Whole-brain glucose metabolic pattern differentiates minimally conscious state from unresponsive wakefulness syndrome.

AIMS: The patient being minimally conscious state (MCS) may benefit from wake-up interventions aimed at improving quality of life and have a higher probability of recovering higher level of consciousness compared to patients with the unresponsive wakefulness syndrome (UWS). However, differentiation of the MCS and UWS poses challenge in clinical practice. This study aimed to explore glucose metabolic pattern (GMP) obtained from 18F-labeled-fluorodeoxyglucose positron emission tomography (18F-FDG-PET) in distinguishing between UWS and MCS. METHODS: Fifty-seven patients with disorders of consciousness (21 cases of UWS and 36 cases of MCS) who had undergone repeated standardized Coma Recovery Scale-Revised (CRS-R) evaluations were enrolled in this prospective study. 18F-FDG-PET was carried out in all patients and healthy controls (HCs). Voxel-based scaled subprofile model/principal component analysis (SSM/PCA) was used to generate GMPs. The expression score of whole-brain GMP was obtained, and its diagnostic accuracy was compared with the standardized uptake value ratio (SUVR). The diagnostic efficiency was validated by one-year later clinical outcomes. RESULTS: UWS-MCS GMP exhibited hypometabolism in the frontal-parietal cortex, along with hypermetabolism in the unilateral lentiform nucleus, putamen, and anterior cingulate gyrus. The UWS-MCS-GMP expression score was significantly higher in UWS compared to MCS patients (0.90 ± 0.85 vs. 0 ± 0.93, p < 0.001). UWS-MCS-GMP expression score achieved an area under the curve (AUC) of 0.77 to distinguish MCS from UWS, surpassing that of SUVR based on the frontoparietal cortex (AUC = 0.623). UWS-MCS-GMP expression score was significantly correlated with the CRS-R score (r = -0.45, p = 0.004) and accurately predicted the one-year outcome in 73.7% of patients. CONCLUSION: UWS and MCS exhibit specific glucose metabolism patterns, the UWS-MCS-GMP expression score significantly distinguishes MCS from UWS, making SSM/PCA a potential diagnostic methods in clinical practice for individual patients.

Diagnostic value of PET imaging in clinically unresponsive patients.

Rapid advancements in the critical care management of acute brain injuries have facilitated the survival of numerous patients who may have otherwise succumbed to their injuries. The probability of conscious recovery hinges on the extent of structural brain damage and the level of metabolic and functional cerebral impairment, which remain challenging to assess via laboratory, clinical, or functional tests. Current research settings and guidelines highlight the potential value of fluorodeoxyglucose-PET (FDG-PET) for diagnostic and prognostic purposes, emphasizing its capacity to consistently illustrate a metabolic reduction in cerebral glucose uptake across various disorders of consciousness. Crucially, FDG-PET might be a pivotal tool for differentiating between patients in the minimally conscious state and those in the unresponsive wakefulness syndrome, a persistent clinical challenge. In patients with disorders of consciousness, PET offers utility in evaluating the degree and spread of functional disruption, as well as identifying irreversible neural damage. Further, studies that capture responses to external stimuli can shed light on residual or revived brain functioning. Nevertheless, the validity of these findings in predicting clinical outcomes calls for additional long-term studies with larger patient cohorts suffering from consciousness impairment. Misdiagnosis of conscious illnesses during bedside clinical assessments remains a significant concern. Based on the clinical research settings, current clinical guidelines recommend PET for diagnostic and/or prognostic purposes. This review article discusses the clinical categories of conscious disorders and the diagnostic and prognostic value of PET imaging in clinically unresponsive patients, considering the known limitations of PET imaging in such contexts.

Review of spinal cord stimulation for disorders of consciousness.

PURPOSE OF REVIEW: High-cervical spinal cord stimulation can alter cortical activity and cerebral metabolism. These effects are potentially beneficial for disorders of consciousness. A better understanding of the effects of clinical application of stimulation is needed. We aimed to evaluate the existing literature to determine the state of available knowledge. We performed a literature review of clinical studies assessing cervical spinal cord epidural stimulation for disorders of consciousness. Only peer-reviewed articles reporting preoperative and postoperative clinical status were included. RECENT FINDINGS: Nineteen studies were included. A total of 532 cases were reported, and 255 patients were considered responsive (47.9%). Considering only studies published after the definition of minimally conscious state (MCS) as an entity, 402 individuals in unresponsive wakefulness syndrome (UWS) and 113 in MCS were reported. Responsiveness to SCS was reported in 170 UWS patients (42.3%) and in 78 MCS cases (69.0%), although the criteria for responsiveness and outcome measures varied among publications. SUMMARY: Cervical SCS yielded encouraging results in patients with disorders of consciousness and seems to be more effective in MCS. More extensive investigation is needed to understand its potential role in clinical practice.

A Family of Spore Lipoproteins Stabilizes the Germination Apparatus by Altering Inner Spore Membrane Fluidity in Bacillus subtilis Spores.

Dormant bacterial spores undergo the process of germination to return to a vegetative state. In most species, germination involves the sensing of nutrient germinants, the release of various cations and a calcium-dipicolinic acid (DPA) complex, spore cortex degradation, and full rehydration of the spore core. These steps are mediated by membrane-associated proteins, and all these proteins have exposure on the outer surface of the membrane, a hydrated environment where they are potentially subject to damage during dormancy. A family of lipoproteins, including YlaJ, which is expressed from the sleB operon in some species, are present in all sequenced Bacillus and Clostridium genomes that contain sleB. B. subtilis possesses four proteins in this family, and prior studies have demonstrated two of these are required for efficient spore germination and these proteins contain a multimerization domain. Genetic studies of strains lacking all combinations of these four genes now reveal all four play roles in ensuring efficient germination, and affect multiple steps in this process. Electron microscopy does not reveal significant changes in spore morphology in strains lacking lipoproteins. Generalized polarization measurements of a membrane dye probe indicate the lipoproteins decrease spore membrane fluidity. These data suggest a model in which the lipoproteins form a macromolecular structure on the outer surface of the inner spore membrane, where they act to stabilize the membrane and potentially interact with other germination proteins, and thus stabilize the function of multiple components of the germination machinery. IMPORTANCE Bacterial spores exhibit extreme longevity and resistance to many killing agents, and are thus problematic agents of several diseases and of food spoilage. However, to cause disease or spoilage, germination of the spore and return to the vegetative state is necessary. The proteins responsible for initiation and progression of germination are thus potential targets for spore-killing processes. A family of membrane-bound lipoproteins that are conserved across most spore-forming species was studied in the model organism Bacillus subtilis. The results indicate that these proteins reduce the membrane fluidity and increase the stability of other membrane associated proteins that are required for germination. Further understanding of such protein interactions on the spore membrane surface will enhance our understanding of the germination process and its potential as a decontamination method target.

Brain Metabolism but Not Gray Matter Volume Underlies the Presence of Language Function in the Minimally Conscious State (MCS): MCS+ Versus MCS- Neuroimaging Differences.

Background. The minimally conscious state (MCS) is subcategorized into MCS- and MCS+, depending on the absence or presence, respectively, of high-level behavioral responses such as command-following. Objective. We aim to investigate the functional and structural neuroanatomy underlying the presence of these responses in MCS- and MCS+ patients. Methods. In this cross-sectional retrospective study, chronic MCS patients were diagnosed using repeated Coma Recovery Scale-Revised assessments. Fluorodeoxyglucose-positron emission tomography data were acquired on 57 patients (16 MCS-; 41 MCS+) and magnetic resonance imaging with voxel-based morphometry analysis was performed on 66 patients (17 MCS-; 49 MCS+). Brain glucose metabolism and gray matter integrity were compared between patient groups and control groups. A metabolic functional connectivity analysis testing the hypothesis of preserved language network in MCS+ compared with MCS- was also done. Results. Patients in MCS+ presented higher metabolism mainly in the left middle temporal cortex, known to be important for semantic processing, compared with the MCS- group. The left angular gyrus was also functionally disconnected from the left prefrontal cortex in MCS- compared with MCS+ group. No significant differences were found in gray matter volume between patient groups. Conclusions. The clinical subcategorization of MCS is supported by differences in brain metabolism but not in gray matter structure, suggesting that brain function in the language network is the main support for recovery of command-following, intelligible verbalization and/or intentional communication in the MCS. Better characterizing the neural correlates of residual cognitive abilities of MCS patients contributes to reduce their misdiagnosis and to adapt therapeutic approaches.

Neural correlates of different behavioral response to transcranial direct current stimulation between patients in the unresponsive wakefulness syndrome and minimally conscious state.

In this study, we used event-related potential (ERP) and 18F-fluorodeoxyglucose-positron emission tomography (FDG-PET) to study the neural correlates of different behavioral response to transcranial direct current stimulation (tDCS) between patients in unresponsive wakefulness syndrome (UWS) and minimally conscious state (MCS). Thirteen patients (eight in UWS and five in MCS) underwent 20 anodal tDCS sessions of the left dorsolateral prefrontal cortex (DLPFC). Before tDCS, all the patients and six age-matched healthy subjects underwent a cerebral FDG-PET scan and ERP test. The coma recovery scale-revised (CRS-R) results revealed that after tDCS, a significant improvement was observed only in the MCS group. The ERP results supported that MCS patients preserved more high-order cortical information processing capacities. The residual brain metabolism in the left DLPFC in MCS patients supported that a residual brain activity in the stimulated area was necessary for a behavioral response to tDCS. Our study also demonstrated that the cerebral metabolic rates of glucose (CMRgl) ratios in intrinsic network were correlated significantly with CRS-R in MCS patients. In addition, the right prefrontal region might be another potential therapeutic target for MCS patients.

Presynaptic dopamine deficit in minimally conscious state patients following traumatic brain injury.

Dopaminergic stimulation has been proposed as a treatment strategy for post-traumatic brain injured patients in minimally conscious state based on a clinical trial using amantadine, a weak dopamine transporter blocker. However, a specific contribution of dopaminergic neuromodulation in minimally conscious state is undemonstrated. In a phase 0 clinical trial, we evaluated 13 normal volunteers and seven post-traumatic minimally conscious state patients using 11C-raclopride PET to estimate dopamine 2-like receptors occupancy in the striatum and central thalamus before and after dopamine transporter blockade with dextroamphetamine. If a presynaptic deficit was observed, a third and a fourth 11C-raclopride PET were acquired to evaluate changes in dopamine release induced by l-DOPA and l-DOPA+dextroamphetamine. Permutation analysis showed a significant reduction of dopamine release in patients, demonstrating a presynaptic deficit in the striatum and central thalamus that could not be reversed by blocking the dopamine transporter. However, administration of the dopamine precursor l-DOPA reversed the presynaptic deficit by restoring the biosynthesis of dopamine from both ventral tegmentum and substantia nigra. The advantages of alternative pharmacodynamic approaches in post-traumatic minimally conscious state patients should be tested in clinical trials, as patients currently refractory to amantadine might benefit from them.

Comprehensive Proteomic Profiling of Patients' Tears Identifies Potential Biomarkers for the Traumatic Vegetative State.

The vegetative state is a complex condition with unclear mechanisms and limited diagnostic, prognostic, and therapeutic methods. In this study, we aimed to explore the proteomic profile of tears from patients in a traumatic vegetative state and identify potential diagnostic markers using tears-a body fluid that can be collected non-invasively. Using iTRAQ quantitative proteomic technology, in the discovery phase, tear samples collected from 16 patients in a traumatic vegetative state and 16 normal individuals were analyzed. Among 1080 identified tear proteins, 57 were upregulated and 15 were downregulated in the patients compared to the controls. Bioinformatics analysis revealed that the differentially-expressed proteins were mainly involved in the wound response and immune response signaling pathways. Furthermore, we verified the levels of 7 differentially-expressed proteins in tears from 50 traumatic vegetative state patients and 50 normal controls (including the samples used in the discovery phase) using ELISA. The results showed that this 7-protein panel had a high discrimination ability for traumatic vegetative state (area under the curve = 0.999). In summary, the altered tear proteomic profile identified in this study provides a basis for potential tear protein markers for diagnosis and prognosis of the traumatic vegetative state and also provides novel insights into the mechanisms of traumatic vegetative state.

Comparing the measured basal metabolic rates in patients with chronic disorders of consciousness to the estimated basal metabolic rate calculated from common predictive equations.

BACKGROUND: Accurately predicting the basal metabolic rate (BMR) of patients in a vegetative state (VS) or minimally conscious state (MCS) is critical to proper nutritional therapy, but commonly used equations have not been shown to be accurate. Therefore, we compared the BMR measured by indirect calorimetry (IC) to BMR values estimated using common predictive equations in VS and MCS patients. METHODS: Body composition variables were measured using the bioelectric impedance analysis (BIA) technique. BMR was measured by IC in 82 patients (64 men and 18 women) with VS or MCS. Patients were classified by body mass index as underweight (<18.5 kg/m2, n = 34) or normal-weight (18.5 ≤ BMI < 25 kg/m2, n = 48). BMR was estimated for each group using the Harris-Benedict (H-B), Schofield, or Cunningham equations and compared to the measured BMR using Bland-Altman analyses. RESULTS: For the underweight group, there was a significant difference between the measured BMR values and the estimated BMR values calculated using the H-B, Schofield, and Cunningham equations (p < 0.05). For the normal-weight group, the BMR values estimated using the H-B and Cunningham equations were different significantly from the measured BMR (p < 0.05 and p < 0.01 respectively). Of the predictive equations, only Schofield was not significantly different from the measured BMR in the normal-weight group. The Schofield equation showed the best concordance (only 41.5%) with the BMR values measured by IC. CONCLUSIONS: None of the commonly used equations to estimate BMR were suitable for the VS or MCS populations. Indirect calorimetry is the preferred way to avoid either over or underestimate of BMR values.

Transependymal Movement of Cerebrospinal Fluid in Neurological and Psychiatric Pathological Conditions.

We retrospectively studied the anamnesis, in particular the etiology, the clinical picture, and computed tomography/magnetic resonance imaging/ultrasound data, in the dynamics of a heterogeneous group of 127 patients with neurological and psychiatric pathological conditions. We were interested in the reasons for the occurrence, the clinical value of various neuroimaging abnormalities in the white matter of the brain, including the periventricular zone, the communication of their occurrence with the possible exit of CSF outside of the limits of the ventricular system. In some of the patients investigations into the cerebral blood flow in dynamics using transcranial Doppler was studied. Also in this regard, indications for and the application of minimally invasive neurosurgery techniques for brain revascularization were investigated.

Clinical Response to tDCS Depends on Residual Brain Metabolism and Grey Matter Integrity in Patients With Minimally Conscious State.

BACKGROUND: Transcranial direct current stimulation (tDCS) was recently shown to promote recovery of voluntary signs of consciousness in some patients in minimally conscious state (MCS). However, it remains unclear why clinical improvement is only observed in a subgroup of patients. OBJECTIVES: In this retrospective study, we investigated the relationship between tDCS responsiveness and neuroimaging data from MCS patients. METHODS: Structural Magnetic Resonance Imaging (MRI), Fluorodeoxyglucose Positron emission tomography (FDG-PET) and clinical electroencephalography (EEG) were acquired in 21 sub-acute and chronic MCS patients (8 tDCS responders) who subsequently (<48 h) received left dorsolateral prefrontal (DLPF) tDCS in a double-blind randomized cross-over trial. The behavioral data have been published elsewhere (Thibaut et al., Neurology, 2014). RESULTS: Grey matter atrophy was observed in non-responders as compared with responders in the left DLPF cortex, the medial-prefrontal cortex, the cingulate cortex, the hippocampi, part of the rolandic regions, and the left thalamus. FDG-PET showed hypometabolism in non-responders as compared with responders in the left DLPF cortex, the medial-prefrontal cortex, the precuneus, and the thalamus. EEG did not show any difference between the two groups. CONCLUSION: Our findings suggest that the transient increase of signs of consciousness following left DLPF tDCS in patients in MCS require grey matter preservation and residual metabolic activity in cortical and subcortical brain areas known to be involved in attention and working memory. These results further underline the critical role of long-range cortico-thalamic connections in consciousness recovery, providing important information for guidelines on the use of tDCS in disorders of consciousness.

Quantitative rates of brain glucose metabolism distinguish minimally conscious from vegetative state patients.

The differentiation of the vegetative or unresponsive wakefulness syndrome (VS/UWS) from the minimally conscious state (MCS) is an important clinical issue. The cerebral metabolic rate of glucose (CMRglc) declines when consciousness is lost, and may reveal the residual cognitive function of these patients. However, no quantitative comparisons of cerebral glucose metabolism in VS/UWS and MCS have yet been reported. We calculated the regional and whole-brain CMRglc of 41 patients in the states of VS/UWS (n=14), MCS (n=21) or emergence from MCS (EMCS, n=6), and healthy volunteers (n=29). Global cortical CMRglc in VS/UWS and MCS averaged 42% and 55% of normal, respectively. Differences between VS/UWS and MCS were most pronounced in the frontoparietal cortex, at 42% and 60% of normal. In brainstem and thalamus, metabolism declined equally in the two conditions. In EMCS, metabolic rates were indistinguishable from those of MCS. Ordinal logistic regression predicted that patients are likely to emerge into MCS at CMRglc above 45% of normal. Receiver-operating characteristics showed that patients in MCS and VS/UWS can be differentiated with 82% accuracy, based on cortical metabolism. Together these results reveal a significant correlation between whole-brain energy metabolism and level of consciousness, suggesting that quantitative values of CMRglc reveal consciousness in severely brain-injured patients.

Brainstem neurons survive the identical ischemic stress that kills higher neurons: insight to the persistent vegetative state.

Global ischemia caused by heart attack, pulmonary failure, near-drowning or traumatic brain injury often damages the higher brain but not the brainstem, leading to a 'persistent vegetative state' where the patient is awake but not aware. Approximately 30,000 U.S. patients are held captive in this condition but not a single research study has addressed how the lower brain is preferentially protected in these people. In the higher brain, ischemia elicits a profound anoxic depolarization (AD) causing neuronal dysfunction and vasoconstriction within minutes. Might brainstem nuclei generate less damaging AD and so be more resilient? Here we compared resistance to acute injury induced from simulated ischemia by 'higher' hippocampal and striatal neurons versus brainstem neurons in live slices from rat and mouse. Light transmittance (LT) imaging in response to 10 minutes of oxygen/glucose deprivation (OGD) revealed immediate and acutely damaging AD propagating through gray matter of neocortex, hippocampus, striatum, thalamus and cerebellar cortex. In adjacent brainstem nuclei, OGD-evoked AD caused little tissue injury. Whole-cell patch recordings from hippocampal and striatal neurons under OGD revealed sudden membrane potential loss that did not recover. In contrast brainstem neurons from locus ceruleus and mesencephalic nucleus as well as from sensory and motor nuclei only slowly depolarized and then repolarized post-OGD. Two-photon microscopy confirmed non-recoverable swelling and dendritic beading of hippocampal neurons during OGD, while mesencephalic neurons in midbrain appeared uninjured. All of the above responses were mimicked by bath exposure to 100 µM ouabain which inhibits the Na+/K+ pump or to 1-10 nM palytoxin which converts the pump into an open cationic channel. Therefore during ischemia the Na+/K+ pump of higher neurons fails quickly and extensively compared to naturally resilient hypothalamic and brainstem neurons. The selective survival of lower brain regions that maintain vital functions will support the persistent vegetative state.

PET in coma and in vegetative state.

Advances in resuscitation and critical care management have resulted in the survival of many patients despite severe brain damage. These patients may remain in coma or in vegetative state. The probability of recovery of conscious function is dependent on the extent of structural brain damage, which is difficult to assess by clinical, laboratory or functional tests. Positron emission tomography (PET) of 18F-fluordeoxyglucose (FDG) can be used to investigate metabolic and functional impairment of the brain. In acute vegetative state (AVS, duration <1 month), overall glucose utilization was significantly reduced in comparison with age-matched controls. In a few cases with locked-in syndrome, cortical metabolism was in the normal range. 11C-Flumazenil (FMZ) measures the density of benzodiazepine receptors (BZRs) and thereby furnishes an estimate of neuronal integrity. PET with this tracer demonstrated a considerable reduction in BZRs in cortical areas, but indicated that the cerebellum was spared from neuronal loss. The comparison of FDG- and FMZ-PET findings in AVS demonstrates that alterations of cerebral glucose consumption do not represent mere functional inactivation, but also irreversible structural damage. In some cases with minimally conscious state, auditory stimuli with emotional valence induced more brain activation (investigated by H215O-PET) than meaningless noise; such studies can be used to detect residual cortical function. To improve prognostication of chances for recovery, a combination of functional activation studies and assessment of the extent of neuronal damage might be the optimal procedure and should be tested in larger cohorts of patients with comatose states of different severity.

Arterial spin labeling and altered cerebral blood flow patterns in the minimally conscious state.

OBJECTIVE: To use arterial spin labeling (ASL) to compare cerebral blood flow (CBF) patterns in minimally conscious state (MCS) patients with those in normal controls in an observational study design. METHODS: Subjects meeting MCS criteria and normal controls were identified. A pseudocontinuous ASL sequence was performed with subjects and controls in the resting awake state. Multiple CBF values for 10 predetermined regions of interest were sampled and average CBF was calculated and compared between controls and subjects. RESULTS: Ten normal controls were identified, with ages ranging from 26 to 54 years. Four subjects met the MCS criteria and received an ASL study, with one patient receiving a second study at a later date. Subjects ranged in age from 19 to 58 years and had traumatic brain injury, stroke, or hypoxic-ischemic encephalopathy. Regional CBF for controls ranged from 21.6 to 57.2 mL/100 g/min, with a pattern of relatively increased blood flow posteriorly including the posterior cingulate, parietal, and occipital cortices. CBF patterns for MCS subjects showed greater variability (from 7.7 to 33.1 mL/100 g/min), demonstrating globally decreased CBF in gray matter compared with that in normal controls, especially in the medial prefrontal and midfrontal regions. In the one subject studied longitudinally, global CBF values increased over time, which correlated with clinical improvement. CONCLUSIONS: We identified globally decreased CBF and a selective reduction of CBF within the medial prefrontal and midfrontal cortical regions as well as gray matter in MCS patients. ASL may serve as an adjunctive method to assess functional reserve in patients recovering from severe brain injuries.

A voxel-based analysis of FDG-PET in traumatic brain injury: regional metabolism and relationship between the thalamus and cortical areas.

The objective was to study the correlations and the differences in glucose metabolism between the thalamus and cortical structures in a sample of severe traumatic brain injury (TBI) patients with different neurological outcomes. We studied 49 patients who had suffered a severe TBI and 10 healthy control subjects using 18F-fluorodeoxyglucose positron emission tomography (18F-FDG-PET). The patients were divided into three groups: a vegetative or minimally-conscious state (MCS&VS) group (n=17), which included patients who were in a vegetative or a minimally conscious state; an In-post-traumatic amnesia (In-PTA) group (n=12), which included patients in PTA; and an Out-PTA group (n=20), which included patients who had recovered from PTA. SPM5 software was used to determine the metabolic differences between the groups. FDG-PET images were normalized and four regions of interest were generated around the thalamus, precuneus, and the frontal and temporal lobes. The groups were parameterized using Student's t-test. Principal component analysis was used to obtain an intensity-estimated-value per subject to correlate the function between the structures. Differences in glucose metabolism in all structures were related to the neurological outcome, and the most severe patients showed the most severe hypometabolism. We also found a significant correlation between the cortico-thalamo-cortical metabolism in all groups. Voxel-based analysis suggests a functional correlation between these four areas, and decreased metabolism was associated with less favorable outcomes. Higher levels of activation of the cortico-cortical connections appear to be related to better neurological condition. Differences in the thalamo-cortical correlations between patients and controls may be related to traumatic dysfunction due to focal or diffuse lesions.

[Treatment outcome of chemotherapy with superselective cerebral artery catheterization in vegetative state patients].

The purpose of the study is to assess the impact of superselective neurotransmitter metabolic therapy in patients in a vegetative state. Superselective intraarterial infusion was conducted on 26 patients with relevant international criteria for the diagnosis of vegetative state. Comprehensive assessment of neurologic symptoms and severity of low metabolism on PET scan allowed to select the vascular pool, for the catheter installation. The catheter was placed either in the carotid or the vertebrobasilar pool. Infusion of neurotransmitter agents was conducted for 7 days continuously. Control of the level of metabolism of labeled glucose in the brain (PET) was performed within 2 weeks after arterial infusion. 14 out of 26 patients showed a positive trend of changes in energy metabolism of the brain. However, only 7 out of 14 patients showed further recovery of consciousness. The data confirms that the delivery path and a combination of medications play a definite role in the effectiveness of vegetative state therapy.

"Relevance vector machine" consciousness classifier applied to cerebral metabolism of vegetative and locked-in patients.

The vegetative state is a devastating condition where patients awaken from their coma (i.e., open their eyes) but fail to show any behavioural sign of conscious awareness. Locked-in syndrome patients also awaken from their coma and are unable to show any motor response to command (except for small eye movements or blinks) but recover full conscious awareness of self and environment. Bedside evaluation of residual cognitive function in coma survivors often is difficult because motor responses may be very limited or inconsistent. We here aimed to disentangle vegetative from "locked-in" patients by an automatic procedure based on machine learning using fluorodeoxyglucose PET data obtained in 37 healthy controls and in 13 patients in a vegetative state. Next, the trained machine was tested on brain scans obtained in 8 patients with locked-in syndrome. We used a sparse probabilistic Bayesian learning framework called "relevance vector machine" (RVM) to classify the scans. The trained RVM classifier, applied on an input scan, returns a probability value (p-value) of being in one class or the other, here being "conscious" or not. Training on the control and vegetative state groups was assessed with a leave-one-out cross-validation procedure, leading to 100% classification accuracy. When applied on the locked-in patients, all scans were classified as "conscious" with a mean p-value of .95 (min .85). In conclusion, even with this relatively limited data set, we could train a classifier distinguishing between normal consciousness (i.e., wakeful conscious awareness) and the vegetative state (i.e., wakeful unawareness). Cross-validation also indicated that the clinical classification and the one predicted by the automatic RVM classifier were in accordance. Moreover, when applied on a third group of "locked-in" consciously aware patients, they all had a strong probability of being similar to the normal controls, as expected. Therefore, RVM classification of cerebral metabolic images obtained in coma survivors could become a useful tool for the automated PET-based diagnosis of altered states of consciousness.