Elsevier

Journal of Psychiatric Research

Volume 82, November 2016, Pages 30-39
Journal of Psychiatric Research

Frequency-specific alterations in functional connectivity in treatment-resistant and -sensitive major depressive disorder

https://doi.org/10.1016/j.jpsychires.2016.07.011Get rights and content

Abstract

Major depressive disorder (MDD) may involve alterations in brain functional connectivity in multiple neural circuits and present large-scale network dysfunction. Patients with treatment-resistant depression (TRD) and treatment-sensitive depression (TSD) show different responses to antidepressants and aberrant brain functions. This study aims to investigate functional connectivity patterns of TRD and TSD at the whole brain resting state. Seventeen patients with TRD, 17 patients with TSD, and 17 healthy controls matched with age, gender, and years of education were recruited in this study. The brain was divided using an automated anatomical labeling atlas into 90 regions of interest, which were used to construct the entire brain functional networks. An analysis method called network-based statistic was used to explore the dysconnected subnetworks of TRD and TSD at different frequency bands. At resting state, TSD and TRD present characteristic patterns of network dysfunction at special frequency bands. The dysconnected subnetwork of TSD mainly lies in the fronto-parietal top-down control network. Moreover, the abnormal neural circuits of TRD are extensive and complex. These circuits not only depend on the abnormal affective network but also involve other networks, including salience network, auditory network, visual network, and language processing cortex. Our findings reflect that the pathological mechanism of TSD may refer to impairment in cognitive control, whereas TRD mainly triggers the dysfunction of emotion processing and affective cognition. This study reveals that differences in brain functional connectivity at resting state reflect distinct pathophysiological mechanisms in TSD and TRD. These findings may be helpful in differentiating two types of MDD and predicting treatment responses.

Introduction

Major depressive disorder (MDD) is a prevalent mental disorder characterized by persistent prevailing low mood and withdrawal from pleasurable activities (Minor et al., 2005). Patients with MDD suffer impairment in domains of emotional processing, cognitive control, affective cognition (cognitive control of emotion), and reward processing (Disner et al., 2011, Kerestes et al., 2014). MDD is one of the top public health concerns worldwide and causes significant disability and disease burden. Although numerous studies have focused on treatments for MDD, approximately one-third of patients with MDD fail to respond to antidepressants and are considered “treatment resistant” (Ionescu et al., 2015).

Previous neurobiological studies confirmed the occurrence of failure on mechanisms for serotonin reuptake inhibition in treatment-resistant depression (TRD) (Coplan et al., 2014) and is regarded as the basis of responses to antidepressant treatments. Changes may be correlated with genetic and biological factors, such as polymorphism of the serotonin transporter gene (Coplan et al., 2014, Santos et al., 2015). Therefore, the pathological mechanism of TRD may differ from that of treatment-sensitive depression (TSD); however, reliable biomarkers used to effectively predict treatment responses and identify two subtypes of depression are lacking (Jentsch et al., 2015).

Functional magnetic resonance imaging (fMRI) techniques are used to investigate the pathophysiology of MDD and identify biomarkers, which can be used to predict treatment responses. To date, the majority of fMRI studies have employed stimulus-driven paradigms, in which certain local brain functional abnormalities are found during cognitive or affective processing. The most consistent findings in MDD studies include decreased frontal cortex activity [primarily involving medial prefrontal cortex (MPFC) and dorsolateral prefrontal cortex (DLPFC)], as well as increased limbic system activity [including anterior cingulate cortex (ACC), amygdala, and hippocampus]at tasking state (Disner et al., 2011, Kerestes et al., 2014, Murray et al., 2011). For example, hyperactivity of amygdala and altered connectivity between the amygdala and ACC were probed under negative emotional face stimuli in MDD (Kerestes et al., 2014, Mingtian et al., 2012).

fMRI is employed in a ‘‘stimulus-free’’ manner, such as in the case of resting state, to reflect the intrinsic activity patterns of brain (Barkhof et al., 2014). A hypothesis proposes that MDD is associated with dysregulated neural networks, rather than disruption of single brain regions (Gong and He, 2015, Palazidou, 2012). Alterations in brain networks, including default mode network (DMN), salience network (SN), cognitive control network (CCN), and affective network (AN), have been identified in MDD (Guo et al., 2014, Kaiser et al., 2015, Luo et al., 2015, Wang et al., 2012, Zeng et al., 2012). Previous studies on functional connectivity (FC) within and between these networks at resting state showed that MDD exhibits hypoconnectivity within the CCN network, which is mainly composed of DLPFC and parts of the parietal lobe and involved in achieving goal-relevant stimuli, regulation of cognitive process, and top-down regulation of attention and emotion (Dichter et al., 2015, Ruge and Wolfensteller, 2015). MDD is also associated with hyperconnectivity within the DMN, which contains several brain regions located in the center of the brain, such as ACC, MPFC, and posterior cingulate cortex (PCC)/precuneus regions. This network is possibly involved in episodic memory and internally oriented, self-referential thought (Guo et al., 2014, Marchetti et al., 2012). Between networks, brain regions belonging to DMN exhibit hyperconnectivity with CCN and SN (insula) (Sawaya et al., 2015). Finally, another robust neuropathology patterns have gained increased research attention. The dysregulation of cortical–limbic–subcortical circuit (sometimes named as affective network, AN) is assumed to perform a vital role in the pathogenesis of depression (Maletic and Raison, 2014, Wang et al., 2012). Furthermore, structures in limbic and subcortical areas show abnormal activation in MDD; these structures include medial thalamus, amygdala, striatum, and hippocampus/parahippocampal, which are possibly involved in emotional perception and function in neural responses to negative stimuli. Cortical regions, such as ACC, and ventromedial prefrontal cortex (vMPFC), are thought to perform a regulatory role over limbic structures, which process emotional stimuli. A breakdown in this circuit could be related to deficits of mood regulation (Maletic and Raison, 2014, Palazidou, 2012). Several studies on MDD reported decreased functional connectivity between ACC and amygdala, pallido striatum, and thalamus (Anand et al., 2009). Decreased functional connectivity between the ACC and a number of cortical areas, including MPFC, superior and inferior frontal cortices, and insula, have also been reported in MDD (Gong and He, 2015, Wang et al., 2016). These findings suggest that MDD may involve alterations in brain connectivity in multiple neural circuits and exhibit large-scale network dysfunction.

Studies have investigated the involvement of alterations in brain function in MDD in responses to antidepressant treatments. For example, TRD shows robust decreased regional homogeneity (ReHo) in the prefrontal cortex and increased ReHo in limbic regions, as well as decreased connectivity within cortical–limbic circuits relative to TSD (Lui et al., 2011, Wu et al., 2011). Furthermore, successful antidepressant treatment of MDD results in increased connectivity among PFC, ACC, and limbic regions (thalamus, striatum, and amygdala) (Anand et al., 2007). By contrast, higher amplitude of low-frequency fluctuations and hyperconnectivity within the DMN was found in TRD relative to TSD. In certain studies, hyperconnectivity within the DMN was normalized in MDD after successful treatment (Guo et al., 2012, Li et al., 2013, Posner et al., 2013). In addition, low connectivity within the CCN predicts poor antidepressant outcomes in MDD (Alexopoulos et al., 2012). However, most previous studies used traditional FC analysis and preselected smaller regions of interest, thereby complicating the process of describing the pattern of brain FC at the whole-brain scale.

In this study, we investigated FC alterations in TSD and TRD by using an analysis method, called network-based statistic (NBS) (Zalesky et al., 2010). NBS is a powerful method when performing this kind of analysis. And can be thought of as a translation of conventional cluster statistics to a graph. NBS differs from cluster-based statistical methods used in mass univariate testing. Rather than in physical space, NBS clusters in topological space, where the most basic equivalent of a cluster is a connected graph component. Hence, FC alterations are identified and modeled as a network. Furthermore, NBS can offer substantially greater power than generic procedures for controlling family-wise error rate (Zalesky et al., 2010). Moreover, recent studies indicate that different frequency bands contribute differently to the low-frequency oscillations (LFOs), and frequency-dependent changes in LFOs have been reported in various brain disorders (Yu et al., 2014). Previous studies have suggested that the functional connectivity abnormalities in spontaneous low frequency (0.01–0.08 Hz) oscillations, but without detailedly description the functional connectivity abnormalities in more narrow frequency bands, for example, slow4 (0.027–0.073 Hz) and slow5 (0.01–0.027 Hz). Zuo have found that LFO amplitudes in slow4 were higher than in the slow5 in many brain regions, such as the basal ganglia, thalamus, and precuneus (Zuo et al., 2010). Han and his colleague (Han et al., 2011) found that amnestic mild cognitive impairment patients have widespread abnormalities in intrinsic brain activity depend on the difference ALFF/fALFF activities in the slow4 and slow5, Luo et al. found major depression disorder patients have abnormal brain network connectivity in different frequency bands (Luo et al., 2015). Indeed, frequency-dependent changes have been found in many diseases (Yu et al., 2014). However, no research focus on the functional connectivity abnormality in slow4 and slow5 based on NBS. Therefore, in the current study, we directly tested whether MDD-related changes in large-scale brain connectivity are dependent on frequency. This exploration will be helpful in discovering neural mechanisms underlying the MDD and provide additional information to improve understanding of the neurobiology of this disorder.

Section snippets

Subjects

A total of 35 right-handed patients with MDD were recruited from the Mental Health Institute, the Second Xiangya Hospital, Central South University, China. All patients were interviewed by two experienced psychiatrists using the Structured Clinical Interview for DSM-IV-TR-Patient Edition (SCID-P, 2/2001 revision, Biometrics Research Department, New York State Psychiatric Institute, USA, Web page: http://www.scid4.org/). DSM-IV criteria for MDD were used for diagnosis. Exclusion criteria include

Demographic and clinical characteristics

Seventeen patients with TRD, 17 patients with TSD, and 17 HCs completed the study. Demographic information and clinical characteristics are presented in Table 2. No significant difference in age (ANOVA, F = 0.81, P = 0.451), gender (Chi-square test, Chi-square value = 0, P = 1), years of education (ANOVA, F = 1.53, P = 0.236) was found among the three groups. Moreover, patient groups did not differ significantly in HAMD scores (t-test, T = −0.92, P = 0.363) and age of first episode (t-test,

Discussion

This study used a new approach, namely, NBS, to identify FC alterations at the resting state of two MDD subtypes, which are modeled as a network. In addition, we directly tested MDD-related changes in large-scale brain connectivity at different frequency bands. In support of previous studies suggesting that MDD may present large-scale network dysfunction, our results showed that brain dysfunction in MDD involves alterations in brain connectivity in multiple neural circuits, as well as

Contributors

Design the study: Huafu Chen, Jingping Zhao and Xujun Duan.

Recruit subjects and collect clinical information: Zongling He, Qian Cui, Yajing Pang.

MRI scan and acquire imaging data: Jiao Li, Xiao Wang, Shaoqiang Han, Yifeng Wang.

MRI data preprocessing: Junjie Zheng, Qing Gao and Zhiliang Long.

Data analysis and write the paper: Zongling He, Qian Cui.

Acknowledgments

We thank the 863 project (2015AA020505), the 973 project (2012CB517901), the Natural Science Foundation of China (61533006 and 31400901), the Fundamental Research Funds, and the Scientific research project of Sichuan Medical Association (S15012), which provided the funding for collecting data, paying remuneration and traffic expenditures fee to participants; the Youth Innovation Project of Sichuan Provincial Medical Association (Q14014), the Science Foundation of Ministry of Education of China (

References (48)

  • B. Li et al.

    A treatment-resistant default mode subnetwork in major depression

    Biol. Psychiatry

    (2013)
  • N.V. Malykhin et al.

    Hippocampal neuroplasticity in major depressive disorder

    Neuroscience

    (2015)
  • Z. Mingtian et al.

    Elevated amygdala activity to negative faces in young adults with early onset major depressive disorder

    Psychiatry Res.

    (2012)
  • K.L. Minor et al.

    Stability of DSM-IV criterion symptoms for major depressive disorder

    J. Psychiatr. Res.

    (2005)
  • M. Santos et al.

    FAS -670A>G genetic polymorphism is associated with treatment resistant depression

    J. Affect. Disord.

    (2015)
  • L. Wang et al.

    A systematic review of resting-state functional-MRI studies in major depression

    J. Affect Disord.

    (2012)
  • Y.-L. Wang et al.

    Altered functional interaction hub between affective network and cognitive control network in patients with major depressive disorder

    Behav. Brain Res.

    (2016)
  • A. Zalesky et al.

    Network-based statistic: identifying differences in brain networks

    Neuroimage

    (2010)
  • A. Zalesky et al.

    Disrupted axonal fiber connectivity in schizophrenia

    Biol. Psychiatry

    (2011)
  • X.-N. Zuo et al.

    The oscillating brain: complex and reliable

    NeuroImage

    (2010)
  • A. Anand et al.

    Reciprocal effects of antidepressant treatment on activity and connectivity of the mood regulating circuit: an FMRI study

    J. Neuropsychiatry Clin. Neurosci.

    (2007)
  • F. Barkhof et al.

    Resting-state functional MR imaging: a new window to the brain

    Radiology

    (2014)
  • J.D. Coplan et al.

    A neurobiological hypothesis of treatment-resistant depression – mechanisms for selective serotonin reuptake inhibitor non-efficacy

    Front. Behav. Neurosci.

    (2014)
  • S.G. Disner et al.

    Neural mechanisms of the cognitive model of depression

    Nat. Rev. Neurosci.

    (2011)
  • Cited by (35)

    View all citing articles on Scopus
    View full text