Elsevier

NeuroImage

Volume 129, 1 April 2016, Pages 233-246
NeuroImage

Dynamic coupling of complex brain networks and dual-task behavior

https://doi.org/10.1016/j.neuroimage.2016.01.028Get rights and content

Highlights

  • Dynamics of brain network topology predict fluctuations in dual-task performance.

  • Network integration differentially correlates with behavior in each component task.

  • Global integration of dynamic brain networks predicts decrease in visual accuracy.

  • Global integration of dynamic brain networks predicts increase in speech accuracy.

  • Dynamic integration of PCC with brain network modules underlies dual-task behavior.

Abstract

Multi-tasking is a familiar situation where behavioral performance is often challenged. To date, fMRI studies investigating the neural underpinning of dual-task interference have mostly relied on local brain activation maps or static brain connectivity networks. Here, based on task fMRI we explored how fluctuations in behavior during concurrent performance of a visuospatial and a speech task relate to alternations in the topology of dynamic brain connectivity networks. We combined a time-resolved functional connectivity and complex network analysis with a sliding window approach applied to the trial by trial behavioral responses to investigate the coupling between dynamic brain networks and dual-task behavior at close temporal proximity. Participants showed fluctuations in their dual-task behavior over time, with the accuracy in the component tasks being statistically independent from one another. On the global level of brain networks we found that dynamic changes of network topology were differentially coupled with the behavior in each component task during the course of dual-tasking. While momentary decrease in the global efficiency of dynamic brain networks correlated with subsequent increase in visuospatial accuracy, better speech performance was preceded by higher global network efficiency and was followed by an increase in between-module connectivity over time. Additionally, dynamic alternations in the modular organization of brain networks at the posterior cingulate cortex were differentially predictive for the visuospatial as compared to the speech accuracy over time. Our results provide the first evidence that, during the course of dual-tasking, each component task is supported by a distinct topological configuration of brain connectivity networks. This finding suggests that the failure of functional brain connectivity networks to adapt to an optimal topology supporting the performance in both component tasks at the same time contributes to the moment to moment fluctuations in dual-task behavior.

Introduction

Multi-tasking is often concurrent with impairment in performance, and introduces a behavioral challenge. The underlying cognitive architecture of the brain's capacity limit in multi-tasking has been investigated by several behavioral and neuroimaging studies ( Marois and Ivanoff, 2005, Lien et al., 2006, Just et al., 2008, Magen and Cohen, 2010, Remy et al., 2010, Huestegge et al., 2014, Nijboer et al., 2014, Watanabe and Funahashi, 2014). To this end, previous functional magnetic resonance imaging (fMRI) studies have mostly relied on local changes in blood–oxygen-level-dependent (BOLD) activations obtained under different experimental conditions. Recent BOLD-activation studies suggest that the brain's capacity limit in dual-tasking results from interference between neural processes ( Remy et al., 2010, Cohen et al., 2014, Nijboer et al., 2014). Here, we investigate whether alternations in the topological organization of brain functional connectivity networks over time relate to trial by trial variability in dual-task performance.

Recent functional connectivity and graph-theoretical studies suggest that performing two tasks simultaneously or in close succession relies on the flexible reconfiguration of the brain networks ( Ekman et al., 2012, Alavash et al., 2015). In Alavash, Hilgetag, et al. (2015) we showed that (1) topological overlap between single-task network modules was associated with higher dual-task interference and (2) topological reconfiguration of each single-task network modules in adaptation to the dual-task condition was associated with lower dual-task interference. For this, in our previous analyses we assumed that the topology of brain networks changes from one task condition to the other, but is static under each task condition. Thus, brain graphs were constructed based on time-averaged connectivity matrices (i.e. static networks) and between-subject differences in modular reorganization were correlated with between-subject differences in behavioral dual-task costs. As such, the focus of the previous investigations has been on the static topology of brain networks built upon the time-averaged connectivity patterns which disregard the fluctuations in functional brain network organization over time ( Fox et al., 2007, Allen et al., 2012, Zalesky et al., 2014).

In recent years there has been a growing interest in studying the dynamics of the functional connectivity patterns observed in large-scale brain networks ( Hutchison et al., 2013, Lindquist et al., 2014). Several studies have documented that the organization of brain networks are not stable, but dynamically change over time and following changes in task conditions ( Wang et al., 2012, Bassett et al., 2013, Mantzaris et al., 2013, Bola and Sabel, 2015). Many authors have suggested that such patterns are potentially relevant to the fluctuations in cognition and behavior ( Spoormaker et al., 2010, Allen et al., 2012, Jones et al., 2012, Bassett et al., 2013, Tagliazucchi et al., 2013, Schaefer et al., 2014, Zalesky et al., 2014, Sadaghiani et al., 2015). Thus, studying the dual-task interference might profit from the analysis of topological reconfiguration of brain network modules within short time bins (instead of nine-minute runs as it was done in Alavash, Hilgetag, et al., 2015 ), or preferably at a higher temporal resolution. If changes in the modular structure of brain networks across task conditions are potentially relevant to behavior, then fluctuations in behavior within a dual-task condition might also be associated with changes in brain network modularity within short time windows. Thus, while previous studies provide insights into complex brain network correlates of behavioral dual-task costs, it is unknown whether trial by trial variability in dual-task behavior relates to the fluctuations in functional brain network organization during the course of task performance. The aim of the present study is to investigate the coupling of such dynamic network patterns with the dynamics of dual-task behavior at close temporal proximity.

In the present study, we revisited the data of our previous fMRI experiment where subjects performed a visuospatial and a speech task in parallel ( Alavash, Hilgetag, et al., 2015 ; Fig. 1 A). In contrast to the previous static network analysis, here we asked how fluctuations in behavior during the course of the dual-task condition relate to fluctuations in the topology of dynamic brain connectivity networks. To answer this question, we combined a time-resolved correlation analysis ( Zalesky et al., 2014 ) with a sliding window approach applied to the trial by trial behavioral responses to capture the possible temporal coupling of complex brain network metrics with the dual-task behavior ( Fig. 2 ). To best of our knowledge, this is the first study investigating the relation between dynamic brain connectivity networks and dynamics of behavioral performance. A recent study conducted in our lab suggests that performance in different cognitive tasks profits from specific patterns of functional brain network integration at different topological scales, i.e. local, intermediate or global network integration ( Alavash, Doebler, et al., 2015 ). Accordingly, here we expected that distinct yet dynamic patterns of network integration generated at different topological scales correlate with the performance in each dual-task component (i.e. visuospatial or speech task processing). Thus, we hypothesized that moment to moment fluctuations in the dual-task behavior stem from the failure of the brain networks to maintain two different topological configurations at the same time, each supporting the performance in one of the dual-task components.

Section snippets

Subjects

Twenty-four healthy, right-handed subjects (12 females, 12 males, all native German speakers) whose age was in the range of 19–32 years (mean = 24.22 years, standard error of mean [SEM] = 0.74 years) participated in the experiment. Two male subjects showed large head movements during scanning (more than 4 mm of translation or degrees of rotation) and were excluded from further analyses. Ethical approval was obtained from the ethics committee of the University of Oldenburg. All procedures were carried

Behavioral data

The performance of the subjects was measured using the average accuracy (%) for responses to the visuospatial or speech targets within each quarter (35 trials) separately. The average accuracies obtained from responses to target stimuli were compared between different task conditions (i.e. visuospatial vs. speech) and quarters (one to four) by means of an analysis of variance (ANOVA) for repeated measures.

The statistical dependency between responses to each component task in the course of

Dynamics of behavioral performance

The average accuracies (%) for responses to the visuospatial or speech targets across quarters are presented in Fig. 1 B. The results from ANOVA revealed a significant effect of task: the accuracy in the speech task was significantly lower than that of visuospatial task (F(1,21) = 35.27, p < 0.001). To investigate the within-subject dependency between visuospatial and speech performance, we computed the Yule's Q coefficient per subject (see ‘ Analysis of behavioral accuracy ’ section). The mean

Dynamics of functional connectivity networks and its specific coupling with behavior

The overall aim of Alavash, Hilgetag, et al. (2015) and the current study is to find complex network correlates of behavioral impairment often observed under dual-task conditions. In Alavash, Hilgetag, et al. (2015) we were able to associate dual-task behavioral costs with changes in the topological configuration of brain network modules. The current study complements our previous work by featuring the following unique aspects. First, in the current study we analyzed the dynamic changes of

Conclusion

To conclude, the results of this study suggest that the successful performance of two tasks in parallel relies on a specific topological configuration of functional brain connectivity networks. Consistent with our hypothesis, we were able to link the fluctuations in dual-task behavior with the dynamic topological structure of the brain functional connectivity networks.

Acknowledgments

M. A. and C. G. were supported by the Hanse-Wissenschaftskolleg. Parts of the analyses were performed at the High Performance Computer Cluster HERO, located at the University of Oldenburg (Germany) and funded by the DFG through its Major Research Instrumentation Program (INST 184/108-1 FUGG) and the Ministry of Science and Culture (MWK) of the Lower Saxony State.

Conflict of interest

The authors declare no competing financial interests.

References (109)

  • K.J. Friston et al.

    Nonlinear responses in fMRI: the balloon model, Volterra kernels, and other hemodynamics

    NeuroImage

    (2000)
  • D.R. Gitelman et al.

    Modeling regional and psychophysiologic interactions in fMRI: the importance of hemodynamic deconvolution

    NeuroImage

    (2003)
  • D.A. Handwerker et al.

    Variation of BOLD hemodynamic responses across subjects and brain regions and their effects on statistical analyses

    NeuroImage

    (2004)
  • D.A. Handwerker et al.

    Periodic changes in fMRI connectivity

    NeuroImage

    (2012)
  • M. Havlicek et al.

    Dynamic modeling of neuronal responses in fMRI using cubature Kalman filtering

    NeuroImage

    (2011)
  • L. Huestegge et al.

    Talking while looking: on the encapsulation of output system representations

    Cogn. Psychol.

    (2014)
  • R.M. Hutchison et al.

    Dynamic functional connectivity: promise, issues, and interpretations

    NeuroImage

    (2013)
  • M.A. Just et al.

    A decrease in brain activation associated with driving when listening to someone speak

    Brain Res.

    (2008)
  • A.M. Kelly et al.

    Competition between functional brain networks mediates behavioral variability

    NeuroImage

    (2008)
  • N.J. Kopell et al.

    Beyond the connectome: the dynome

    Neuron

    (2014)
  • A. Kucyi et al.

    Dynamic functional connectivity of the default mode network tracks daydreaming

    NeuroImage

    (2014)
  • A.N. Landau et al.

    Distributed Attention Is Implemented through Theta-Rhythmic Gamma Modulation

    Curr. Biol.

    (2015)
  • N. Leonardi et al.

    On spurious and real fluctuations of dynamic functional connectivity during rest

    NeuroImage

    (2015)
  • C.H. Liao et al.

    Estimating the delay of the fMRI response

    NeuroImage

    (2002)
  • M.A. Lindquist et al.

    Evaluating dynamic bivariate correlations in resting-state fMRI: a comparison study and a new approach

    NeuroImage

    (2014)
  • R. Marois et al.

    Capacity limits of information processing in the brain

    Trends Cogn. Sci.

    (2005)
  • R.P. Monti et al.

    Estimating time-varying brain connectivity networks from functional MRI time series

    NeuroImage

    (2014)
  • M. Nijboer et al.

    Single-task fMRI overlap predicts concurrent multitasking interference

    NeuroImage

    (2014)
  • J.M. Palva et al.

    Infra-slow fluctuations in electrophysiological recordings, blood–oxygenation-level-dependent signals, and psychophysical time series

    NeuroImage

    (2012)
  • F. Remy et al.

    Dual-task interference during initial learning of a new motor task results from competition for the same brain areas

    Neuropsychologia

    (2010)
  • A. Roebroeck et al.

    The identification of interacting networks in the brain using fMRI: model selection, causality and deconvolution

    NeuroImage

    (2011)
  • M. Rubinov et al.

    Complex network measures of brain connectivity: uses and interpretations

    NeuroImage

    (2010)
  • M. Rubinov et al.

    Weight-conserving characterization of complex functional brain networks

    NeuroImage

    (2011)
  • S. Ryali et al.

    Multivariate dynamical systems models for estimating causal interactions in fMRI

    NeuroImage

    (2011)
  • G. Schöner et al.

    A dynamic pattern theory of behavioral change

    J. Theor. Biol.

    (1988)
  • K. Steinhaeuser et al.

    Identifying and evaluating community structure in complex networks

    Pattern Recogn. Lett.

    (2010)
  • N. Tzourio-Mazoyer et al.

    Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain

    NeuroImage

    (2002)
  • M.P. van den Heuvel et al.

    Exploring the brain network: a review on resting-state fMRI functional connectivity

    Eur. Neuropsychopharmacol.

    (2010)
  • M.P. van den Heuvel et al.

    Network hubs in the human brain

    Trends Cogn. Sci.

    (2013)
  • H. van der Voet

    Comparing the predictive accuracy of models using a simple randomization test

    Chemom. Intell. Lab.

    (1994)
  • Z. Wang et al.

    Changes in the brain intrinsic organization in both on-task state and post-task resting state

    NeuroImage

    (2012)
  • S. Achard et al.

    A resilient, low-frequency, small-world human brain functional network with highly connected association cortical hubs

    J. Neurosci.

    (2006)
  • Y.Y. Ahn et al.

    Link communities reveal multiscale complexity in networks

    Nature

    (2010)
  • M. Alavash et al.

    Persistency and flexibility of complex brain networks underlie dual-task interference

    Hum. Brain Mapp.

    (2015)
  • A.F. Alexander-Bloch et al.

    Disrupted modularity and local connectivity of brain functional networks in childhood-onset schizophrenia

    Front. Syst. Neurosci.

    (2010)
  • E.A. Allen et al.

    Tracking whole-brain connectivity dynamics in the resting state

    Cereb. Cortex

    (2012)
  • D.S. Bassett et al.

    Dynamic reconfiguration of human brain networks during learning

    PNAS

    (2010)
  • D.S. Bassett et al.

    Task-based core-periphery organization of human brain dynamics

    PLoS Comput. Biol.

    (2013)
  • Y. Benjamini et al.

    Controlling the false discovery rate — a practical and powerful approach to multiple testing

    J Roy Stat Soc B Met

    (1995)
  • S. Blondel et al.

    Fast unfolding of communities in large networks

    J. Stat. Mech.

    (2008)
  • Cited by (16)

    • The brain under cognitive workload: Neural networks underlying multitasking performance in the multi-attribute task battery

      2022, Neuropsychologia
      Citation Excerpt :

      Thus, the use of dFC represents one approach to address this gap as the temporal focus is more refined than in sFC which may in turn help detect additional network connections underlying rapid behavioural responses required during multitasking. In our literature review, we have identified four studies that used dFC to analyze fMRI data from a multitasking paradigm (Alavash et al., 2016; Maillet et al., 2019; Petrican and Grady, 2017; Zhang et al., 2019). Collectively, these studies highlight the value of employing dFC analysis because it can help show the neural profile that underlies multitasking, which may not be observed from sFC analysis.

    • Pinning bipartite synchronization for coupled reaction–diffusion neural networks with antagonistic interactions and switching topologies

      2021, Neural Networks
      Citation Excerpt :

      It is well known that the investigation of synchronization among a group of interconnected nodes by means of only local actions has been an active study field in systems and control throughout the past decade, and recently it has attracted much attention in science and technology such as communications networks, cellular networks, disease transmission networks, power grids, social networks, World Wide Web, and food webs. Therefore, analysis and control of synchronization behaviors in complex networks have been extensively investigated by many scholars and experts (Alavash & Thief, 2016; Altafini, 2013; Arenas, Díaz-Guilera, Kurths, Moreno, & Zhou, 2008; Chen, Liu, & Lu, 2007; He, Ji, Zhang, & Wu, 2016; Li, Lou, Wang and Alsaedi, 2018; Lu & Chen, 2019; Nie & Chen, 2010; Wang et al., 2021; Wang, Qin, Wu, Huang and Wei, 2019; Wu, 2005; Yang, Li, & Qiu, 2019). Some interesting results, such as crack synchronization, phase synchronization, impulsive stability and synchronization have been reported in Bao, Cao, Kürths, Alsaedi, and Ahmad (2018), Ji, Lu, Lou, Qiu, and Shi (2020), Jiang, Lu, and Liu (2020), Li (2017), Li, Shen and Rakkiyappan (2018), Liu, Song, Wen, Cao, and Yang (2018) and Wang, Lu and Lou (2019).

    • Large-scale network interactions involved in dividing attention between the external environment and internal thoughts to pursue two distinct goals

      2019, NeuroImage
      Citation Excerpt :

      The first goal of the current study was to provide insight into the nature of such network reconfigurations. Although network interactions subserving internal-external dual-tasking have not been studied extensively, many studies have examined the neural correlates of dual-tasking involving two perceptually-based tasks (Alavash et al., 2015; Alavash et al., 2016; Jiang, 2004; Nijboer et al., 2014; Szameitat et al., 2016). Dual-tasking often comes at a behavioral cost compared to performing tasks in isolation, and a general principle that has emerged from this literature is that such behavioral costs emerge to the extent that the neural correlates of the two single tasks overlap.

    • Principles of dynamic network reconfiguration across diverse brain states

      2018, NeuroImage
      Citation Excerpt :

      Similarly, Shine and colleagues (Shine et al., 2016a) demonstrated that task-related integration in network topology related to fast, effective cognitive performance on the N-back task (Fig. 3). This dissolution of modularity during N-back performance has been corroborated by other groups (Alavash et al., 2016; Chen et al., 2016; Cohen and D'Esposito, 2016; Davison et al., 2015; Shine et al., 2016a; Vatansever et al., 2015), demonstrating that brain-wide integration during cognitively complex tasks may act as a predictive signature of individual differences in executive function (Braun et al., 2015; Gallen et al., 2016; Schultz and Cole, 2016; Shine et al., 2016a). Replicating these fMRI results, a number of studies using electrophysiology during N-back tasks have shown similar increases in network-level integration as a function of cognitive performance (Bola and Borchardt, 2016; de Pasquale et al., 2012; Kitzbichler et al., 2011; Zippo et al., 2016).

    View all citing articles on Scopus
    View full text