Regional cerebral metabolic patterns demonstrate the role of anterior forebrain mesocircuit dysfunction in the severely injured brain

Abstract

Although disorders of consciousness (DOCs) demonstrate widely varying clinical presentations and patterns of structural injury, global down-regulation and bilateral reductions in metabolism of the thalamus and frontoparietal network are consistent findings. We test the hypothesis that global reductions of background synaptic activity in DOCs will associate with changes in the pattern of metabolic activity in the central thalamus and globus pallidus. We compared 32 [(18)F]fluorodeoxyglucose PETs obtained from severely brain-injured patients (BIs) and 10 normal volunteers (NVs). We defined components of the anterior forebrain mesocircuit on high-resolution T1-MRI (ventral, associative, and sensorimotor striatum; globus pallidus; central thalamus and noncentral thalamus). Metabolic profiles for BI and NV demonstrated distinct changes in the pattern of uptake: ventral and association striatum (but not sensorimotor) were significantly reduced relative to global mean uptake after BI; a relative increase in globus pallidus metabolism was evident in BI subjects who also showed a relative reduction of metabolism in the central thalamus. The reversal of globus pallidus and central thalamus profiles across BIs and NVs supports the mesocircuit hypothesis that broad functional (or anatomic) deafferentation may combine to reduce central thalamus activity and release globus pallidus activity in DOCs. In addition, BI subjects showed broad frontoparietal metabolic down-regulation consistent with prior studies supporting the link between central thalamic/pallidal metabolism and down-regulation of the frontoparietal network. Recovery of left hemisphere frontoparietal metabolic activity was further associated with command following.

Keywords: fronto-striato-thalamic circuit; minimally conscious state; thalamocortical loops; vegetative state.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.

(A and B) Axial anatomic high-resolution MRI in representative NV and BI subjects, respectively, showing locations and manual delimitation for ventral, associative, and sensorimotor striatum (VST, AST, and SMST, respectively), globus pallidus (GP), central and noncentral thalamus (c-TH and non-c-TH, respectively), and color scheme for regions of interest (VST, yellow; AST, green; SMST, red; GP, white; c-TH, magenta; non-c-TH, violet). (C and D) [18F]FDG-PET, T1-MRI fusion in in representative NV and BI subject, respectively. NV demonstrates symmetric pattern of relatively increased c-TH metabolism compared with non-c-TH. A marked asymmetry of thalamic metabolism with loss of contrast in c-TH compared with non-c-TH metabolism is evident in BI subject.

Fig. 2.

Group data displaying mn-UV of glucose metabolism in deep brain structures measured in NV and BI subjects. (A) Box plot. A significant reduction in relative glucose metabolism of the ventral striatum (VST), associative striatum (AST), and central thalamus (c-TH) in BI subjects is seen compared with NV. No difference in sensorimotor striatum (SMST) mn-UV is present between NV and BI subjects. A significant increase in globus pallidus (GP) metabolism is present in the group of BI subjects. (B) Contrast showing the results of the post hoc Bonferroni corrected analysis of a two-way ANOVA (group and ROI). Significant results are shown in blue for NV and red for BI patients, whereas white boxes denote no significant differences; arrows indicate the direction of the significance (i.e., pointing toward the higher mn-UV values). (C) Bivariate scattergram demonstrates an inverse linear correlation between glucose metabolic rate of the c-TH (x axis) and the GP (y axis), P < 0.001 (between groups).

Fig. 3.

Group data displaying mn-UV of regional cortical glucose metabolism in NV and BI subjects. (A) Box plot for the left hemisphere cortical regions showing a significant reduction of glucose metabolism across anterior frontal cortices, medial cingulate cortex (mCGc)/precuneus, and posterior parietal cortices in BI subjects. Note a distinct pattern of rostrocaudal progression of increasing mn-UV glucose metabolic values in BI patients as arranged [orbitofrontal, ventromedial, and dorsolateral prefrontal cortices, premotor cortices and primary sensorimotor (OFC/vmPFC/dlPFc/PMC/SM1, respectively)] that converges with NV mn-UV at SM1. BI subjects demonstrate significant decreases in superior posterior parietal cortex (sPPC), posterior cingulate cortex (pCGc), precuneus, and V1. (B) Box plot for the right hemisphere showing similar pattern of rostrocaudal progression of increasing mn-UV glucose metabolic values in BI patients as arranged (OFC/vmPFC/ dlPFc/PMC/SM1). Note that right pCGc does not show differences in observed for left hemisphere in posterior structures (for significance, see Table S1).

Fig. 4.

Group data displaying mn-UV of glucose metabolism for deep brain and cortical structures comparing subsets of severe BI patients demonstrating command following (CF; green), those in the noncommand following (N-CF) group (gray), and NV (blue). (A) Box plot of two deep brain structures showing a significant difference in glucose metabolism. Increases of mn-UV are seen in c-TH for the CF group compared with N-CF, whereas decreases in VST are present. (B) Box plot of the ratio c-TH/GP of raw standard uptake values body weight normalized (i_c-TH/GP); index separates the CF and N-CF groups with increased ratio in CF group that remains lower than NV. (C) Box plot of four cortical structures showing a significant difference in glucose metabolism between the two groups. A significant increase of glucose metabolism in the CF group is present for SM1, iPPc, and sPPc in the left hemisphere. An opposite metabolic pattern is seen for the right OFC, which shows a reduction of metabolism for CF group. This finding is consistent with that observed in the VST (A).