Functional and Structural Integrity of Frontoparietal Connectivity in Traumatic and Anoxic Coma

Abstract

Objectives: Recovery from coma might critically depend on the structural and functional integrity of frontoparietal networks. We aimed to measure this integrity in traumatic brain injury and anoxo-ischemic (cardiac arrest) coma patients by using an original multimodal MRI protocol.

Design: Prospective cohort study.

Setting: Three Intensive Critical Care Units affiliated to the University in Toulouse (France).

Patients: We longitudinally recruited 43 coma patients (Glasgow Coma Scale at the admission < 8; 29 cardiac arrest and 14 traumatic brain injury) and 34 age-matched healthy volunteers. Exclusion criteria were disorders of consciousness lasting more than 30 days and focal brain damage within the explored brain regions. Patient assessments were conducted at least 2 days (5 ± 2 d) after complete withdrawal of sedation. All patients were followed up (Coma Recovery Scale-Revised) 3 months after acute brain injury.

Interventions: None.

Measurements and main results: Functional and structural MRI data were recorded, and the analysis was targeted on the posteromedial cortex, the medial prefrontal cortex, and the cingulum. Univariate analyses and machine learning techniques were used to assess diagnostic and predictive values. Coma patients displayed significantly lower medial prefrontal cortex-posteromedial cortex functional connectivity (area under the curve, 0.94; 95% CI, 0.93-0.95). Cardiac arrest patients showed specific structural disturbances within posteromedial cortex. Significant cingulum architectural disturbances were observed in traumatic brain injury patients. The machine learning medial prefrontal cortex-posteromedial cortex multimodal classifier had a significant predictive value (area under the curve, 0.96; 95% CI, 0.95-0.97), best combination of subregions that discriminates a binary outcome based on Coma Recovery Scale-Revised).

Conclusions: This exploratory study suggests that frontoparietal functional disconnections are specifically observed in coma and their structural counterpart provides information about brain injury mechanisms. Multimodal MRI biomarkers of frontoparietal disconnection predict 3-month outcome in our sample. These findings suggest that fronto-parietal disconnection might be particularly relevant for coma outcome prediction and could inspire innovative precision medicine approaches.

Figure 1.

Functional and structural regions of interests (ROIs). A, Medial prefrontal cortex (mPFC) and posteromedial cortex (PMC). Leftmost shows a 3D rendering of the whole mPFC and PMC ROIs. Rightmost shows the outline of the 11 mPFC subregions and the three PMC subregions from the Willard atlas on a standard T1 template. B, Cingulum. Leftmost shows a 3D rendering of the whole cingulum as defined (10). Rightmost shows the outline of the five cingulum subregions (see methods).

Figure 2.

Univariate analysis of regions of interests (ROIs) subregions. A, Cortex: boxplot of the two indexes pertaining to the cortex, gray matter density (GMD) and mean diffusivity (MD), broken down by medial prefrontal cortex (mPFC) and posteromedial cortex (PMC) subregions. MD values were multiplied by a constant for easiness of visualization and comparison with GMD values. B, Cingulum: boxplot of the three indexes pertaining to the cingulum, fractional anisotropy (FA), MD, and radial diffusivity (RD) broken down by mPFC and PMC subregions. FA, MD, and RD values were Z transformed for easiness of visualization and comparisons between indexes. C, Functional connectivity (FC) (measured as Fisher-transformed Pearson r) between the mPFC and PMC subregions. *Significant diagnosis effect (healthy controls vs coma); #Significant etiology effect (traumatic brain injury [TBI] vs anoxic).

Figure 3.

Discrimination: brain injury mechanism. A, Two indexes/regions of interests (ROIs) that led to the best etiology classification. Structural indexes (in this case radial diffusivity [RD] and mean diffusivity [MD]) are represented by a full colored ROI. B, Decision surface of the best model (logistic regression). C, Fifteen most selected features (out of 100 repetitions) in the model combining all features. Please note that the five most relevant features (dashed red box) are exclusively structural parameters. FC = functional connectivity, mPFC = medial prefrontal cortex, PMC = posteromedial cortex, TBI = traumatic brain injury, VBM = voxel-based morphometry.

Figure 4.

Prediction: neurologic outcome. A, Leftmost shows the two indexes/regions of interests (ROIs) that led to the best prediction of the outcome as measured by Coma Recovery Scale-Revised (CRS-R). Functional connectivity (FC) between two regions is represented using ROIs outline of the same color. Central shows the scatterplot of the relationship between CRS-R predictions obtained with the best model (linear regression) and actual CRS-R. Rightmost shows the most selected features (out of 100 repetitions) in the model combining all features (support vector regression). B, Leftmost shows the two indexes/ROIs that led to the best prediction of the binary outcome. FC between two regions is represented using ROIs outline of the same color. Central shows the decision surface of the best model (logistic regression), with the color reflecting the variable using for binary prediction: recovered (minimally conscious state [MCS]– and MCS+) in blue and not recovered (vegetative state [VS]/unresponsive wakefulness syndrome [UWS]) in red. The marker shape reflecting the specific 3 mo clinical status as VS/UWS (circle), MCS– (triangle), and MCS+ (square). Rightmost shows the most selected features (out of 100 repetitions) in the model combining all features. Please note that the five most relevant features (red dashed box) are FC parameters. mPFC = medial prefrontal cortex, PMC = posteromedial cortex.