Changes in resting connectivity during recovery from severe traumatic brain injury

https://doi.org/10.1016/j.ijpsycho.2011.03.011Get rights and content

Abstract

In the present study we investigate neural network changes after moderate and severe traumatic brain injury (TBI) through the use of resting state functional connectivity (RSFC) methods. Using blood oxygen level dependent functional MRI, we examined RSFC at 3 and 6 months following resolution of posttraumatic amnesia. The goal of this study was to examine how regional off-task connectivity changes during a critical period of recovery from significant neurological disruption. This was achieved by examining regional changes in the intrinsic, or “resting”, BOLD fMRI signal in separate networks: 1) regions linked to goal-directed (or external-state) networks and 2) default mode (or internal-state) networks. Findings here demonstrate significantly increased resting connectivity internal-state networks in the TBI sample during the first 6 months following recovery. The most consistent finding was increased connectivity in both internal and external state networks to the insula and medial temporal regions during recovery. These findings were dissociable from repeat measurements in a matched healthy control sample.

Highlights

► Resting brain connectivity was examined during recovery from brain injury. ► Results reveal changes between “internal” and “goal-directed” networks during recovery. ► The most consistent finding was increased connectivity within the insula. ► Insula connectivity may permit fluid negotiation between internal states and external demands.

Introduction

There is growing interest in the use of functional neuroimaging, and in particular blood oxygen level dependent functional magnetic resonance imaging (BOLD fMRI), to document brain changes associated with traumatic brain injury (TBI). This developing literature has focused primarily on the task-induced changes that differentiate clinical and healthy samples during cognitive, motor and sensory tasks. In one specific literature examining working memory (WM; or the ability to maintain a small amount of information “in mind” for online use) deficits after TBI, several consistent findings have emerged. Investigators have almost universally observed increased involvement of the regions critical for WM, including prefrontal cortex (PFC) and anterior cingulate cortex (ACC) and occasionally parietal regions in TBI (Christodoulou et al., 2001, Hillary et al., 2010, Hillary et al., 2011, McAllister et al., 1999, McAllister et al., 2001, Medaglia et al., 2011, Newsome et al., 2007, Scheibel et al., 2007). There is also rich literature examining the role of frontal systems disruption and the critical contribution of cognitive control to deficit after TBI (McDowell et al., 1997; Hillary et al., 2010, Perlstein et al., 2004, Larson et al., 2006, Larson et al., 2007 Scheibel et al., 2007, Scheibel et al., 2009). These studies offer heuristics as for how large-scale neuronal activity might adapt to TBI, including the potentially critical role of anterior networks and those involved in cognitive (or attentional) control.

While early studies of TBI have helped clarify some of the basic “activation” changes associated with injury and afford the opportunity to link specific cognitive deficits to brain activation changes, there are a number of important future directions for this literature. First, a majority of the studies to date have been cross-sectional observations which pose significant methodological challenges for investigators (e.g., differential task performance between groups) (Price et al., 2006, Price and Friston, 2002) and often do not permit the examination of critical within-subject dynamics. As a remedy to this, the current study makes use of a longitudinal design during a critical window of recovery (i.e., between 3 and 6 months post injury) in order to examine within-subject brain changes.

Second, most studies to date using functional imaging methods to examine the consequences associated with TBI have focused on task-related brain activation. The study of task “activation” offers the opportunity to examine task-specific alterations in neural networks after TBI, but such approaches limit the scope of study (focusing on task induced regions of interest (ROIs), instead of whole brain function) and are burdened by design challenges that are often difficult to resolve. For example, examining task-related activation requires the creation of appropriate control tasks and the need to guarantee equivalent task performance between groups [e.g., task accuracy in TBI vs. healthy control (HC)]. Moreover, the work to date has focused almost solely on the magnitude of the signal (i.e., topographical activation differences) as opposed to communication within the network or covariance in (and between) ROIs.

In pioneering work conducted over 15 years ago, Biswal et al. demonstrated that covariance in voxels comprising the primary motor cortex during rest showed spatial overlap with the observable change in the BOLD response during motor stimulation. These influential findings were the foundation to an intriguing method for investigating brain function through the use of “resting” BOLD data to understand underlying neural connectivity, or “resting state functional connectivity” (RSFC) throughout the brain (Biswal et al., 1995). In fMRI work, RSFC methods often focus on isolating the covariance in very low frequency (~ 0.1 Hz) fluctuations in the BOLD signal, thus permitting analysis of non-task related brain activity which likely plays a non-trivial role in both on-task and off-task functioning.

This early work was extended by Lowe et al. (1997) by demonstrating similar effects in larger regions of sensorimotor cortex (i.e., across multiple slices) and other examiners used these methods to examine relationships between motor and association cortex (Xiong et al., 1998, Xiong et al., 1999). Of critical importance in this literature is the demonstration that task-induced activation maps underestimate the size and number of functionally connected regions and that functional networks are more fully revealed by RSC analysis (Biswal et al., 1995, Xiong et al., 1998, Xiong et al., 1999). These studies established the foundation for “resting-state functional connectivity studies” using fMRI (Biswal et al., 1995, Greicius and Menon, 2004, Gusnard and Raichle, 2001, Hampson et al., 2002, Lowe et al., 1997) and a literature examining task negative or “default mode” networks (Fox et al., 2005; Raichle et al., 2001, Raichle and Snyder, 2007). In the case of the latter, examiners identified distinct “off-task” networks operating in concert as one transitions in and out of goal-directed behavior.

One general interpretation differentiating task-on and task-off networks is that they are reciprocal so that at moments where goal-directed behavior is necessary, the “inward” or self-reflective default mode network remits, giving way to neural activity relevant to task. However, interpreting this relationship as an opponent process may oversimplify this relationship; separate investigations have demonstrated that the default mode activity plays a role in task and the magnitude of deactivation in default mode regions contribute to task performance (Cole et al., 2010, Hampson et al., 2010). These findings offer guiding principles for understanding the role of resting states in healthy neural systems, but questions remain regarding how significant neural network disruption, such as that observed in TBI, might influence the interplay between task-related positive and negative brain activation.

The examination of both RSFC and default-mode networks in clinical samples remains novel, but there are already several findings that provide a framework for understanding how neurological disruption influences the resting signal and for developing expectations in TBI. To date, resting connectivity has been used to examine network changes in a number of clinical disorders including schizophrenia (Camchong et al., 2009, Rotarska-Jagiela et al., 2010, Zhou et al., 2010), normal aging (Koch et al., 2010), stroke (Carter et al., 2010), mood disorders (Chepenik et al., 2010, Hamilton et al., 2010, Sheline et al., 2010), multiple sclerosis (Rocca et al., 2010) and dementias (J. Zhou et al., 2010). There have also been whole brain analyses using resting data to examine alterations in cerebral blood flow (Kim et al., 2010) and we recently applied graph theory to examine “small-worldness” in networks after TBI (Nakamura et al., 2009). In one of the more intriguing applications of resting connectivity to date, Vanhaudenhuyse et al. (2010) used baseline BOLD measures to differentiate cognitively intact and comatose non-communicative brain injured patients. Not surprisingly, the outcome of these studies has varied and this is likely due as much to methodological differences as the effects of distinct pathophysiology in the clinical samples represented. Even so, two important findings emerge from this literature that may be relevant for TBI in the current study. The first is that neurological compromise has been demonstrated to influence resting connectivity (broadly defined). Second, one consequence for global brain connectivity is that connections between critical nodes may be greatly diminished or even unobservable after neurological disruption (see Ongur et al., 2010, Skudlarski et al., 2010, Vanhaudenhuyse et al., 2010).

We anticipate that network disruption results in less coherence in resting connectivity during periods of goal-directed behavior and, therefore, we should observe increased connectivity in internal-state networks during recovery. That is, significant neurological disruption of frontal systems (often observed in TBI), may result in a failure to effectively transition between self-reflective processing and outward goal-directed behavior.

The goal of this study is to examine resting state connectivity to determine if there are systematic changes in whole-brain connectivity during recovery from TBI. We aim to examine the changes in resting fMRI connectivity during the first 6 months following injury when behavioral recovery is known to occur (Millis et al., 2001, Pagulayan et al., 2006). The use of resting fMRI circumvents the methodological dilemmas that arise when using fMRI in clinical samples including difficulty guaranteeing task compliance and assumptions surrounding cognitive subtraction and pure insertion, where contributing components to a task are presumed to be linear and/or additive (Hillary, 2008, Price et al., 2006, Price and Friston, 2002). Moreover, the influences of diffuse neurological disruption (like that observed in TBI) on whole-brain neural networks remain essentially unknown. Traditional fMRI studies in TBI have excluded much of the operating brain in order to isolate specific task-related networks. While this approach offers advantages for examining discrete cognitive deficits, it offers very little information about global brain changes secondary to injury. Consequently, there is very little work documenting how large-scale neural networks adapt to neural disruption and resting connectivity offers one approach to address this issue. For these reasons, resting connectivity is ideal for identifying alterations in the BOLD signal in the recovering brain. This approach may provide additional insight into how neural plasticity is expressed in the injured brain and offer context for findings in cross-sectional activation studies to date, including determining the meaning of increased neural involvement repeatedly observed in activation studies (for review see Hillary, 2008).

Of note, RSFC and the DMN and even the notion of the brain at “rest” have contextual meanings. For the purpose of this study, we will focus on the intrinsic, or resting, BOLD signal during off-task blocks that flank a visual working memory task. In this sense, we are not isolating the off-task “deactivations” that are often the focus in traditional studies of the DMN. Instead, we focus on covariance between four seeded ROIs and the intrinsic BOLD signal during these off-task blocks.

Section snippets

Subjects

Ten participants with moderate and severe TBI between the ages of 19 and 56 years and ten healthy adults of comparable age underwent MRI scanning at separate time points for this study. Individuals with TBI underwent MRI data acquisition at 3 and 6 months after emerging from posttraumatic amnesia (PTA). These study participants were included from an original sample of 15 subjects. The data from five subjects were not included in the current study due to attrition (n = 1), an inability to adequately

Demographic, clinical descriptors

The groups were well-matched for age and gender, but there were significant between-group differences in education (see Table 1).

fMRI results: behavioral data

The current study does not focus on task-related BOLD signal change, however, task performance is a reasonable indicator of cognitive improvement from 3 to 6 months and here we compared the second run of each time point (to permit task acclimation during the first run and minimize the influence of early task practice effects). Compared to the HC sample, accuracy was

Discussion

The current study aimed to examine intrinsic, or “resting” brain connectivity during a period known to be of critical importance for recovery following moderate and severe TBI. The approach used here in a group of individuals sustaining moderate and severe TBI used BOLD fMRI to examine fluctuations in the “off-task” BOLD signal. Primary findings reveal increased involvement of internal-state networks during recovery from TBI (elaborated below). We arrived at this finding by examining two

Study limitations and future directions

The current study holds the advantage of examining TBI during a known window of recovery at separate time points and it is the first to do so using resting connectivity methods. Even so, there are several limitations to this study that require mention. The most significant shortcoming to this study is the small sample size for each of the groups; certainly to conduct sub-group analyses (e.g., injuries to right vs. left hemisphere), additional subjects would be required. In addition, the control

References (75)

  • S. McDowell et al.

    Working memory impairments in traumatic brain injury: evidence from a dual-task paradigm

    Neuropsychologia

    (1997)
  • M. Mennes et al.

    Inter-individual differences in resting-state functional connectivity predict task-induced BOLD activity

    Neuroimage

    (2010)
  • K. Murphy et al.

    The impact of global signal regression on resting state correlations: are anti-correlated networks introduced?

    Neuroimage

    (2009)
  • M.R. Newsome et al.

    Working memory brain activation following severe traumatic brain injury

    Cortex

    (2007)
  • D. Ongur et al.

    Default mode network abnormalities in bipolar disorder and schizophrenia

    Psychiatry Res.

    (2010)
  • K.F. Pagulayan et al.

    A longitudinal study of health-related quality of life after traumatic brain injury

    Arch. Phys. Med. Rehabil.

    (2006)
  • M.E. Raichle et al.

    A default mode of brain function: a brief history of an evolving idea

    Neuroimage

    (2007)
  • A. Rotarska-Jagiela et al.

    Resting-state functional network correlates of psychotic symptoms in schizophrenia

    Schizophr. Res.

    (2010)
  • R. Sanchez-Carrion et al.

    A longitudinal fMRI study of working memory in severe TBI patients with diffuse axonal injury

    Neuroimage

    (2008)
  • P. Skudlarski et al.

    Brain connectivity is not only lower but different in schizophrenia: a combined anatomical and functional approach

    Biol. Psychiatry

    (2010)
  • S.M. Smith et al.

    Advances in functional and structural MR image analysis and implementation as FSL

    Neuroimage

    (2004)
  • R.N. Spreng et al.

    Default network activity, coupled with the frontoparietal control network, supports goal-directed cognition

    Neuroimage

    (2010)
  • G. Teasdale et al.

    Assessment of command impaired consciousness: a practical scale

    Lancet

    (1974)
  • Q. Zou et al.

    Static and dynamic characteristics of cerebral blood flow during the resting state

    Neuroimage

    (2009)
  • M.G. Beaupre et al.

    Cross-cultural emotion recognition among Canadian ethnic groups

    J. Cross Cult. Psy.

    (2005)
  • B.B. Biswal et al.

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

    Magn. Reson. Med.

    (1995)
  • B.B. Biswal et al.

    Toward discovery science of human brain function

    Proc. Natl. Acad. Sci. U. S. A.

    (2010)
  • A. Büki et al.

    All roads lead to disconnection? — Traumatic axonal injury revisited

    Acta Neurochir.

    (2006)
  • J. Camchong et al.

    Altered functional and anatomical connectivity in schizophrenia

    Schizophr. Bull.

    (2009)
  • A.R. Carter et al.

    Resting interhemispheric functional magnetic resonance imaging connectivity predicts performance after stroke

    Ann. Neurol.

    (2010)
  • C. Christodoulou et al.

    Functional magnetic resonance imaging of working memory impairment after traumatic brain injury

    J. Neurol. Neurosurg. Psychiatry

    (2001)
  • A.D. Craig

    How do you feel—now? The anterior insula and human awareness

    Nat. Rev. Neurosci.

    (2009)
  • A. Di Martino et al.

    Functional connectivity of human striatum: a resting state FMRI study

    Cereb. Cortex

    (2008)
  • M.D. Fox et al.

    The human brain is intrinsically organized into dynamic, anticorrelated functional networks

    Proc. Natl. Acad. Sci. U. S. A.

    (2005)
  • M.D. Greicius et al.

    Default-mode activity during a passive sensory task: uncoupled from deactivation but impacting activation

    J. Cogn. Neurosci.

    (2004)
  • M.D. Greicius et al.

    Functional connectivity in the resting brain: a network analysis of the default mode hypothesis

    Proc. Natl. Acad. Sci.

    (2003)
  • D.A. Gusnard et al.

    Searching for a baseline: functional imaging and the resting human brain

    Nat. Rev. Neurosci.

    (2001)
  • Cited by (111)

    • Traumatic brain injury alterations in the functional connectome are associated with neuroinflammation but not tau in a P30IL tauopathy mouse model

      2022, Brain Research
      Citation Excerpt :

      This pattern of decreased functional connectivity post-TBI was consistent with studies in humans of mixed severity TBI (52% prevalence of contusions and diffuse axonal injuries among the studied subjects) (Pandit et al., 2013), mild TBI (Xiong et al., 2016), and repetitive sub-concussive head traumas (Johnson et al., 2014). Other studies have also found increased or mixed increased and decreased functional connectivity (Harris et al., 2016; Hillary et al., 2011; Shumskaya et al., 2017) though the split could be depending on whether TBI subjects had an underlying white matter injury (Sharp et al., 2011). DTI imaging in the same TBI model strongly supported the presence of axonal injury post-TBI (Soni et al., 2021).

    View all citing articles on Scopus
    View full text