Elsevier

NeuroImage

Volume 39, Issue 1, 1 January 2008, Pages 527-537
NeuroImage

Competition between functional brain networks mediates behavioral variability

https://doi.org/10.1016/j.neuroimage.2007.08.008Get rights and content

Abstract

Increased intraindividual variability (IIV) is a hallmark of disorders of attention. Recent work has linked these disorders to abnormalities in a “default mode” network, comprising brain regions routinely deactivated during goal-directed cognitive tasks. Findings from a study of the neural basis of attentional lapses suggest that a competitive relationship between the “task-negative” default mode network and regions of a “task-positive” attentional network is a potential locus of dysfunction in individuals with increased IIV. Resting state studies have shown that this competitive relationship is intrinsically represented in the brain, in the form of a negative correlation or antiphase relationship between spontaneous activity occurring in the two networks. We quantified the negative correlation between these two networks in 26 subjects, during active (Eriksen flanker task) and resting state scans. We hypothesized that the strength of the negative correlation is an index of the degree of regulation of activity in the default mode and task-positive networks and would be positively related to consistent behavioral performance. We found that the strength of the correlation between the two networks varies across individuals. These individual differences appear to be behaviorally relevant, as interindividual variation in the strength of the correlation was significantly related to individual differences in response time variability: the stronger the negative correlation (i.e., the closer to 180° antiphase), the less variable the behavioral performance. This relationship was moderately consistent across resting and task conditions, suggesting that the measure indexes moderately stable individual differences in the integrity of functional brain networks. We discuss the implications of these findings for our understanding of the behavioral significance of spontaneous brain activity, in both healthy and clinical populations.

Introduction

The identification of the neural substrates of interindividual differences in behavior and cognition is a fundamental objective for cognitive neuroscience. These investigations have typically focused on individual differences in measures of central tendency such as mean response time (RT) or accuracy. More recently, several behavioral investigations have shown that measures of dispersion or intraindividual variability (IIV) can provide information about performance that is not detectable by the mean (e.g., Hervey et al., 2006, Leth-Steensen et al., 2000). Such studies have found that measures of IIV (including standard deviation, SD, and coefficient of variation, CV) are better able to differentiate cognitively impaired (e.g., following sleep deprivation) and clinical groups (e.g., Attention-Deficit/Hyperactivity Disorder—ADHD) from healthy controls than measures of central tendency such as mean accuracy, error rates, or RT (Klein et al., 2006), giving rise to an increased focus on individual differences IIV as an indicator of neurological dysfunction (Castellanos et al., 2005, Chuah et al., 2006, Lim et al., 2007, MacDonald et al., 2006, Stuss et al., 2003). Cognitively, IIV is thought to index the efficiency with which attentional resources are allocated in the face of demands on cognitive control (Stuss et al., 2003, Bellgrove et al., 2004) and IIV is negatively correlated with general intelligence, g (Jensen, 1992). Frontal lobe circuits are implicated in IIV, and increased IIV is a hallmark of several conditions in which attention regulation and frontal lobe circuitry are compromised, including old age (West, 2001, West et al., 2002), Attention-Deficit/Hyperactivity Disorder (ADHD, Castellanos et al., 2005, Klein et al., 2006), traumatic brain injury (Stuss et al., 2003, Stuss et al., 1989) and schizophrenia (Schwartz et al., 1989; see Stuss et al., 2003, MacDonald et al., 2006).

Investigating the brain’s activity at rest provides a novel basis for examining individual differences in brain function (Buckner and Vincent, 2007). Raichle and colleagues (2001) first stimulated interest in resting state neural activity by describing a “default mode” network of brain regions that show strongly coherent spontaneous activity at rest, activity which is suppressed during performance of attention-demanding cognitive tasks. Recent studies have demonstrated abnormalities in the default mode network in conditions including ADHD (Tian et al., 2006, Yu-Feng et al., 2006), Alzheimer’s disease (Buckner et al., 2005, Greicius et al., 2004, Wang et al., 2006) and schizophrenia (Liang et al., 2006, Liu et al., 2006). Increased IIV is also observed in these conditions, suggesting that the examination of default mode activity may be relevant for the study of IIV.

A study by Weissman et al. (2006) supports this suggestion. They demonstrated that the failure to suppress default mode activity was associated with attentional lapses, indexed by longer response times (RTs), during a selective attention task. Attentional lapses were also associated with decreased task-related activity in prefrontal, anterior cingulate and visual cortices. Behavioral analyses (such as those utilizing Ex-Gaussian distributional measures) have shown that, in clinical populations characterized by increased IIV, the shape of individual response time distributions is predominantly determined by the presence of a large number of abnormally long RTs—an exaggerated positive skew, reflecting attentional lapses (Hervey et al., 2006, Leth-Steensen et al., 2000). Taken together, these findings suggest that behavioral variability may reflect trial-to-trial variability in the competitive balance of activity across regions that support task performance and those that are task-negative, rather than in the default mode network per se.

Intriguingly, examinations of functional connectivity during resting states show that this competitive relationship is intrinsically represented in the brain, in the form of a strong negative correlation (an antiphase relationship, also known as an “anticorrelation”, Fox et al., 2005) between the default mode network and a network of brain regions commonly activated during the performance of goal-directed cognitive tasks (e.g., working memory, attention; Fox et al., 2005, Fransson, 2005, Greicius et al., 2003), termed the “task-positive” network. The “task-positive’ network includes regions previously identified in various task-based neuroimaging studies as the “dorsal attention system,” in addition to dorso- and ventrolateral prefrontal cortex, dorsomedial prefrontal cortex and insula. This intrinsic negative relationship appears to replicate the common observation of task-related activity increases in frontal and parietal cortical regions (referred to as “task-positive” because they are typically active during task performance) and concomitant task-related activity decreases in the medial prefrontal, posterior cingulate, anterior temporal and lateral parietal cortices (i.e., the default mode network; referred to as “task-negative” because these regions are typically deactivated during task performance). The task-positive and task-negative networks are therefore thought to subserve opposing or competitive processes: task-focused attention (task-positive) and stimulus independent thought (task-negative, Fox et al., 2005). These opposing processes are hypothesized to compete for processing resources, consistent with Weissman et al.’s (2006) observation that decreased task-related suppression of activity in the default mode (task-negative) network is associated with poorer task performance (attentional lapses). That the same competitive relationship between task-positive and task-negative areas that is present during task performance is represented intrinsically in the brain’s spontaneous activity suggests that it reflects a fundamental functional characteristic of the brain and may therefore affect goal-directed behavior. However, its behavioral significance has not yet been examined.

This study aimed to evaluate the impact of the intrinsic competitive relationship between the default mode network and the task-positive attentional network on behavioral performance. Prior work suggests that cognitive performance may be facilitated when these two networks are disjunctively, rather than simultaneously, active (i.e., default mode activity is suppressed and task-positive areas are activated; Fransson, 2006, Greicius et al., 2003, Greicius and Menon, 2004, Weissman et al., 2006). Accordingly, we interpret the strength of an individual’s negative correlation between the default mode and task-positive networks as an index of the degree of regulation of activity in those networks. We hypothesized that the strength of the negative correlation would be positively related to consistent behavioral performance. That is, we hypothesized that a strong negative correlation (approaching r =  1, indicating that activity in the two networks is nearing 180° antiphase) would be associated with more consistent (less variable) performance. Conversely, we predicted that a weaker negative correlation (indicating a greater likelihood that the two networks are simultaneously active at times) would be associated with greater behavioral variability. We also expected that this link would be more evident under the more attentionally demanding incongruent condition, which was expected to increase the requirement to regulate activity between the networks.

Section snippets

Subjects and experimental paradigm

Twenty-six right-handed adults (mean age: 28.1 ± 8.5 years) participated in this event-related fMRI study. Participants provided signed informed consent prior to participation, in accordance with the institutional review boards of NYU and the NYU School of Medicine, and received monetary compensation for their participation. On each trial of a slow-paced Eriksen flanker task (inter-trial interval (ITI) varied between 8 s and 14 s; mean ITI = 12 s), participants used one of two buttons on a response

Behavioral and task activation data

The behavioral data exhibited the standard flanker effect (n = 26; mean congruent RT = 595.8 ms, SD (of mean RT population) = 146.2 ms, mean incongruent RT = 771.4 ms, SD = 215.8 ms, t(25) =  6.5694, p < 0.001; error rate < 1%). Consistent with previous work (e.g., Bunge et al., 2002, Fassbender et al., 2006, van Veen et al., 2001), the task elicited activation across a distributed frontoparietal network, with greater activation for incongruent than congruent trials (Fig. 2, yellow–red). Task performance also

Discussion

We investigated the potential functional significance of the intrinsic antiphase relationship between the default mode and task-positive networks by evaluating its association with behavioral performance. Three main findings emerged: (1) the strength of the negative correlation between the default mode and task-positive networks varies among individuals; (2) this antiphase relationship is largely consistent within an individual; (3) in line with our hypotheses, these individual differences

Acknowledgments

The authors would like to thank Daniel Margulies and Zarrar Shehzad for their help with data collection, and Amy Krain for helpful suggestions. This study was partially supported by 5T32MH067763, 5R21MH066393, and the Stavros S. Niarchos Foundation.

References (65)

  • C. Klein et al.

    Intra-subject variability in attention-deficit hyperactivity disorder

    Biol. Psychiatry

    (2006)
  • C. Leth-Steensen et al.

    Mean response times, variability, and skew in the responding of ADHD children: a response time distributional approach

    Acta. Psychol. (Amst)

    (2000)
  • S.W. MacDonald et al.

    Intra-individual variability in behavior: links to brain structure, neurotransmission and neuronal activity

    Trends Neurosci.

    (2006)
  • D.S. Margulies et al.

    Mapping the functional connectivity of anterior cingulate cortex

    NeuroImage

    (2007)
  • M.P. Milham et al.

    The relative involvement of anterior cingulate and prefrontal cortex in attentional control depends on nature of conflict

    Brain Res. Cogn. Brain Res.

    (2001)
  • L. Pessoa et al.

    Neural correlates of visual working memory: fMRI amplitude predicts task performance

    Neuron

    (2002)
  • D.N. Ruskin et al.

    Cocaine or selective block of dopamine transporters influences multisecond oscillations in firing rate in the globus pallidus

    Neuropsychopharmacology

    (2001)
  • B. Rypma et al.

    Neural correlates of cognitive efficiency

    NeuroImage

    (2006)
  • F. Schwartz et al.

    Reaction time impairment in schizophrenia and affective illness: the role of attention

    Biol. Psychiatry

    (1989)
  • L. Tian et al.

    Altered resting-state functional connectivity patterns of anterior cingulate cortex in adolescents with attention deficit hyperactivity disorder

    Neurosci. Lett.

    (2006)
  • W.A. Truccolo et al.

    Trial-to-trial variability of cortical evoked responses: implications for the analysis of functional connectivity

    Clin. Neurophysiol.

    (2002)
  • V. van Veen et al.

    Anterior cingulate cortex, conflict monitoring, and levels of processing

    NeuroImage

    (2001)
  • L. Wang et al.

    Changes in hippocampal connectivity in the early stages of Alzheimer’s disease: evidence from resting state fMRI

    NeuroImage

    (2006)
  • R. West et al.

    Lapses of intention and performance variability reveal age-related increases in fluctuations of executive control

    Brain Cogn.

    (2002)
  • E. Zarahn et al.

    Empirical analyses of BOLD fMRI statistics. I. Spatially unsmoothed data collected under null-hypothesis conditions

    NeuroImage

    (1997)
  • C.F. Beckmann et al.

    Investigations into resting-state connectivity using independent component analysis

    Philos. Trans. R. Soc. Lond., B Biol. Sci.

    (2005)
  • B. Biswal et al.

    Functional connectivity in the motor cortex of resting human brain using echo-planar MRI

    Magn. Reson. Med.

    (1995)
  • B. Biswal et al.

    Reduction of physiological fluctuations in fMRI using digital filters

    Magn. Reson. Med.

    (1996)
  • R.L. Buckner et al.

    Unrest at rest: the importance of default activity and spontaneous network correlations

    NeuroImage

    (2007)
  • R.L. Buckner et al.

    Molecular, structural, and functional characterization of Alzheimer’s disease: evidence for a relationship between default activity, amyloid, and memory

    J. Neurosci.

    (2005)
  • G. Buzsaki

    Rhythms of the Brain

    (2006)
  • F.X. Castellanos et al.

    Neuroscience of attention-deficit/hyperactivity disorder: the search for endophenotypes

    Nat. Rev., Neurosci.

    (2002)
  • Cited by (1056)

    View all citing articles on Scopus
    View full text