Evaluation of Prognosis in Patients with Severe Traumatic Brain Injury Using Resting-State Functional Magnetic Resonance Imaging

Neuroimaging in Disorders of Consciousness and Recovery

There is a clinical need for more accurate diagnosis and prognostication in patients with disorders of consciousness (DoC). There are several neuroimaging modalities that enable detailed, quantitative assessment of structural and functional brain injury, with demonstrated diagnostic and prognostic value. Additionally, longitudinal neuroimaging studies have hinted at quantifiable structural and functional neuroimaging biomarkers of recovery, with potential implications for the management of DoC.

Integrating fractional amplitude of low-frequency fluctuation and functional connectivity to investigate the mechanism and prognosis of severe traumatic brain injury

Objective

Functional magnetic resonance imaging (fMRI) has been used for evaluating residual brain function and predicting the prognosis of patients with severe traumatic brain injury (sTBI). This study aimed to integrate the fractional amplitude of low-frequency fluctuation (fALFF) and functional connectivity (FC) to investigate the mechanism and prognosis of patients with sTBI.

Methods

Sixty-five patients with sTBI were included and underwent fMRI scanning within 14 days after brain injury. The patient's outcome was assessed using the Glasgow Outcome Scale-Extended (GOSE) at 6 months post-injury. Of the 63 patients who met fMRI data analysis standards, the prognosis of 18 patients was good (GOSE scores ≥ 5), and the prognosis of 45 patients was poor (GOSE scores ≤  4). First, we apply fALFF to identify residual brain functional differences in patients who present different prognoses and conjoined it in regions of interest (ROI)-based FC analysis to investigate the residual brain function of sTBI at the acute phase of sTBI. Then, the area under the curve (AUC) was used to evaluate the predictive ability of the brain regions with the difference of fALFF and FC values.

Results

Patients who present good outcomes at 6 months post-injury have increased fALFF values in the Brodmann area (7, 18, 31, 13, 39 40, 42, 19, 23) and decreased FC values in the Brodmann area (28, 34, 35, 36, 20, 28, 34, 35, 36, 38, 1, 2, 3, 4, 6, 13, 40, 41, 43, 44, 20, 28 35, 36, 38) at the acute phase of sTBI. The parameters of these alterations can be used for predicting the long-term outcomes of patients with sTBI, of which the fALFF increase in the temporal lobe, occipital lobe, precuneus, and middle temporal gyrus showed the highest predictive ability (AUC = 0.883).

Conclusion

We provide a compensatory mechanism that several regions of the brain can be spontaneously activated at the acute phase of sTBI in those who present with a good prognosis in the 6-month follow-up, that is, a destructive mode that increases its fALFF in the local regions and weakens its FC to the whole brain. These findings provide a theoretical basis for developing early intervention targets for sTBI patients.

Common Data Elements for Disorders of Consciousness: Recommendations from the Working Group on Neuroimaging

BACKGROUND

Over the past 5 decades, advances in neuroimaging have yielded insights into the pathophysiologic mechanisms that cause disorders of consciousness (DoC) in patients with severe brain injuries. Structural, functional, metabolic, and perfusion imaging studies have revealed specific neuroanatomic regions, such as the brainstem tegmentum, thalamus, posterior cingulate cortex, medial prefrontal cortex, and occipital cortex, where lesions correlate with the current or future state of consciousness. Advanced imaging modalities, such as diffusion tensor imaging, resting-state functional magnetic resonance imaging (fMRI), and task-based fMRI, have been used to improve the accuracy of diagnosis and long-term prognosis, culminating in the endorsement of fMRI for the clinical evaluation of patients with DoC in the 2018 US (task-based fMRI) and 2020 European (task-based and resting-state fMRI) guidelines. As diverse neuroimaging techniques are increasingly used for patients with DoC in research and clinical settings, the need for a standardized approach to reporting results is clear. The success of future multicenter collaborations and international trials fundamentally depends on the implementation of a shared nomenclature and infrastructure.

METHODS

To address this need, the Neurocritical Care Society's Curing Coma Campaign convened an international panel of DoC neuroimaging experts to propose common data elements (CDEs) for data collection and reporting in this field.

RESULTS

We report the recommendations of this CDE development panel and disseminate CDEs to be used in neuroimaging studies of patients with DoC.

CONCLUSIONS

These CDEs will support progress in the field of DoC neuroimaging and facilitate international collaboration.

Alteration in ventral tegmental area and default mode network interplay and prediction of coma recovery in patients with sTBI

Purpose

To investigate the role of VTA and DMN in modulating human consciousness in patient with sTBI.

Methods

We mapped an atlas of VTA in the brainstem and a total of 19 region of interests in the ventral and dorsal DMN onto functional magnetic resonance imaging in 28 patients with sTBI and 28 healthy controls. We assessed the functional connectivity alteration in subcortical VTA and cortical DMN nodes in patients of coma. We evaluated the spatially distribution of FC alteration in VTA and DMN nodes after sTBI and evaluated their predictive value for coma recovery.

Results

There was a decrease in FC between VTA and DMN in patients compared to controls. After decomposition, the FC between VTA and 10 DMN nodes were decreased whereas the FC within 2 DMN nodes were increased in patients with acute coma. The FC alteration in DMN nodes provided useful information for the early prediction of 6-month coma recovery in patients with sTBI.

Conclusions

We provide initial evidence for the decreased FC between VTA and massive DMN nodes in patients with coma in acute phase of sTBI. We found that the FC alteration within DMN is more useful than the FC alteration between VTA and DMN for predicting coma recovery in patients with sTBI. VTA and DMN connectivity mapping provides an opportunity to advance the cortical-subcortical mechanism of human consciousness.

Identification of chronic mild traumatic brain injury using resting state functional MRI and machine learning techniques

Mild traumatic brain injury (mTBI) is a major public health concern that can result in a broad spectrum of short-term and long-term symptoms. Recently, machine learning (ML) algorithms have been used in neuroscience research for diagnostics and prognostic assessment of brain disorders. The present study aimed to develop an automatic classifier to distinguish patients suffering from chronic mTBI from healthy controls (HCs) utilizing multilevel metrics of resting-state functional magnetic resonance imaging (rs-fMRI). Sixty mTBI patients and forty HCs were enrolled and allocated to training and testing datasets with a ratio of 80:20. Several rs-fMRI metrics including fractional amplitude of low-frequency fluctuation (fALFF), regional homogeneity (ReHo), degree centrality (DC), voxel-mirrored homotopic connectivity (VMHC), functional connectivity strength (FCS), and seed-based FC were generated from two main analytical categories: local measures and network measures. Statistical two-sample t-test was employed comparing between mTBI and HCs groups. Then, for each rs-fMRI metric the features were selected extracting the mean values from the clusters showing significant differences. Finally, the support vector machine (SVM) models based on separate and multilevel metrics were built and the performance of the classifiers were assessed using five-fold cross-validation and via the area under the receiver operating characteristic curve (AUC). Feature importance was estimated using Shapley additive explanation (SHAP) values. Among local measures, the range of AUC was 86.67-100% and the optimal SVM model was obtained based on combined multilevel rs-fMRI metrics and DC as a separate model with AUC of 100%. Among network measures, the range of AUC was 80.42-93.33% and the optimal SVM model was obtained based on the combined multilevel seed-based FC metrics. The SHAP analysis revealed the DC value in the left postcentral and seed-based FC value between the motor ventral network and right superior temporal as the most important local and network features with the greatest contribution to the classification models. Our findings demonstrated that different rs-fMRI metrics can provide complementary information for classifying patients suffering from chronic mTBI. Moreover, we showed that ML approach is a promising tool for detecting patients with mTBI and might serve as potential imaging biomarker to identify patients at individual level.

Clinical trial registration

[clinicaltrials.gov], identifier [NCT03241732].

Aberrant brain functional hubs convergence in the acute severe traumatic brain injury patients with rapidly recovering

PURPOSE

We aimed to identify the aberrant functional hubs in patients with acute severe traumatic brain injury (sTBI) and investigate whether they could help inform prognosis.

METHODS

Twenty-eight sTBI patients and health controls underwent imaging scanning. The graph-theoretical measure of degree centrality (DC) was applied to identify the abnormal brain functional hubs and conjoined with regions of interest-based analysis to investigate their interaction and impact on whole-brain. We further split sTBI patients into two subgroups according to their recovery to explore whether the fractional amplitude of low-frequency fluctuation (fALFF) roles in functional connectivity (FC) differential areas to help inform the patients' long-term prognosis.

RESULTS

We identified the part of prefrontal cortex (PFC), precentral and postcentral gyrus (Pre-/Post-CG), cingulate gyrus (CgG), posterior medial cortex (PMC), and brainstem that could be core hubs whose DC was significantly increased in patients with acute sTBI. The interaction strength of the paired hubs could be enhanced (CG-PFC, CgG-PFC, CG-brainstem, CgG-brainstem, PMC-brainstem, and PFC-brainstem) and weakened (CG-CgG, CG-PMC, CgG-PMC, and PMC-PFC), compared with healthy controls. We also found abnormal FC in 5 hubs to whole-brain. The spontaneous brain activities in the FC differential regions [e.g., the fALFF and mean fALFF value] were valid to predict outcome at 6-month in patients with sTBI.

CONCLUSION

We demonstrated a compensatory mechanism that part of brain regions will converge into abnormal functional hubs in patients with acute sTBI, which provides a potential approach to objectively predicting patients' long-term outcome.

Correcting cardiorespiratory noise in resting-state functional MRI data acquired in critically ill patients

Resting-state functional MRI is being used to develop diagnostic, prognostic and therapeutic biomarkers for critically ill patients with severe brain injuries. In studies of healthy volunteers and non-critically ill patients, prospective cardiorespiratory data are routinely collected to remove non-neuronal fluctuations in the resting-state functional MRI signal during analysis. However, the feasibility and utility of collecting cardiorespiratory data in critically ill patients on a clinical MRI scanner are unknown. We concurrently acquired resting-state functional MRI (repetition time = 1250 ms) and cardiac and respiratory data in 23 critically ill patients with acute severe traumatic brain injury and in 12 healthy control subjects. We compared the functional connectivity results from two approaches that are commonly used to correct cardiorespiratory noise: (i) denoising with cardiorespiratory data (i.e. image-based method for retrospective correction of physiological motion effects in functional MRI) and (ii) standard bandpass filtering. Resting-state functional MRI data in 7 patients could not be analysed due to imaging artefacts. In 6 of the remaining 16 patients (37.5%), cardiorespiratory data were either incomplete or corrupted. In patients (n = 10) and control subjects (n = 10), the functional connectivity results corrected with the image-based method for retrospective correction of physiological motion effects in functional MRI did not significantly differ from those corrected with bandpass filtering of 0.008-0.125 Hz. Collectively, these findings suggest that, in critically ill patients with severe traumatic brain injury, there is limited feasibility and utility to denoising the resting-state functional MRI signal with prospectively acquired cardiorespiratory data.

Investigating the mechanism and prognosis of patients with disorders of consciousness on the basis of brain networks between the thalamus and whole-brain

Purpose

This study aimed to investigate the changes in the functional connectivity between the bilateral thalamus and the whole-brain in patients with severe traumatic brain injury (sTBI) patients suffering from disorders of consciousness (DOC) and to explore their potential prognostic representation capacity.

Methods

The sTBI patients suffering from DOC and healthy controls underwent functional magnetic resonance imaging. We defined patients with the Extended Glasgow Outcome Score (GOS-E) ≥ 3 as the wake group and GOS-E = 2 as the coma group. The differences in functional connectivity between sTBI and healthy controls and between wake and coma groups were compared. Based on the brain regions with altered functional connectivity between wake and coma groups, they were divided into 26 regions of interest. Based on the Z-values of regions of interest, the receiver operating characteristic analysis was conducted to classify the prognosis of patients.

Results

A total of 28 patients and 15 healthy controls were finally included. Patients who had DOC indicated a significant disruption of functional connectivity between the bilateral thalamus and the whole-brain (FDR corrected, P < 0.0007). The functional connectivity strength (bilateral thalamus to whole-brain) was significantly different between coma patients who went on to wake and those who were eventually non-awake at 6 months after sTBI (Alphasim corrected, P < 0.05). Furthermore, the 26 regions of interest had a similar or even better prognostic distinction ability than the admission Glasgow coma score.

Conclusion

The thalamus-based system of consciousness of sTBI patients suffering from DOC is disrupted. There are differences in the thalamus-to-whole-brain network between wake and coma groups and these differences have potential prognostic characterization capability.

Applications of Advanced MRI to Disorders of Consciousness

Disorder of consciousness (DoC) after severe brain injury presents numerous challenges to clinicians, as the diagnosis, prognosis, and management are often uncertain. Magnetic resonance imaging (MRI) has long been used to evaluate brain structure in patients with DoC. More recently, advances in MRI technology have permitted more detailed investigations of the brain's structural integrity (via diffusion MRI) and function (via functional MRI). A growing literature has begun to show that these advanced forms of MRI may improve our understanding of DoC pathophysiology, facilitate the identification of patient consciousness, and improve the accuracy of clinical prognostication. Here we review the emerging evidence for the application of advanced MRI for patients with DoC.

Abnormal default mode network could be a potential prognostic marker in patients with disorders of consciousness

OBJECTIVES

The study aimed to investigate disorders of consciousness (DOC) mechanisms of patients with severe traumatic brain injury (sTBI) related to default mode network (DMN) and to introduce a machine learning model that predicts the prognosis of these patients for 6 months.

METHODS

The sTBI patients suffering from DOC and healthy controls underwent functional magnetic resonance imaging. We defined patients with Extended Glasgow Outcome Score ≥ 5 as good outcome group, otherwise they were poor outcome group. The differences of DMN between sTBI and healthy controls and between good and poor outcome groups were compared. Based on the brain regions with altered functional connectivity between good and poor outcome groups, they were divided into 8 regions of interests according to side. The Z values of the regions of interests were extracted by Rest 1.8. Based on Z values, the Subspace K-Nearest Neighbor (Subspace KNN) was conducted to classify prognosis of sTBI patients suffering from DOC.

RESULTS

A total of 84 DMNs derived from patients and 45 DMNs from healthy controls were finally analyzed. The connectivity of the DMN was significantly decreased in sTBI patients suffering from DOC (Alphasim corrected, P < 0.05). In addition, compared with the poor outcome group (DMN samples = 60), the brain regions of DMN with decreased functional connectivity in the good outcome group (DMN samples = 24) the following bilateral areas: brodman Area 11, anterior cingulate and paracingulate gyri, brodman Area 25, olfactory cortex (Alphasim corrected, P < 0.05). The ability of Subspace KNN machine learning to distinguish the prognosis of patients (area under curve) was 0.97.

CONCLUSIONS

The interruption of DMN may be one of the reasons for DOC in patients with sTBI. Furthermore, based on early DMN (1-4 weeks), Subspace KNN machine learning has the potential value to distinguish the prognosis (6 months after brain trauma) of sTBI patients suffering from DOC.

Resting-State Functional MRI Metrics in Patients With Chronic Mild Traumatic Brain Injury and Their Association With Clinical Cognitive Performance

Mild traumatic brain injury (mTBI) accounts for more than 80% of people experiencing brain injuries. Symptoms of mTBI include short-term and long-term adverse clinical outcomes. In this study, resting-state functional magnetic resonance imaging (rs-fMRI) was conducted to measure voxel-based indices including fractional amplitude of low-frequency fluctuation (fALFF), regional homogeneity (ReHo), and functional connectivity (FC) in patients suffering from chronic mTBI; 64 patients with chronic mTBI at least 3 months post injury and 40 healthy controls underwent rs-fMRI scanning. Partial correlation analysis controlling for age and gender was performed within mTBI cohort to explore the association between rs-fMRI metrics and neuropsychological scores. Compared with controls, chronic mTBI patients showed increased fALFF in the left middle occipital cortex (MOC), right middle temporal cortex (MTC), and right angular gyrus (AG), and increased ReHo in the left MOC and left posterior cingulate cortex (PCC). Enhanced FC was observed from left MOC to right precuneus; from right MTC to right superior temporal cortex (STC), right supramarginal, and left inferior parietal cortex (IPC); and from the seed located at right AG to left precuneus, left superior medial frontal cortex (SMFC), left MTC, left superior temporal cortex (STC), and left MOC. Furthermore, the correlation analysis revealed a significant correlation between neuropsychological scores and fALFF, ReHo, and seed-based FC measured from the regions with significant group differences. Our results demonstrated that alterations of low-frequency oscillations in chronic mTBI could be representative of disruption in emotional circuits, cognitive performance, and recovery in this cohort.

Improving Diagnosis and Prognosis in Acute Severe Brain Injury: A Multimodal Imaging Protocol

Multi-modal neuroimaging techniques have the potential to dramatically improve the diagnosis of the level consciousness and prognostication of neurological outcome for patients with severe brain injury in the intensive care unit (ICU). This protocol describes a study that will utilize functional Magnetic Resonance Imaging (fMRI), electroencephalography (EEG), and functional Near Infrared Spectroscopy (fNIRS) to measure and map the brain activity of acute critically ill patients. Our goal is to investigate whether these modalities can provide objective and quantifiable indicators of good neurological outcome and reliably detect conscious awareness. To this end, we will conduct a prospective longitudinal cohort study to validate the prognostic and diagnostic utility of neuroimaging techniques in the ICU. We will recruit 350 individuals from two ICUs over the course of 7 years. Participants will undergo fMRI, EEG, and fNIRS testing several times over the first 10 days of care to assess for residual cognitive function and evidence of covert awareness. Patients who regain behavioral awareness will be asked to complete web-based neurocognitive tests for 1 year, as well as return for follow up neuroimaging to determine which acute imaging features are most predictive of cognitive and functional recovery. Ultimately, multi-modal neuroimaging techniques may improve the clinical assessments of patients' level of consciousness, aid in the prediction of outcome, and facilitate efforts to find interventional methods that improve recovery and quality of life.

Altered resting-state functional connectivity within the developing social brain after pediatric traumatic brain injury

Traumatic brain injury (TBI) in childhood and adolescence can interrupt expected development, compromise the integrity of the social brain network (SBN) and impact social skills. Yet, no study has investigated functional alterations of the SBN following pediatric TBI. This study explored functional connectivity within the SBN following TBI in two independent adolescent samples. First, 14 adolescents with mild complex, moderate or severe TBI and 16 typically developing controls (TDC) underwent resting‐state functional magnetic resonance imaging 12–24 months post‐injury. Region of interest analyses were conducted to compare the groups' functional connectivity using selected SBN seeds. Then, replicative analysis was performed in an independent sample of adolescents with similar characteristics (9 TBI, 9 TDC). Results were adjusted for age, sex, socioeconomic status and total gray matter volume, and corrected for multiple comparisons. Significant between‐group differences were detected for functional connectivity in the dorsomedial prefrontal cortex and left fusiform gyrus, and between the left fusiform gyrus and left superior frontal gyrus, indicating positive functional connectivity for the TBI group (negative for TDC). The replication study revealed group differences in the same direction between the left superior frontal gyrus and right fusiform gyrus. This study indicates that pediatric TBI may alter functional connectivity of the social brain. Frontal‐fusiform connectivity has previously been shown to support affect recognition and changes in the function of this network could relate to more effortful processing and broad social impairments.

Default Mode Network Connectivity Predicts Emotion Recognition and Social Integration After Traumatic Brain Injury

Moderate-severe traumatic brain injury (TBI) may result in difficulty with emotion recognition, which has negative implications for social functioning. As aspects of social cognition have been linked to resting-state functional connectivity (RSFC) in the default mode network (DMN), we sought to determine whether DMN connectivity strength predicts emotion recognition and level of social integration in TBI. To this end, we examined emotion recognition ability of 21 individuals with TBI and 27 healthy controls in relation to RSFC between DMN regions. Across all participants, decreased emotion recognition ability was related to increased connectivity between dorsomedial prefrontal cortex (dmPFC) and temporal regions (temporal pole and parahippocampal gyrus). Furthermore, within the TBI group, connectivity between dmPFC and parahippocampal gyrus predicted level of social integration on the Community Integration Questionnaire, an important index of post-injury social functioning in TBI. This finding was not explained by emotion recognition ability, indicating that DMN connectivity predicts social functioning independent of emotion recognition. These results advance our understanding of the neural underpinnings of emotional and social processes in both healthy and injured brains, and suggest that RSFC may be an important marker of social outcomes in individuals with TBI.