporalparietal junction (TPJ) (Corbetta et al., 2008). Using resting-state functional magnetic resonance imaging (rs-fMRI), studies have demonstrated that depression in adults was associated with abnormal VAN function. Meta-analysis of resting-state functional connectivity in major depressive disorder (MDD) suggested that MDD was related to
Corresponding author at: Key Laboratory of Mental Health, Institute of Psychology, Chinese Academy of Science, Beijing, China. Corresponding author at: Center for Brain Disorders and Cognitive Neuroscience, Shenzhen University, Shenzhen, China. E-mail addresses: [email protected] (X. Li), [email protected] (Y. Luo).
⁎⁎
https://doi.org/10.1016/j.jad.2019.04.033 Received 18 August 2018; Received in revised form 8 March 2019; Accepted 7 April 2019 Available online 08 April 2019 0165-0327/ © 2019 Published by Elsevier B.V.
Journal of Affective Disorders 252 (2019) 55–59
J. Liu, et al.
2.4. Image preprocessing
hypoconnectivity between the VAN seeds and the precuneus, extending to the occipital and posterior cingulate cortex (Kaiser et al., 2015). Another rs-fMRI meta-analysis indicated lower activation in the TPJ but increased activation in the VLPFC in patients with depression than in healthy controls (Fitzgerald et al., 2008). Because the brain changes significantly as adolescents mature, brain abnormalities in adolescent depression could be different from those in adults. Investigating the neurobiology of the VAN during the early stages of the disease is thus critical for understanding the etiology of depression. However, few studies have investigated the neurobiology of the VAN in adolescents. Moreover, directional effective connectivity within the VAN is poorly understood. In the current study, we adopted stochastic dynamic causal modeling (sDCM) for rs-fMRI in a large sample of adolescents to estimate the effective connectivity between regions within the VAN. The sDCM approach allows endogenous or random fluctuations in hidden states, such as neuronal activity, or hemodynamic states, such as local perfusion and deoxyhemoglobin content, and is this suitable for identifying the casual and directed nature of coupling between intrinsic modes of brain activity (Daunizeau et al., 2009; Riera et al., 2004). We applied sDCM to characterize the within-network effective connectivity of the VAN in healthy adolescents and to determine how it is related to depression (non-clinical). Given that previous task-fMRI studies showed both lower activity in the VLPFC and superior parietal lobule in adults with depression who were required to shift their attentional focus away from negative stimuli (Beevers et al., 2010; Fales et al., 2008), we hypothesized that individuals with higher depression scores might exhibit greater within-VAN effective connectivity.
Data preprocessing was performed using DPARSF software (Yan and Zang, 2010, http://www.restfmri.net). The T1 images were segmented into gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) partitions. The functional images were preprocessed with the following steps: 1) deletion of the first 10 volumes; 2) correction of slice timing; 3) assignment of each T1 image to a functional image; 4) normalization to Montreal Neurological Institute (MNI) space with solution of 3 × 3 × 3 mm; 5) spatial smoothing by convolution with an isotropic Gaussian kernel (FWHM = 6 mm); 6) linear trend removing; 7) bandpass (0.01–0.1 Hz) filtering; and 8) data scrubbing. The scrubbing procedure was performed following Power et al. (2012). We excluded any volume whose frame-wise dependent (FD) value exceeded 0.2 mm, along with the previous volume and the two following volumes. Fiftythree participants were excluded from the subsequent analysis because less than 150 of their volumes were left after the scrubbing procedure. The mean FD of the remaining participants was 0.07 ( ± 0.03). Further nuisance variables for rs-fMRI measures, such as motion parameters, the signal averaged over the individual segmented CSF and white matter (WM) regions, as well as the global signal (Murphy and Fox, 2017), were regressed out with a general linear model at the whole-brain level before the volume of interest extraction (see details in “Selection and extraction of volumes of interest”). 2.5. Selection and extraction of volumes of interest (VOI) The locations of the key cortical regions in the VAN were identified with spatial independent component analysis (ICA). The preprocessed images were analyzed using the Group ICA of the fMRI toolbox (GIFT) software package (Medical Image Analysis Lab, University of New Mexico, Albuquerque, New Mexico, USA) with the infomax Algorithm (Calhoun and Adali, 2004). Using spatial ICA, we separated spatially independent patterns from the time courses and generated 41 independent components for each participant. To identify the components that corresponded to parts of the VAN, spatial regression was performed to match with a pre-existing VAN template (Yeo et al., 2011). Components with the highest spatial regression coefficient between the 41-group mean component maps and the VAN template were selected as the network functional templates for VOI selection (seed regions for the following effective connectivity analysis). Four brain region masks, including bilateral TPJ and bilateral inferior frontal gyrus, were separately exacted from this VAN template. The VOIs were centered on the local maximum in the group mean map of the corresponding component within the four brain region masks mentioned above. For instance, the VLPFC VOI was selected based on the spatial regression of the ICA that identified a maximum ventral attention activation component within the VLPFC brain mask. Before dynamic casual modeling, we built a general liner model and then extracted the time-series data from the results. The general linear model included nuisance variables for rs-fMRI measures, such as 24 motion parameters (6 head-motion parameters, 6 head-motion parameters one time point before, and the 12 corresponding squared items), the signal averaged over the individual segmented CSF and white matter (WM) regions, as well as the global signal (Yan et al., 2013). For each participant, the principal eigenvariate of the VOI (the first principal component of the local multivariate time series over all voxels in the VOI) was defined as the seed of the models (radius = 6 mm).
2. Methods 2.1. Participants Two hundred and sixteen healthy adolescents (15.72 ± 0.94 years, 112 males) were recruited from the local community. All participants were healthy individuals without brain injuries or neurological diseases. This study was approved by the Institutional Review Board of the Institute of Psychology of the Chinese Academy of Sciences and the Institutional Review Board of Beijing MRI Center for Brain Research. 2.2. Questionnaires Participant self-reported depression levels were assessed by the Chinese version of the Children's Depression Inventory (CDI) (Kovacs, 1985). The Cronbach alpha was 0.85 for CDI. As depression shares some negative emotion characteristics with anxiety, to ensure this did not confound the issue, we also included the Chinese version of the Spielberger's Trait Anxiety Inventory (STAI-T) (Spielberger et al., 1970), which uses 20 items to measure self-reported anxiety levels. The Cronbach alpha was 0.87 for STAI-T in the current study. 2.3. MRI data acquisition All participants completed a six-minute rs-fMRI scan on a 3-Tesla SIEMENS MRI scanner (Beijing, China). Whole-brain T2*-weighted blood oxygenation level-dependent (BOLD) contrast images were acquired with single-shot gradient-recalled echo planar imaging (GREPI) sequences (repetition time = 2000 ms, echo time = 30 ms, flip angle = 90°, acquisition matrix = 64 × 64, field of view = 22 cm, slice thickness = 3 mm, alignment = AC-PC line). High-resolution T1weighted images were also acquired for each participant using a magnetization-prepared rapid gradient echo (MPRAGE) sequence (repetition time = 2530 ms, echo time = 3.37 ms, acquisition matrix = 256 × 192, field of view = 25.6 × 19.2 cm, flip angle = 7°, slice thickness = 1.33 mm, alignment = AC-PC line).
2.6. Specification and estimation of the sDCM at the first level The effective connectivity within the VAN was analyzed using sDCM implemented in SPM12 (version 6906, Wellcome Department of Imaging Neurosciences, University College London, UK, http://www. fil.ion.ucl.ac.uk/spm). The mathematical model of the underlying 56
Journal of Affective Disorders 252 (2019) 55–59
J. Liu, et al.
neuronal connectivity among DCM nodes is a system of bilinear differential state equations with coefficients specified by three matrices (A matrix, B matrix, and C matrix) (Friston et al., 2003). sDCM is different from conventional deterministic DCM. It seeks to improve model estimation by modeling endogenous or random fluctuations in unobserved (hidden) neuronal and physiological states in the differential equations of the neuronal states (Li et al., 2011a). sDCM can be used to explore the effectivity functional connectivity of rs-fMRI data. A full reciprocally connected model (full model) between each of the four seed regions, as well as the self-connections of each of the four seed regions, was constructed for each participant. Generalized filtering was then used for model inversion and parameter estimation. 2.7. Second level analysis of the sDCM Group level post-hoc optimization was conducted by selecting all estimated models. A Bayesian network discovery scheme was applied to search for the best model in the model space (Friston et al., 2011). Evidence of all the possible reduced models in the model space was obtained as indicators of the best model after network discovery. The model with the highest model evidence was selected as the winning model. Bayesian parameter averaging (BPA) was performed over the participants to determine the group-level optimal sparse model. The endogenous connectivity in DCMs is quantified by the “A” matrix parameters, which measure the effective connectivity strength between nodes. Thus, the endogenous connection strength between each node in the winning model was extracted in the DCM “A” matrix and used for the subsequent analysis for each participant. We performed correlation analysis between the connection strengths obtained in the DCM analysis and the depression scores. Statistical analysis was carried out using SPSS (IBM Corp., version 20. Armonk, NY, USA). Because depression shares some negative emotion characteristics with trait anxiety (r = 0.82, p < 0.001 in our data sample), to investigate the specific effect for depression, partial correlation analysis was adopted to control for anxiety. In addition, we also took demographic variables such as sex, age, and mean FD as covariates in the partial correlation analysis.
In the VAN, all 16 connections within the full model were significantly different from zero. After controlling for sex, age, mean FD, and the trait anxiety scores, the positive correlations were significant or marginally significant for the following connections: right VLPFC → left TPJ (r = 0.24, Bonferroni corrected p = 0.03), left TPJ → right VLPFC (r = 0.22, Bonferroni corrected p = 0.08) (Table 1). The partial correlation between trait anxiety and any effective connectivity within the VAN was not significant, after controlling for sex, age, mean FD, and the CDI scores: right VLPFC → left TPJ (r = −0.19, Bonferroni corrected p = 0.32), left TPJ → right VLPFC (r = −0.21, Bonferroni corrected p = 0.14).
2.8. Validation analysis
3.4. Validation results
Previous studies have suggested that global signals might introduce widespread negative functional connectivity and may alter the intrinsic correlation structure of brain networks (Ciric et al., 2017; Murphy et al., 2009; Weissenbacher et al., 2009). Thus, we repeated our general linear model and sDCM analysis without global signal regression.
The results of validation analysis that processed the data without global signal regression remained consistent with the main results. After controlling for sex, age, trait anxiety, and individual mean FD, the effective connectivity from the right VLPFC to the left TPJ could significantly predict the scores from the depression questionnaire
Fig. 1. ICA components selected by spatial regression.
connections of the four VOIs. The evidence showed that the best model was the full model (posterior probability > 0.99). The BPA result showed that all connections in the model were significantly greater or less than zero. 3.3. Within network effective connectivity and depression
3. Results
Table 1 Correlation between within network effective connectivity and depression.
3.1. Group ICA and selected VOIs Forty-one independent components were obtained from the group ICA analysis. We chose the intrinsic VAN template from Yeo et al. (2011). We then performed spatial regression with the template to identify the network-related component. After spatial sorting, we selected the top rank-ordered components with the highest spatial regression coefficient as the network-related component (regression coefficient = 0.18). The brain template and the corresponding selected component are presented in Fig. 1. The peak activation within the selected component and each brain region were considered the volumes of interest (VOIs). The result was two VOIs, bilateral VLPFC (MNI coordinates [−27, 36, −12] and [48, 42, 3]) and bilateral TPJ ([−60, −27, 18] and [63, −24, 27]).
Brain network
Effective connectivity
Partial correlation r p
VAN
Left TPJ –> left TPJ Right TPJ–> left TPJ Left VLPFC –> left TPJ Right VLPFC –> left TPJ Left TPJ–> right TPJ Right TPJ –> right TPJ Left VLPFC –>right TPJ Right VLPFC –> right TPJ Left TPJ –> left VLPFC Right TPJ –> left VLPFC Left VLPFC –> left VLPFC Right VLPFC –> left VLPFC Left TPJ –> right VLPFC Right TPJ –> right VLPFC Left VLPFC –> right VLPFC Right VLPFC –> right VLPFC
0.003 −0.021 0.193 0.240* −0.073 −0.001 −0.090 0.011 0.149 −0.037 0.014 0.046 0.224 0.087 0.009 0.058
3.2. sDCM ⁎
In the VANs, we searched all possible models generated from the 57
Bonferroni corrected p < 0.05.
0.972 0.794 0.015 0.002 0.359 0.989 0.259 0.895 0.060 0.645 0.856 0.568 0.005 0.276 0.911 0.468
Journal of Affective Disorders 252 (2019) 55–59
J. Liu, et al.
depression. They directed measured the attention bias to negative stimulus, and found positive correlation between threat bias and the VAN RSFC (between right VLPFC and right TPJ) for both the children with depression history and health control, which is in accordance with our result that showed higher level of depression associated with greater effective connectivity within VAN. However, they reported children with a history of depression had significantly reduced resting-state functional connectivity among the right regions comprising the VAN, and negative results was found among the left regions comprising the VAN. The difference might arise from the different experimental samples. In Sylvester et al. (2013), children with a history of depression or anxiety in this study had nonsignificant lower threat bias (i.e., threat avoidance) compared to subjects with no history of psychiatric illness. A general mechanism could be that lower threat bias was related reduced RSFC within the VAN. Thus children with histories of anxiety/ depression were observed decreased VAN resting state connectivity in Sylvester et al. (2013), but children with higher current depression might have the increased effective connectivity in the current study. During adolescence, the neurodevelopmental trajectory largely alters the brain, and the current results indicate that for some individuals, brain response patterns to unexpected stimuli might be prone to evolve into attention bias to negative stimulus, and these people might be at higher risk of depression. Overall, using a stochastic dynamic causal model, we discovered that depression scores were correlated to resting-state directional effective connectivity within the ventral attention networks of healthy adolescents. The significant positive correlation found in effective connectivity within the ventral attention network might serve as the neural basis for attention bias in depression. These findings might have clinical implications for treatment interventions, and suggest that monitoring neural changes within bottom-up processing networks might be useful.
(r = 0.25, Bonferroni corrected p = 0.03). The correlation between effective connectivity from the left TPJ to the right VLPFC and to CDI scores was marginally significant (r = 0.22, Bonferroni Corrected p = 0.08). The partial correlation between trait anxiety and any effective connectivity within the VAN was not significant, after controlling for sex, age, mean FD, and the CDI scores: right VLPFC → left TPJ (r = −0.19, Bonferroni corrected p = 0.24), left TPJ → right VLPFC (r = −0.20, Bonferroni corrected p = 0.22). 4. Discussion Although several studies have explored the neurobiology of the VAN, the key brain network for attention bias, in the depression of adults, few studies have investigated the neurobiology of the VAN in adolescents, especially the effective connectivity within the VAN. We investigated the within-VAN effective connectivity that is assumed to be related to stimulus-driven attention in a large sample of adolescents. We observed a significant positive correlation between depression scores and the within-network effective functional connectivity. Specifically, higher depression scores were related with increased bidirectional connection between the VLPFC and the TPJ. To the best of our knowledge, this is the first study to evaluate within-network directional effective connectivity related to the propensity for depression in healthy adolescents. Behavioral studies have shown that depressed adults (e.g., Joormann and Gotlib, 2007; Caseras et al., 2007) and adolescents (Sylvester et al., 2016) can be characterized by problems in disengagement from negative stimuli. Depressed individuals may not automatically orient their attention toward negative information in the environment, but once such information is noticed, they have greater difficulty disengaging from it. A series of neuroimaging results have suggested that this attentional bias is related to a neural mechanism underlying depression. For instance, this kind of negative attention bias is associated with several abnormal neural responses, such as decreased activation in the TPJ (rs-fMRI study; Fitzgerald et al., 2008), decreased activity in the right VLPFC and right superior parietal cortex (task-fMRI studies; Beevers et al., 2010; Fales et al., 2008), as well as increased magnitude of specific early P300, a characteristic event-related electrical potential seen on electroencephalograms (EEGs) (Bruder et al., 2002; Li et al., 2011b). Our finding indicates a role for the VAN in adolescents that was previously unknown. The increased positive correlation between left TPJ and right VLPFC for adolescents who reported higher depression scores might be due to the co-decrease of activity in these two brain regions at rest in participants who report more depressive symptoms. Given that the VLPFC has been shown to be related to stimulus-driven attention processing in both adults (Asplund et al., 2010) and adolescents (White et al., 2016), and that the TPJ has been reported to be related to post-perceptual processes involved in contextual updating and adjustments of top-down expectations (Geng and Vossel, 2013), we infer that during negative stimulus-driven attention, unexpected stimuli might lead to abnormally decreased right VLPFC activation in adolescents with higher risk of depression due to negative attention bias. These individuals tend to fixate on the negative stimulus. It might be harder for them to update/disengage their attention according to contextual information and top-down expectations, which would lead to decreased activation in the left TPJ. Thus, the effective connectivity from the right VLPFC to the left TPJ might be increased for more depressive adolescents. The effective connectivity from the left TPJ to the right VLPFC was probably the feedback signal from the right VLPFC→ left TPJ connection. The feedback signal might co-increase, but it would be relatively weaker because of signal attenuation. Thus, the effective connectivity from the left TPJ to the right VLPFC might also increase in more depressive adolescents. Sylvester et al. (2013) investigated the resting-state functional connectivity of the VAN in children aged 8–12 years with a history of
Conflict of interest We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work, there is no professional or other personal interest of any nature or kind in any product, service and/or company that could be construed as influencing the position presented in, or the review of, the manuscript entitled. Role of the funding source The research conducted is supported by The National Key Basic Research Program of China (Project no. 2014CB846100) and The Natural Science Foundation of China (Project nos. 31530031, 31700977). The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding source. Acknowledgments We thank the National Key Basic Research Program of China 2014CB846100, as well as the Natural Science Foundation of China (Project nos. 31530031, 31700977) for the financial support. Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.jad.2019.04.033. References Asplund, C.L., Todd, J.J., Snyder, A.P., Marois, R., 2010. A central role for the lateral prefrontal cortex in goal-directed and stimulus-driven attention. Nature Neurosci. 13 (4), 507.
58
Journal of Affective Disorders 252 (2019) 55–59
J. Liu, et al.
Kessler, R.C., Walters, E.E., 1998. Epidemiology of DSM-III-R major depression and minor depressionamong adolescents and young adults in the National Comorbidity Survey. Depress. Anxiety 7 (1), 3–14. Kovacs, M., 1985. The children's depression inventory (CDI). Psychopharmacol. Bull. 21, 995–998. Li, B., Daunizeau, J., Stephan, K.E., Penny, W., Hu, D., Friston, K., 2011a. Generalised filtering and stochastic DCM for fMRI. Neuroimage 58 (2), 442–457. Li, Y., Hu, Y., Liu, T., Wu, D., 2011b. Dipole source analysis of auditory P300 response in depressive and anxiety disorders. Cogn. Neurodyn. 5 (2), 221–229. Murphy, K., Birn, R.M., Handwerker, D.A., Jones, T.B., Bandettini, P.A., 2009. The impact of global signal regression on resting state correlations: are anti-correlated networks introduced? Neuroimage 44 (3), 893–905. Murphy, K., Fox, M.D., 2017. Towards a consensus regarding global signal regression for resting state functional connectivity MRI. Neuroimage 154, 169–173. Peckham, A.D., McHugh, R.K., Otto, M.W., 2010. A meta analysis of the magnitude of biased attention in depression. Depress. Anxiety 27 (12), 1135–1142. Power, J.D., Barnes, K.A., Snyder, A.Z., Schlaggar, B.L., Petersen, S.E., 2012. Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion. Neuroimage 59 (3), 2142–2154. Riera, J.J., Watanabe, J., Kazuki, I., Naoki, M., Aubert, E., Ozaki, T., Kawashima, R., 2004. A state-space model of the hemodynamic approach: nonlinear filtering of BOLD signals. NeuroImage 21 (2), 547–567. Spielberger, C.D., Gorsuch, R.L., Lushene, R.E., 1970. Manual for the State-Trait Anxiety Inventory. Sylvester, C.M., Barch, D.M., Corbetta, M., Power, J.D., Schlaggar, B.L., Luby, J.L., 2013. Resting state functional connectivity of the ventral attention network in children with a history of depression or anxiety. J. Am. Acad. Child Adolesc. Psychiatry 52 (12), 1326–1336. Sylvester, C.M., Hudziak, J.J., Gaffrey, M.S., Barch, D.M., Luby, J.L., 2016. Stimulusdriven attention, threat bias, and sad bias in youth with a history of an anxiety disorder or depression. J. Abnorm. Child Psychol. 44 (2), 219–231. Teachman, B.A., Joormann, J., Steinman, S.A., Gotlib, I.H., 2012. Automaticity in anxiety disorders and major depressive disorder. Clin. Psychol. Rev. 32 (6), 575–603. Weissenbacher, A., Kasess, C., Gerstl, F., Lanzenberger, R., Moser, E., Windischberger, C., 2009. Correlations and anticorrelations in resting-state functional connectivity MRI: a quantitative comparison of preprocessing strategies. Neuroimage 47 (4), 1408–1416. White, L.K., Britton, J.C., Sequeira, S., Ronkin, E.G., Chen, G., Bar-Haim, Y., ... Pine, D.S., 2016. Behavioral and neural stability of attention bias to threat in healthy adolescents. Neuroimage 136, 84–93. Yan, C., Zang, Y., 2010. DPARSF: a MATLAB toolbox for" pipeline" data analysis of resting-state fMRI. Front. Syst. Neurosci. 4, 13. Yan, C.G., Cheung, B., Kelly, C., Colcombe, S., Craddock, R.C., Di Martino, A., ... Milham, M.P., 2013. A comprehensive assessment of regional variation in the impact of head micromovements on functional connectomics. Neuroimage 76, 183–201. Yeo, Thomas, T., B., Krienen, F.M., Sepulcre, J., Sabuncu, M.R., Lashkari, D., Hollinshead, M., …, Buckner, R.L., 2011. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J. Neurophysiol. 106 (3), 1125–1165. https:// doi.org/10.1152/jn.00338.2011. Zisook, S., Lesser, I., Stewart, J.W., Wisniewski, S.R., Balasubramani, G.K., Fava, M., ... Trivedi, M.H., 2007. Effect of age at onset on the course of major depressive disxsorder. Am. J. Psychiatry 164 (10), 1539–1546.
Beevers, C.G., Clasen, P., Stice, E., Schnyer, D., 2010. Depression symptoms and cognitive control of emotion cues: a functional magnetic resonance imaging study. Neuroscience 167 (1), 97–103. Berndt, E.R., Koran, L.M., Finkelstein, S.N., Gelenberg, A.J., Kornstein, S.G., Miller, I.M., Thase, M.E., Trapp, G.A., Keller, M.B., 2000. Lost human capital from early-onset chronic depression. Am. J. Psychiatry 157 (6), 940–947. Bruder, G.E., Kayser, J., Tenke, C.E., Leite, P., Schneier, F.R., Stewart, J.W., Quitkin, F.M., 2002. Cognitive ERPs in depressive and anxiety disorders during tonal and phonetic oddball tasks. Clin. Electroencephalogr. 33 (3), 119–124. Calhoun, V.D., Adali, T., 2004. Group ICA of fMRI Toolbox (GIFT). Online at http://icatb. sourceforge. net. Caseras, X., Garner, M., Bradley, B.P., Mogg, K., 2007. Biases in visual orienting to negative and positive scenes in dysphoria: an eye movement study. J. Abnorm. Psychol. 116 (3), 491. Ciric, R., Wolf, D.H., Power, J.D., Roalf, D.R., Baum, G.L., Ruparel, K., ... Gur, R.C., 2017. Benchmarking of participant-level confound regression strategies for the control of motion artifact in studies of functional connectivity. Neuroimage 154, 174–187. Cisler, J.M., Koster, E.H., 2010. Mechanisms of attentional biases towards threat in anxiety disorders: an integrative review. Clin. Psychol. Rev. 30 (2), 203–216. Corbetta, M., Patel, G., Shulman, G.L., 2008. The reorienting system of the human brain: from environment to theory of mind. Neuron 58 (3), 306–324. Daunizeau, J., Kiebel, S.J., Friston, K.J., 2009. Dynamic causal modelling of distributed electromagnetic responses. NeuroImage 47 (2), 590–601. Disner, S.G., Beevers, C.G., Haigh, E.A., Beck, A.T., 2011. Neural mechanisms of the cognitive model of depression. Nat. Rev. Neurosci. 12 (8), 467. Eizenman, M., Lawrence, H.Y., Grupp, L., Eizenman, E., Ellenbogen, M., Gemar, M., Levitan, R.D., 2003. A naturalistic visual scanning approach to assess selective attention in major depressive disorder. Psychiatry Res. 118 (2), 117–128. Fales, C.L., Barch, D.M., Rundle, M.M., Mintun, M.A., Snyder, A.Z., Cohen, J.D., ... Sheline, Y.I., 2008. Altered emotional interference processing in affective and cognitive-control brain circuitry in major depression. Biol. Psychiatry 63 (4), 377–384. Fitzgerald, P.B., Laird, A.R., Maller, J., Daskalakis, Z.J., 2008. A meta-analytic study of changes in brain activation in depression. Human Brain Mapp. 29 (6), 683–695. Friston, K.J., Harrison, L., Penny, W., 2003. Dynamic causal modelling. Neuroimage 19 (4), 1273–1302. Friston, K.J., Li, B., Daunizeau, J., Stephan, K.E., 2011. Network discovery with DCM. Neuroimage 56 (3), 1202–1221. Geng, J.J., Vossel, S., 2013. Re-evaluating the role of TPJ in attentional control: contextual updating? Neurosci. Biobehav. Rev. 37 (10), 2608–2620. Gotlib, I.H., Krasnoperova, E., Yue, D.N., Joormann, J., 2004. Attentional biases for negative interpersonal stimuli in clinical depression. J. Abnorm. Psychol. 113 (1), 127. Hollon, S.D., Shelton, R.C., Wisniewski, S., Warden, D., Biggs, M.M., Friedman, E.S., ... Shores-Wilson, K., 2006. Presenting characteristics of depressed outpatients as a function of recurrence: preliminary findings from the STAR* D clinical trial. J. Psychiatr. Res. 40 (1), 59–69. Joormann, J., Gotlib, I.H., 2007. Selective attention to emotional faces following recovery from depression. J. Abnorm. Psychol. 116 (1), 80. Kaiser, R.H., Andrews-Hanna, J.R., Wager, T.D., Pizzagalli, D.A., 2015. Large-scale network dysfunction in major depressive disorder. JAMA Psychiatry 72 (6), 603. Kellough, J.L., Beevers, C.G., Ellis, A.J., Wells, T.T., 2008. Time course of selective attention in clinically depressed young adults: an eye tracking study. Behav. Res. Ther. 46 (11), 1238–1243.
59
Contents lists available at ScienceDirect
Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad
Research paper
Ventral attention-network effective connectivity predicts individual differences in adolescent depression ⁎
Jie Liua,b,c, Pengfei Xua,c, Jingyuan Zhanga, Nengzhi Jiangb,d, Xinying Lib,d, , Yuejia Luoa,e,f,
T ⁎⁎
a
College of Psychology and Sociology, Shenzhen University, Shenzhen, China Key Laboratory of Mental Health, Institute of Psychology, Chinese Academy of Sciences, Beijing, China c Center for Brain Disorders and Cognitive Neuroscience, Shenzhen University, Shenzhen, China d Department of Psychology, University of Chinese Academy of Sciences, Beijing, China e Centre for Emotion and Brain, Shenzhen Institute of Neuroscience, Shenzhen, China f Department of Psychology, Southern Medical University, Guangzhou, China b
A R T I C LE I N FO
A B S T R A C T
Keywords: Ventral attention network Depression Attention bias Resting-state fMRI Stochastic dynamic causal modelling
Background: Stimulus-driven negative attention bias is a central deficit in depression and might play an important role in vulnerability to depression Adolescents are susceptible to depression. Thus, investigating the neural correlates of attention bias in adolescents is a critical step for identifying neural markers of early onset of depression. Previous studies have shown that the ventral attention network (VAN), which includes bilateral ventrolateral prefrontal cortex (VLPFC) and bilateral temporal-parietal junction (TPJ), is the key brain network for stimulus-driven attention. However, the relationship between depression and effective connectivity within the VAN in adolescents is poorly understood. Method: We employed resting-state fMRI to assess the relationship between directional effective connectivity within the VAN and depression scores in 216 healthy adolescents. Results: Using stochastic dynamic modeling, we found that individuals who exhibited higher self-reported depression showed stronger effective connectivity between right VLPFC and left TPJ within the VAN. Limitation: The level of depression in this study was assessed with self-reported questionnaire. This measure might be more influenced by current mood in adolescents than that in adults. Future studies should emplo more objective measures to index levels of depression. Conclusions: Our findings indicate that effective connectivity between right VLPFC and left TPJ could at least partially serve as a biomarker for bottom-up processing of depression in adolescents.
1. Introduction Depression is a prevalent and pervasive mental health disorder that is often not optimally treated (Disner et al., 2011). Adolescents are susceptible to depression, with a prevalence that can reach up to 14% in adolescents (Kessler andWalters, 1998). Adolescent-onset depression is associated with greater severity of symptoms, likelihood of relapse, and suicidal tendencies throughout life than adult-onset depression (Berndt et al., 2000; Zisook et al., 2007; Hollon et al., 2006). A central cognitive impairment that occurs in depression is the inability to allocate attention to appropriate emotional cues (Gotlib et al., 2004; Peckham et al., 2010). Adults with depression showed increased attention for negative stimuli and decreased attention for positive stimuli compared with nondepressed individuals (e.g., Eizenman et al., 2003; Kellough et al.,
⁎
2008). Moreover, although the speed with which adults with depression orient to sad stimuli is comparable to that of non-depressed individuals (Cisler and Koster 2010; Teachman et al., 2012), the amount of time they spend paying attention to them is longer by several seconds (Peckham et al., 2010). Similar attention bias can also be observed in adolescents with depression (e.g., Sylvester et al., 2016). The ventral attention network (VAN) is associated with the orientation of stimulus-driven attention, and its key brain regions are the bilateral ventrolateral prefrontal cortex (VLPFC) and bilateral temporalparietal junction (TPJ) (Corbetta et al., 2008). Using resting-state functional magnetic resonance imaging (rs-fMRI), studies have demonstrated that depression in adults was associated with abnormal VAN function. Meta-analysis of resting-state functional connectivity in major depressive disorder (MDD) suggested that MDD was related to
Corresponding author at: Key Laboratory of Mental Health, Institute of Psychology, Chinese Academy of Science, Beijing, China. Corresponding author at: Center for Brain Disorders and Cognitive Neuroscience, Shenzhen University, Shenzhen, China. E-mail addresses: [email protected] (X. Li), [email protected] (Y. Luo).
⁎⁎
https://doi.org/10.1016/j.jad.2019.04.033 Received 18 August 2018; Received in revised form 8 March 2019; Accepted 7 April 2019 Available online 08 April 2019 0165-0327/ © 2019 Published by Elsevier B.V.
Journal of Affective Disorders 252 (2019) 55–59
J. Liu, et al.
2.4. Image preprocessing
hypoconnectivity between the VAN seeds and the precuneus, extending to the occipital and posterior cingulate cortex (Kaiser et al., 2015). Another rs-fMRI meta-analysis indicated lower activation in the TPJ but increased activation in the VLPFC in patients with depression than in healthy controls (Fitzgerald et al., 2008). Because the brain changes significantly as adolescents mature, brain abnormalities in adolescent depression could be different from those in adults. Investigating the neurobiology of the VAN during the early stages of the disease is thus critical for understanding the etiology of depression. However, few studies have investigated the neurobiology of the VAN in adolescents. Moreover, directional effective connectivity within the VAN is poorly understood. In the current study, we adopted stochastic dynamic causal modeling (sDCM) for rs-fMRI in a large sample of adolescents to estimate the effective connectivity between regions within the VAN. The sDCM approach allows endogenous or random fluctuations in hidden states, such as neuronal activity, or hemodynamic states, such as local perfusion and deoxyhemoglobin content, and is this suitable for identifying the casual and directed nature of coupling between intrinsic modes of brain activity (Daunizeau et al., 2009; Riera et al., 2004). We applied sDCM to characterize the within-network effective connectivity of the VAN in healthy adolescents and to determine how it is related to depression (non-clinical). Given that previous task-fMRI studies showed both lower activity in the VLPFC and superior parietal lobule in adults with depression who were required to shift their attentional focus away from negative stimuli (Beevers et al., 2010; Fales et al., 2008), we hypothesized that individuals with higher depression scores might exhibit greater within-VAN effective connectivity.
Data preprocessing was performed using DPARSF software (Yan and Zang, 2010, http://www.restfmri.net). The T1 images were segmented into gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) partitions. The functional images were preprocessed with the following steps: 1) deletion of the first 10 volumes; 2) correction of slice timing; 3) assignment of each T1 image to a functional image; 4) normalization to Montreal Neurological Institute (MNI) space with solution of 3 × 3 × 3 mm; 5) spatial smoothing by convolution with an isotropic Gaussian kernel (FWHM = 6 mm); 6) linear trend removing; 7) bandpass (0.01–0.1 Hz) filtering; and 8) data scrubbing. The scrubbing procedure was performed following Power et al. (2012). We excluded any volume whose frame-wise dependent (FD) value exceeded 0.2 mm, along with the previous volume and the two following volumes. Fiftythree participants were excluded from the subsequent analysis because less than 150 of their volumes were left after the scrubbing procedure. The mean FD of the remaining participants was 0.07 ( ± 0.03). Further nuisance variables for rs-fMRI measures, such as motion parameters, the signal averaged over the individual segmented CSF and white matter (WM) regions, as well as the global signal (Murphy and Fox, 2017), were regressed out with a general linear model at the whole-brain level before the volume of interest extraction (see details in “Selection and extraction of volumes of interest”). 2.5. Selection and extraction of volumes of interest (VOI) The locations of the key cortical regions in the VAN were identified with spatial independent component analysis (ICA). The preprocessed images were analyzed using the Group ICA of the fMRI toolbox (GIFT) software package (Medical Image Analysis Lab, University of New Mexico, Albuquerque, New Mexico, USA) with the infomax Algorithm (Calhoun and Adali, 2004). Using spatial ICA, we separated spatially independent patterns from the time courses and generated 41 independent components for each participant. To identify the components that corresponded to parts of the VAN, spatial regression was performed to match with a pre-existing VAN template (Yeo et al., 2011). Components with the highest spatial regression coefficient between the 41-group mean component maps and the VAN template were selected as the network functional templates for VOI selection (seed regions for the following effective connectivity analysis). Four brain region masks, including bilateral TPJ and bilateral inferior frontal gyrus, were separately exacted from this VAN template. The VOIs were centered on the local maximum in the group mean map of the corresponding component within the four brain region masks mentioned above. For instance, the VLPFC VOI was selected based on the spatial regression of the ICA that identified a maximum ventral attention activation component within the VLPFC brain mask. Before dynamic casual modeling, we built a general liner model and then extracted the time-series data from the results. The general linear model included nuisance variables for rs-fMRI measures, such as 24 motion parameters (6 head-motion parameters, 6 head-motion parameters one time point before, and the 12 corresponding squared items), the signal averaged over the individual segmented CSF and white matter (WM) regions, as well as the global signal (Yan et al., 2013). For each participant, the principal eigenvariate of the VOI (the first principal component of the local multivariate time series over all voxels in the VOI) was defined as the seed of the models (radius = 6 mm).
2. Methods 2.1. Participants Two hundred and sixteen healthy adolescents (15.72 ± 0.94 years, 112 males) were recruited from the local community. All participants were healthy individuals without brain injuries or neurological diseases. This study was approved by the Institutional Review Board of the Institute of Psychology of the Chinese Academy of Sciences and the Institutional Review Board of Beijing MRI Center for Brain Research. 2.2. Questionnaires Participant self-reported depression levels were assessed by the Chinese version of the Children's Depression Inventory (CDI) (Kovacs, 1985). The Cronbach alpha was 0.85 for CDI. As depression shares some negative emotion characteristics with anxiety, to ensure this did not confound the issue, we also included the Chinese version of the Spielberger's Trait Anxiety Inventory (STAI-T) (Spielberger et al., 1970), which uses 20 items to measure self-reported anxiety levels. The Cronbach alpha was 0.87 for STAI-T in the current study. 2.3. MRI data acquisition All participants completed a six-minute rs-fMRI scan on a 3-Tesla SIEMENS MRI scanner (Beijing, China). Whole-brain T2*-weighted blood oxygenation level-dependent (BOLD) contrast images were acquired with single-shot gradient-recalled echo planar imaging (GREPI) sequences (repetition time = 2000 ms, echo time = 30 ms, flip angle = 90°, acquisition matrix = 64 × 64, field of view = 22 cm, slice thickness = 3 mm, alignment = AC-PC line). High-resolution T1weighted images were also acquired for each participant using a magnetization-prepared rapid gradient echo (MPRAGE) sequence (repetition time = 2530 ms, echo time = 3.37 ms, acquisition matrix = 256 × 192, field of view = 25.6 × 19.2 cm, flip angle = 7°, slice thickness = 1.33 mm, alignment = AC-PC line).
2.6. Specification and estimation of the sDCM at the first level The effective connectivity within the VAN was analyzed using sDCM implemented in SPM12 (version 6906, Wellcome Department of Imaging Neurosciences, University College London, UK, http://www. fil.ion.ucl.ac.uk/spm). The mathematical model of the underlying 56
Journal of Affective Disorders 252 (2019) 55–59
J. Liu, et al.
neuronal connectivity among DCM nodes is a system of bilinear differential state equations with coefficients specified by three matrices (A matrix, B matrix, and C matrix) (Friston et al., 2003). sDCM is different from conventional deterministic DCM. It seeks to improve model estimation by modeling endogenous or random fluctuations in unobserved (hidden) neuronal and physiological states in the differential equations of the neuronal states (Li et al., 2011a). sDCM can be used to explore the effectivity functional connectivity of rs-fMRI data. A full reciprocally connected model (full model) between each of the four seed regions, as well as the self-connections of each of the four seed regions, was constructed for each participant. Generalized filtering was then used for model inversion and parameter estimation. 2.7. Second level analysis of the sDCM Group level post-hoc optimization was conducted by selecting all estimated models. A Bayesian network discovery scheme was applied to search for the best model in the model space (Friston et al., 2011). Evidence of all the possible reduced models in the model space was obtained as indicators of the best model after network discovery. The model with the highest model evidence was selected as the winning model. Bayesian parameter averaging (BPA) was performed over the participants to determine the group-level optimal sparse model. The endogenous connectivity in DCMs is quantified by the “A” matrix parameters, which measure the effective connectivity strength between nodes. Thus, the endogenous connection strength between each node in the winning model was extracted in the DCM “A” matrix and used for the subsequent analysis for each participant. We performed correlation analysis between the connection strengths obtained in the DCM analysis and the depression scores. Statistical analysis was carried out using SPSS (IBM Corp., version 20. Armonk, NY, USA). Because depression shares some negative emotion characteristics with trait anxiety (r = 0.82, p < 0.001 in our data sample), to investigate the specific effect for depression, partial correlation analysis was adopted to control for anxiety. In addition, we also took demographic variables such as sex, age, and mean FD as covariates in the partial correlation analysis.
In the VAN, all 16 connections within the full model were significantly different from zero. After controlling for sex, age, mean FD, and the trait anxiety scores, the positive correlations were significant or marginally significant for the following connections: right VLPFC → left TPJ (r = 0.24, Bonferroni corrected p = 0.03), left TPJ → right VLPFC (r = 0.22, Bonferroni corrected p = 0.08) (Table 1). The partial correlation between trait anxiety and any effective connectivity within the VAN was not significant, after controlling for sex, age, mean FD, and the CDI scores: right VLPFC → left TPJ (r = −0.19, Bonferroni corrected p = 0.32), left TPJ → right VLPFC (r = −0.21, Bonferroni corrected p = 0.14).
2.8. Validation analysis
3.4. Validation results
Previous studies have suggested that global signals might introduce widespread negative functional connectivity and may alter the intrinsic correlation structure of brain networks (Ciric et al., 2017; Murphy et al., 2009; Weissenbacher et al., 2009). Thus, we repeated our general linear model and sDCM analysis without global signal regression.
The results of validation analysis that processed the data without global signal regression remained consistent with the main results. After controlling for sex, age, trait anxiety, and individual mean FD, the effective connectivity from the right VLPFC to the left TPJ could significantly predict the scores from the depression questionnaire
Fig. 1. ICA components selected by spatial regression.
connections of the four VOIs. The evidence showed that the best model was the full model (posterior probability > 0.99). The BPA result showed that all connections in the model were significantly greater or less than zero. 3.3. Within network effective connectivity and depression
3. Results
Table 1 Correlation between within network effective connectivity and depression.
3.1. Group ICA and selected VOIs Forty-one independent components were obtained from the group ICA analysis. We chose the intrinsic VAN template from Yeo et al. (2011). We then performed spatial regression with the template to identify the network-related component. After spatial sorting, we selected the top rank-ordered components with the highest spatial regression coefficient as the network-related component (regression coefficient = 0.18). The brain template and the corresponding selected component are presented in Fig. 1. The peak activation within the selected component and each brain region were considered the volumes of interest (VOIs). The result was two VOIs, bilateral VLPFC (MNI coordinates [−27, 36, −12] and [48, 42, 3]) and bilateral TPJ ([−60, −27, 18] and [63, −24, 27]).
Brain network
Effective connectivity
Partial correlation r p
VAN
Left TPJ –> left TPJ Right TPJ–> left TPJ Left VLPFC –> left TPJ Right VLPFC –> left TPJ Left TPJ–> right TPJ Right TPJ –> right TPJ Left VLPFC –>right TPJ Right VLPFC –> right TPJ Left TPJ –> left VLPFC Right TPJ –> left VLPFC Left VLPFC –> left VLPFC Right VLPFC –> left VLPFC Left TPJ –> right VLPFC Right TPJ –> right VLPFC Left VLPFC –> right VLPFC Right VLPFC –> right VLPFC
0.003 −0.021 0.193 0.240* −0.073 −0.001 −0.090 0.011 0.149 −0.037 0.014 0.046 0.224 0.087 0.009 0.058
3.2. sDCM ⁎
In the VANs, we searched all possible models generated from the 57
Bonferroni corrected p < 0.05.
0.972 0.794 0.015 0.002 0.359 0.989 0.259 0.895 0.060 0.645 0.856 0.568 0.005 0.276 0.911 0.468
Journal of Affective Disorders 252 (2019) 55–59
J. Liu, et al.
depression. They directed measured the attention bias to negative stimulus, and found positive correlation between threat bias and the VAN RSFC (between right VLPFC and right TPJ) for both the children with depression history and health control, which is in accordance with our result that showed higher level of depression associated with greater effective connectivity within VAN. However, they reported children with a history of depression had significantly reduced resting-state functional connectivity among the right regions comprising the VAN, and negative results was found among the left regions comprising the VAN. The difference might arise from the different experimental samples. In Sylvester et al. (2013), children with a history of depression or anxiety in this study had nonsignificant lower threat bias (i.e., threat avoidance) compared to subjects with no history of psychiatric illness. A general mechanism could be that lower threat bias was related reduced RSFC within the VAN. Thus children with histories of anxiety/ depression were observed decreased VAN resting state connectivity in Sylvester et al. (2013), but children with higher current depression might have the increased effective connectivity in the current study. During adolescence, the neurodevelopmental trajectory largely alters the brain, and the current results indicate that for some individuals, brain response patterns to unexpected stimuli might be prone to evolve into attention bias to negative stimulus, and these people might be at higher risk of depression. Overall, using a stochastic dynamic causal model, we discovered that depression scores were correlated to resting-state directional effective connectivity within the ventral attention networks of healthy adolescents. The significant positive correlation found in effective connectivity within the ventral attention network might serve as the neural basis for attention bias in depression. These findings might have clinical implications for treatment interventions, and suggest that monitoring neural changes within bottom-up processing networks might be useful.
(r = 0.25, Bonferroni corrected p = 0.03). The correlation between effective connectivity from the left TPJ to the right VLPFC and to CDI scores was marginally significant (r = 0.22, Bonferroni Corrected p = 0.08). The partial correlation between trait anxiety and any effective connectivity within the VAN was not significant, after controlling for sex, age, mean FD, and the CDI scores: right VLPFC → left TPJ (r = −0.19, Bonferroni corrected p = 0.24), left TPJ → right VLPFC (r = −0.20, Bonferroni corrected p = 0.22). 4. Discussion Although several studies have explored the neurobiology of the VAN, the key brain network for attention bias, in the depression of adults, few studies have investigated the neurobiology of the VAN in adolescents, especially the effective connectivity within the VAN. We investigated the within-VAN effective connectivity that is assumed to be related to stimulus-driven attention in a large sample of adolescents. We observed a significant positive correlation between depression scores and the within-network effective functional connectivity. Specifically, higher depression scores were related with increased bidirectional connection between the VLPFC and the TPJ. To the best of our knowledge, this is the first study to evaluate within-network directional effective connectivity related to the propensity for depression in healthy adolescents. Behavioral studies have shown that depressed adults (e.g., Joormann and Gotlib, 2007; Caseras et al., 2007) and adolescents (Sylvester et al., 2016) can be characterized by problems in disengagement from negative stimuli. Depressed individuals may not automatically orient their attention toward negative information in the environment, but once such information is noticed, they have greater difficulty disengaging from it. A series of neuroimaging results have suggested that this attentional bias is related to a neural mechanism underlying depression. For instance, this kind of negative attention bias is associated with several abnormal neural responses, such as decreased activation in the TPJ (rs-fMRI study; Fitzgerald et al., 2008), decreased activity in the right VLPFC and right superior parietal cortex (task-fMRI studies; Beevers et al., 2010; Fales et al., 2008), as well as increased magnitude of specific early P300, a characteristic event-related electrical potential seen on electroencephalograms (EEGs) (Bruder et al., 2002; Li et al., 2011b). Our finding indicates a role for the VAN in adolescents that was previously unknown. The increased positive correlation between left TPJ and right VLPFC for adolescents who reported higher depression scores might be due to the co-decrease of activity in these two brain regions at rest in participants who report more depressive symptoms. Given that the VLPFC has been shown to be related to stimulus-driven attention processing in both adults (Asplund et al., 2010) and adolescents (White et al., 2016), and that the TPJ has been reported to be related to post-perceptual processes involved in contextual updating and adjustments of top-down expectations (Geng and Vossel, 2013), we infer that during negative stimulus-driven attention, unexpected stimuli might lead to abnormally decreased right VLPFC activation in adolescents with higher risk of depression due to negative attention bias. These individuals tend to fixate on the negative stimulus. It might be harder for them to update/disengage their attention according to contextual information and top-down expectations, which would lead to decreased activation in the left TPJ. Thus, the effective connectivity from the right VLPFC to the left TPJ might be increased for more depressive adolescents. The effective connectivity from the left TPJ to the right VLPFC was probably the feedback signal from the right VLPFC→ left TPJ connection. The feedback signal might co-increase, but it would be relatively weaker because of signal attenuation. Thus, the effective connectivity from the left TPJ to the right VLPFC might also increase in more depressive adolescents. Sylvester et al. (2013) investigated the resting-state functional connectivity of the VAN in children aged 8–12 years with a history of
Conflict of interest We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work, there is no professional or other personal interest of any nature or kind in any product, service and/or company that could be construed as influencing the position presented in, or the review of, the manuscript entitled. Role of the funding source The research conducted is supported by The National Key Basic Research Program of China (Project no. 2014CB846100) and The Natural Science Foundation of China (Project nos. 31530031, 31700977). The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding source. Acknowledgments We thank the National Key Basic Research Program of China 2014CB846100, as well as the Natural Science Foundation of China (Project nos. 31530031, 31700977) for the financial support. Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.jad.2019.04.033. References Asplund, C.L., Todd, J.J., Snyder, A.P., Marois, R., 2010. A central role for the lateral prefrontal cortex in goal-directed and stimulus-driven attention. Nature Neurosci. 13 (4), 507.
58
Journal of Affective Disorders 252 (2019) 55–59
J. Liu, et al.
Kessler, R.C., Walters, E.E., 1998. Epidemiology of DSM-III-R major depression and minor depressionamong adolescents and young adults in the National Comorbidity Survey. Depress. Anxiety 7 (1), 3–14. Kovacs, M., 1985. The children's depression inventory (CDI). Psychopharmacol. Bull. 21, 995–998. Li, B., Daunizeau, J., Stephan, K.E., Penny, W., Hu, D., Friston, K., 2011a. Generalised filtering and stochastic DCM for fMRI. Neuroimage 58 (2), 442–457. Li, Y., Hu, Y., Liu, T., Wu, D., 2011b. Dipole source analysis of auditory P300 response in depressive and anxiety disorders. Cogn. Neurodyn. 5 (2), 221–229. Murphy, K., Birn, R.M., Handwerker, D.A., Jones, T.B., Bandettini, P.A., 2009. The impact of global signal regression on resting state correlations: are anti-correlated networks introduced? Neuroimage 44 (3), 893–905. Murphy, K., Fox, M.D., 2017. Towards a consensus regarding global signal regression for resting state functional connectivity MRI. Neuroimage 154, 169–173. Peckham, A.D., McHugh, R.K., Otto, M.W., 2010. A meta analysis of the magnitude of biased attention in depression. Depress. Anxiety 27 (12), 1135–1142. Power, J.D., Barnes, K.A., Snyder, A.Z., Schlaggar, B.L., Petersen, S.E., 2012. Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion. Neuroimage 59 (3), 2142–2154. Riera, J.J., Watanabe, J., Kazuki, I., Naoki, M., Aubert, E., Ozaki, T., Kawashima, R., 2004. A state-space model of the hemodynamic approach: nonlinear filtering of BOLD signals. NeuroImage 21 (2), 547–567. Spielberger, C.D., Gorsuch, R.L., Lushene, R.E., 1970. Manual for the State-Trait Anxiety Inventory. Sylvester, C.M., Barch, D.M., Corbetta, M., Power, J.D., Schlaggar, B.L., Luby, J.L., 2013. Resting state functional connectivity of the ventral attention network in children with a history of depression or anxiety. J. Am. Acad. Child Adolesc. Psychiatry 52 (12), 1326–1336. Sylvester, C.M., Hudziak, J.J., Gaffrey, M.S., Barch, D.M., Luby, J.L., 2016. Stimulusdriven attention, threat bias, and sad bias in youth with a history of an anxiety disorder or depression. J. Abnorm. Child Psychol. 44 (2), 219–231. Teachman, B.A., Joormann, J., Steinman, S.A., Gotlib, I.H., 2012. Automaticity in anxiety disorders and major depressive disorder. Clin. Psychol. Rev. 32 (6), 575–603. Weissenbacher, A., Kasess, C., Gerstl, F., Lanzenberger, R., Moser, E., Windischberger, C., 2009. Correlations and anticorrelations in resting-state functional connectivity MRI: a quantitative comparison of preprocessing strategies. Neuroimage 47 (4), 1408–1416. White, L.K., Britton, J.C., Sequeira, S., Ronkin, E.G., Chen, G., Bar-Haim, Y., ... Pine, D.S., 2016. Behavioral and neural stability of attention bias to threat in healthy adolescents. Neuroimage 136, 84–93. Yan, C., Zang, Y., 2010. DPARSF: a MATLAB toolbox for" pipeline" data analysis of resting-state fMRI. Front. Syst. Neurosci. 4, 13. Yan, C.G., Cheung, B., Kelly, C., Colcombe, S., Craddock, R.C., Di Martino, A., ... Milham, M.P., 2013. A comprehensive assessment of regional variation in the impact of head micromovements on functional connectomics. Neuroimage 76, 183–201. Yeo, Thomas, T., B., Krienen, F.M., Sepulcre, J., Sabuncu, M.R., Lashkari, D., Hollinshead, M., …, Buckner, R.L., 2011. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J. Neurophysiol. 106 (3), 1125–1165. https:// doi.org/10.1152/jn.00338.2011. Zisook, S., Lesser, I., Stewart, J.W., Wisniewski, S.R., Balasubramani, G.K., Fava, M., ... Trivedi, M.H., 2007. Effect of age at onset on the course of major depressive disxsorder. Am. J. Psychiatry 164 (10), 1539–1546.
Beevers, C.G., Clasen, P., Stice, E., Schnyer, D., 2010. Depression symptoms and cognitive control of emotion cues: a functional magnetic resonance imaging study. Neuroscience 167 (1), 97–103. Berndt, E.R., Koran, L.M., Finkelstein, S.N., Gelenberg, A.J., Kornstein, S.G., Miller, I.M., Thase, M.E., Trapp, G.A., Keller, M.B., 2000. Lost human capital from early-onset chronic depression. Am. J. Psychiatry 157 (6), 940–947. Bruder, G.E., Kayser, J., Tenke, C.E., Leite, P., Schneier, F.R., Stewart, J.W., Quitkin, F.M., 2002. Cognitive ERPs in depressive and anxiety disorders during tonal and phonetic oddball tasks. Clin. Electroencephalogr. 33 (3), 119–124. Calhoun, V.D., Adali, T., 2004. Group ICA of fMRI Toolbox (GIFT). Online at http://icatb. sourceforge. net. Caseras, X., Garner, M., Bradley, B.P., Mogg, K., 2007. Biases in visual orienting to negative and positive scenes in dysphoria: an eye movement study. J. Abnorm. Psychol. 116 (3), 491. Ciric, R., Wolf, D.H., Power, J.D., Roalf, D.R., Baum, G.L., Ruparel, K., ... Gur, R.C., 2017. Benchmarking of participant-level confound regression strategies for the control of motion artifact in studies of functional connectivity. Neuroimage 154, 174–187. Cisler, J.M., Koster, E.H., 2010. Mechanisms of attentional biases towards threat in anxiety disorders: an integrative review. Clin. Psychol. Rev. 30 (2), 203–216. Corbetta, M., Patel, G., Shulman, G.L., 2008. The reorienting system of the human brain: from environment to theory of mind. Neuron 58 (3), 306–324. Daunizeau, J., Kiebel, S.J., Friston, K.J., 2009. Dynamic causal modelling of distributed electromagnetic responses. NeuroImage 47 (2), 590–601. Disner, S.G., Beevers, C.G., Haigh, E.A., Beck, A.T., 2011. Neural mechanisms of the cognitive model of depression. Nat. Rev. Neurosci. 12 (8), 467. Eizenman, M., Lawrence, H.Y., Grupp, L., Eizenman, E., Ellenbogen, M., Gemar, M., Levitan, R.D., 2003. A naturalistic visual scanning approach to assess selective attention in major depressive disorder. Psychiatry Res. 118 (2), 117–128. Fales, C.L., Barch, D.M., Rundle, M.M., Mintun, M.A., Snyder, A.Z., Cohen, J.D., ... Sheline, Y.I., 2008. Altered emotional interference processing in affective and cognitive-control brain circuitry in major depression. Biol. Psychiatry 63 (4), 377–384. Fitzgerald, P.B., Laird, A.R., Maller, J., Daskalakis, Z.J., 2008. A meta-analytic study of changes in brain activation in depression. Human Brain Mapp. 29 (6), 683–695. Friston, K.J., Harrison, L., Penny, W., 2003. Dynamic causal modelling. Neuroimage 19 (4), 1273–1302. Friston, K.J., Li, B., Daunizeau, J., Stephan, K.E., 2011. Network discovery with DCM. Neuroimage 56 (3), 1202–1221. Geng, J.J., Vossel, S., 2013. Re-evaluating the role of TPJ in attentional control: contextual updating? Neurosci. Biobehav. Rev. 37 (10), 2608–2620. Gotlib, I.H., Krasnoperova, E., Yue, D.N., Joormann, J., 2004. Attentional biases for negative interpersonal stimuli in clinical depression. J. Abnorm. Psychol. 113 (1), 127. Hollon, S.D., Shelton, R.C., Wisniewski, S., Warden, D., Biggs, M.M., Friedman, E.S., ... Shores-Wilson, K., 2006. Presenting characteristics of depressed outpatients as a function of recurrence: preliminary findings from the STAR* D clinical trial. J. Psychiatr. Res. 40 (1), 59–69. Joormann, J., Gotlib, I.H., 2007. Selective attention to emotional faces following recovery from depression. J. Abnorm. Psychol. 116 (1), 80. Kaiser, R.H., Andrews-Hanna, J.R., Wager, T.D., Pizzagalli, D.A., 2015. Large-scale network dysfunction in major depressive disorder. JAMA Psychiatry 72 (6), 603. Kellough, J.L., Beevers, C.G., Ellis, A.J., Wells, T.T., 2008. Time course of selective attention in clinically depressed young adults: an eye tracking study. Behav. Res. Ther. 46 (11), 1238–1243.
59