Clinical application of neuromodulation therapy in patients with disorder of consciousness: A pooled analysis of 544 participants

Abstract

BACKGROUND:

The number of patients with disorders of consciousness (DoC) has increased dramatically with the advancement of intensive care and emergency medicine, which brings tremendous economic burdens and even ethical issues to families and society.

OBJECTIVE:

To evaluate the effectiveness of neuromodulation therapy for patients with DoC.

METHODS:

First, we conducted a literature review of individual patient data (IPD) on PubMed, EMBASE, and Cochrane-controlled trials following PRISMA guidelines. Then, we collected neuromodulation cases from our institution. Finally, we conducted a pooled analysis using the participants from the medical literature (n = 522) and our local institutions (n = 22).

RESULTS:

In this pooled analysis of 544 patients with DoC with a mean age of 46.33 years, our results revealed that patients have improved CRS-R scores [1.0 points (95% CI, 0.57–1.42)] after neuromodulation. Among them, patients have better effectiveness in traumatic than non-traumatic etiology (P < 0.05). The effectiveness of consciousness improvement could be affected by the age, baseline consciousness state, and duration of stimulation. Compared with non-invasive intervention, an invasive intervention can bring more behavioral improvement (P < 0.0001) to MCS rather than UWS/VS patients. Importantly, neuromodulation is a valuable therapy even years after the onset of DoC.

CONCLUSION:

This pooled analysis spotlights that the application of neuromodulation can improve the behavioral performance of patients with DoC. A preliminary trend is that age, etiology, baseline consciousness state, and stimulation duration could impact its effectiveness.

Keywords

Neuromodulation brain injury disorder of consciousness pooled analysis

1 Introduction

Disorders of consciousness (DoC) are characterized by a deficiency in arousal and/or awareness, including coma, unresponsive wakefulness syndrome (formerly known as ‘vegetative state’) (UWS/VS), minimally conscious state (MCS), and emergence from MCS (EMCS) (Giacino et al., 2014). The number of patients with DoC has increased dramatically with the advancement of intensive care and emergency medicine, which brings tremendous economic burdens and even ethical issues to families and society. Considering the limitations of traditional therapy, including pharmacology and rehabilitative strategies, growing interest has been paid to neuromodulation therapy for consciousness promotion (Zheng et al., 2022).

The American Academy of Neurology (AAN) (Giacino et al., 2018) and the European Academy of Neurology (EAN) (Kondziella et al., 2020) guidelines for managing patients with DoC were issued in 2018 and 2020, respectively. However, how to apply neuromodulatory therapies appropriately and effectively in patients with DoC is still lacking in these guidelines, although several randomized controlled trials (RCTs) are ongoing to provide more evidence-based suggestions, of which transcranial direct current stimulation (tDCS) has shown class II evidence (Thibaut et al., 2014). A recent review suggested that patients with prolonged DoC could benefit from neuromodulation interventions even years after injury. It also clarified the advantages of non-invasive interventions (no side effects) and proposed the disadvantages of invasive interventions (risk of infection and clinical deterioration) (Thibaut et al., 2019). Still, the clinical profile of patients who could benefit from specific interventions and the optimal target or duration of stimulation need to be determined.

In the past, sporadic recommendations on implementing non-invasive neural regulation have come up: tDCS has beneficial effects on DoC, especially in regulating the left dorsolateral prefrontal cortex (DLPFC) in MCS (Zhang & Song, 2018); MCS could benefit more than UWS/VS from tDCS or TMS (Shou et al., 2021). However, most of them are qualitative rather than quantitative suggestions. An evidence-based recommendations are still lacking, and the precise treatment choice and beneficiaries of specific intervention types are still uncertain. The good news is that systematic reviews and meta-analyses of neuromodulation RCT studies provided the highest level of neuromodulation practice (Delahaye et al., 1991). Even so, the neural correlation of consciousness has yet to be established, and only a few clinical trials on neuromodulation are ongoing, leading to uncertainty in selecting its optimal solution. In recent years, clinicians realized that an individual patient data (IPD) analysis that extracts patient-level data from publications allows us to overcome the problem of insufficient data volume (Stewart et al., 2015), and a pooled analysis could bring helpful information and thus to provide experience from neuromodulation cases with heterogeneity (Taioli & Bonassi, 2002).

In this study, we conducted a pooled analysis of existing IPD evidence and our single-center neuromodulation experience in patients with DoC, thus providing a reference for clinical practice and future evidence-based medicine studies on this promising therapy.

2 Methods

2.1 Literature review of IPD

First, a PRISMA method was used for literature retrieval in PubMed, Embase, and the Cochrane Central Register of Controlled Trials (from Jan 2007 to Feb 2022). We combined two components into each search scenario: 1) identified studies with the keywords “disorder of consciousness”, “coma”, “vegetative state”, “unresponsive wakefulness syndrome” or “minimally conscious state”, and 2) identified papers with the keywords about the neuromodulation of interest (that is, “transcranial magnetic stimulation”, “transcranial direct current stimulation”, “vagal nerve stimulation”, “deep brain stimulation” or “spinal cord stimulation”).

Studies that met the broad retrieval scope were reviewed in detail. The inclusion criteria were: 1) English writing human studies, 2) following a PIS strategy, that is, a combination of participants, interventions, and study type, 3) designs included case reports, cohort studies, cross-sectional studies, case-control studies, or randomized controlled trials, 4) non-compound neuromodulation (such as a simultaneous tDCS and rTMS), 5) no redundant objectives (such as diagnostic application, indicator evaluation, or responder differentiation), and 6) using CRS-R for consciousness assessment. The exclusion criteria were: 1) narrative or systematic reviews or meta-analysis studies, 2) not human studies, and 3) lack of CRS-R scores. Further steps were taken to ensure search comprehensiveness: 1) references from included studies were reviewed, 2) studies with weak or null findings were captured to avoid bias toward positive bias inherent in the selected, and 3) the publication bias was investigated through funnel plots (eFigs. 1–3 in the Supplement).

The following study-specific IPD were extracted: the patients’ clinical characteristics, including the age, gender, etiology of injury, and DoC duration, and the stimulation protocol, including neuromodulation types, stimulation duration, stimulation targets, any reported adverse effects (eTable 2 in the Supplement). A pair of reviewers [including at least one member of Disorders of Consciousness: Special Interest Group (DOC-SIG), International Brain Injury Association (IBIA) (Xh.W. and Hb.D.)] discussed all studies at each stage of the flowchart. All participants were discussed until a consensus decision had been reached.

2.2 Participants in our institution

After obtaining approval from the ethics committee and informed consent from the patient’s family, the patients diagnosed with MCS or UWS/VS were recruited for neuromodulation in our institution (from Jan 2007 to Feb 2022). We performed TMS and DBS on our DoC patients (Clinicaltrials.gov: NCT02667899), and their detailed neuromodulation protocols are available in the Supplementary Information. Eligible criteria are 1) no use of cortical excitatory, sedative, or neuromuscular blocker drugs, 2) no history of systemic or severe neurological disease, 3) no critical condition (hemodynamic and respiratory instability), 4) no CT or/and MRI scans showed any focal lesions in the stimulation target, and 5) no surgical contraindication for invasive intervention, and exclusion criteria are 1) incomplete patient information, and 2) lack of CRS-R scores.

2.3 Consciousness assessment

The CRS-R scale can effectively measure patients with DoC and has excellent test-retest reliability in distinguishing between MCS and UWS/VS (Giacino, Kalmar, & Whyte, 2004; Giacino et al., 1991). Previous multicenter observational studies have shown that the linear measurement effect derived from the CRS-R score aligns with scientific measurement principles and is sufficiently reliable for consciousness assessment (La Porta et al., 2013). Meanwhile, the sub-scores can overcome the limitation of low sensitivity and lack of diagnostic accuracy of the CRS-R total scores (Annen et al., 2019). Accordingly, CRS-R scores can reflect the quantitative effectiveness of neuromodulation therapy, and thus, the magnitude of it was comparable across participants. We retained the CRS-R score as the primary outcome indicator if multiple behavior indicators were reported in one study. We collected all CRS-R data in the recent records following intervention (post-neuromodulation CRS-R scores) to capture the relatively complete behavior scores in all studies (eTable 3 in the Supplement). These data are used for further subgroup analysis to evaluate the effectiveness of neuromodulation.

2.4 Statistical analysis

Considering the heterogeneity in study design, studies recorded from different sources are sometimes not comparable. We applied the four key steps to make our analysis more persuasive (Stewart et al., 2015). First, we performed a homogeneity test over participants after standard feature decomposition and accordingly extracted individual data for comparison. Second, we retained all literature-recorded CRS-R information to a maximum extent, even if there was no improvement. Third, we adopted the latest CRS-R scores after the treatment as our primary outcome measures. Fourth, we also incorporated the information on neuromodulation cases from our institution.

Of the 544 participants (522 from the literature and 22 from our institution), categorical variables are reported as percentages and were assessed with the χ² test or Fisher’s exact test as appropriate; continuous variables were expressed as means±standard deviation (SD) or medians and 25% interquartile range (IQR). Normally and non-normally distributed continuous variables were compared with Student’s t-test, Mann-Whitney U test, one-way ANOVA analysis, or Tukey’s multiple comparisons test, respectively. Statistical tests were baseline adjusted as needed from one-tailed or two-tailed settings, and p values <0.05 were considered statistically significant. Prism statistical software (version 9.3.1, https://www.graphpad.com) was used to perform all statistical analyses, and RevMan software (version 5.3, https://tech.cochrane.org/revman) was used to perform meta-analysis.

3 Results

In this study, 544 patients with DoC were eligible for further pooled analysis: 522 patients from the database and 22 patients from our institution (Fig. 1). The general demographic and clinical characteristics of the included participants are summarized in Table 1. The distribution and weight of included studies are displayed using forest plots (eFig. 4 in the Supplement). On the whole, the overall effect of all neuromodulation techniques was z = 4.61 (P < 0.00001, 95% CI), and the improvement in CRS-R scores was 1.0 point (95% CI, 0.57–1.42) from baseline to first evaluation post-neuromodulation. Significantly, there were significant differences in the baseline and post-neuromodulation CRS-R scores (P < 0.001) (Fig. 2A).

Fig. 1

Flowchart depicting the identification and inclusion of participants.

Fig. 2

Characteristic differences of behavioral effectiveness. A. Comparison of CRS-R in pre- and post-neuromodulation. B. Comparison of age in patients with improved CRS-R and patients without improved CRS-R. C. Comparison of the duration of stimulation in patients with improved CRS-R and patients without improved CRS-R. D. Comparison of DoC duration in patients with improved CRS-R and patients without improved CRS-R. E. Comparison of CRS-R D value in patients with different etiology. F. Comparison of behavior improvement in patients with different baseline consciousness states. G. Comparison of CRS-R D value in patients with different baseline consciousness states (in patients with CRS-R D > 0). H. Correlation analysis between the duration of stimulation (in days) and the CRS-R D value. I. Correlation matrix between baseline CRS-R total score, baseline CRS-R sub-score, and CRS-R D value. (Bars represent mean±standard deviation, ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001, CRS-R(+) means: CRS-R scores improvement).

Table 1

Patient’s demographic and clinical characteristics

Variables	Value
Age (years) (mean±SD)		46.33±16.42
Baseline diagnosis	EMCS	42.80±17.01
	MCS	46.28±16.47
	UWS/VS	46.97±16.60
Intervention	tDCS	47.59±16.42
	TMS	44.21±16.22
	VNS	43.26±19.55
	DBS	44.24±14.41
	SCS	39.48±13.84
Gender (N, %)
Male		346 (63.6%)
Female		198 (36.4%)
DoC duration (months) [median, IQR]
Baseline diagnosis	EMCS	10 [5.6–36.8]
	MCS	13 [4–36]
	UWS/VS	11 [4–21]
Intervention	tDCS	10 [3–825.6]
	TMS	16 [7–21]
	VNS	8 [3.2–12]
	DBS	20 [7–53]
	SCS	8 [4–11]
Etiology of DoC (N, %)
Traumatic		219 (40.3%)
Non-traumatic		325 (59.7%)
CRS-R (points) [median, IQR]
Pre-neuromodulation		8 [6–11]
Post-neuromodulation		9 [6–13]

Hence, we define a CRS-R difference (CRS-R D) value generated by baseline CRS-R scores subtracted from the post-neuromodulation CRS-R scores to achieve a quantitative comparison of effectiveness. Accordingly, the patients who presented with relatively younger ages (44.75 years vs. 47.71 years) and received a longer duration of stimulation (P < 0.001) tended to show higher CRS-R D values (Fig. 2B, C). There were no significant differences in the duration of DoC, indicating that neuromodulation is still an effective intervention for months to years (0.2 to 631 months) after the onset of DoC (Fig. 2D). The patients who presented with traumatic etiology had higher CRS-R D values than those with non-traumatic etiology (P < 0.001), of which patients with hemorrhagic etiology had higher CRS-R D values than those with anoxic etiology (P < 0.001) (Fig. 2E). The improvement in consciousness scores of patients in the MCS group was significantly higher than that in the UWS/VS group (P < 0.001) (Fig. 2F, G). Although there was no trend of higher baseline CRS-R scores tending to have higher improvement in DoC, there was a time-dependent effectiveness of neuromodulation; that is, the longer the duration of stimulation, the higher the D value (Fig. 2H, I) (eFig. 5 in the Supplement).

We also found that patients who underwent the invasive intervention (e.g., DBS and SCS) showed significantly higher CRS-R D values than those who received the non-invasive intervention (e.g., tDCS, TMS, and VNS) (P < 0.0001), especially in patients with traumatic etiology (P < 0.0001). A significant improvement in behavioral scores usually tends to present in patients with hemorrhagic etiology (P < 0.01) following a non-invasive intervention, while in patients with traumatic etiology (P < 0.01) following an invasive intervention. We also found that compared with non-invasive intervention, the patients with UWS/VS have higher behavioral score improvement following invasive intervention (P < 0.01) (Fig. 3A–C). Second, despite there being more studies on tDCS, our multiple comparison results showed that the DBS (P < 0.0001) and TMS (P < 0.001) approaches have better performance than others. When focusing on tDCS (clinical III evidence) (Thibaut et al., 2019), our results showed that CRS-R D values were significantly higher in patients with MCS than in patients with UWS/VS (P < 0.01) (Fig. 3D–F). Third, except that the CRS-R D values of patients who receive TMS at the temporal parietal junction (TPJ) area are higher than that of others (P < 0.05), there were no significant differences in the different stimulation target selection for given neuromodulation type (Fig. 3G–I).

Fig. 3

Difference in methodological selection. A. Comparison of CRS-R D value in non-invasive and invasive neuromodulation. B. Comparison of CRS-R D value in patients with different etiology following non-invasive or invasive neuromodulation. C. Comparison of CRS-R D value in patients with different baseline consciousness following non-invasive or invasive neuromodulation. D. Comparison of CRS-R D value in different types of neuromodulation. E. Comparison of CRS-R D value in patients with different baseline consciousness following tDCS. F. Comparison of CRS-R D value in patients with different baseline consciousness following TMS. G. Comparison of CRS-R D value in select different tDCS targets. H. Comparison of CRS-R D value in select different TMS targets. I. Comparison of CRS-R D value in select different surgical targets. (Bars represent mean±standard deviation, ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001).

Furthermore, we analyzed the neuromodulation approach over two severe types of DoC, including 517 patients diagnosed with MCS and UWS/VS. Our results showed that the differences in CRS-R D value between invasive and non-invasive intervention were insignificant, and it was higher in TMS than in tDCS groups in MCS patients (P < 0.001). The CRS-R D value was higher in invasive than non-invasive intervention groups in UWS/VS patients (P < 0.05), of which invasive intervention seems to have better performance than tDCS (P < 0.01) (Fig. 4A, B). Meanwhile, our correlation analysis results showed that: 1) there was a correlation between the duration of tDCS and the CRS-R D value in patients with MCS (P < 0.001) rather than in patients with UWS/VS (Fig. 4C, D); and 2) there was a correlation between the duration of TMS and the CRS-R D value in patients with UWS/VS (P < 0.0001) rather than in patients with MCS (Fig. 4E, F).

Fig. 4

Difference between MCS and UWS/VS. A. Comparison of CRS-R D value in patients with MCS following different types of neuromodulation. B. Comparison of CRS-R D value in patients with UWS/VS following different types of neuromodulation. C.E. Correlation analysis between CRS-R D value and assumed methodological factors in patients with MCS following tDCS, TMS, and invasive intervention. D.F. Correlation analysis between CRS-R D value and assumed methodological factors in patients with UWS/VS following tDCS, TMS, and invasive intervention. (Bars represent mean±standard deviation, ^*p < 0.05, ^**p < 0.01, ^***p < 0.001).

4 Discussion

The management of patients with DoC needs to be better understood. Previously proposed therapies (e.g., rehabilitation training, hyperbaric oxygen, stem cell therapies) were far from meeting the standard for clinical application extensively; medications (e.g., amantadine, class II; level B) may also have associated side effects (Edlow et al., 2021; Giacino et al., 2018). Neuromodulation therapy is promising in having a few side effects and has been increasingly explored for consciousness promotion in patients with DoC (Thibaut et al., 2019). However, almost all clinical neuromodulation studies were limited by small sample sizes and had insufficient evidence to support or refute their clinical use. Therefore, we performed a pooled analysis of 544 participants who received neuromodulation for consciousness-promoting therapy in this study.

Despite the heterogeneity in their protocols, our pooled analysis results favored that current neuromodulation therapy can improve behavioral performance in patients with DoC. We conducted stratified analyses to assess whether the DoC characteristics factors would influence this behavioral improvement. We found that patients with MCS (vs. UWS/VS) status, traumatic (vs. non-traumatic) etiology, and relatively younger age have better consciousness improvement following neuromodulation. The regularity of these was in line with the rehabilitation performance of different DoC etiologies: 50% of MCS patients and 3% of UWS/VS patients showed behavioral improvement; 38% of patients with traumatic MCS and only 2% of patients with non-traumatic UWS/VS have no response within one year (Hirschberg & Giacino, 2011). In addition, traumatic etiology tends to have higher recovery potential than those with anoxic (Cullen, Crescini, & Bayley, 2009) and hypoxic-ischemic etiology (Harbinson, Zarshenas, & Cullen, 2017). The differences may result from the different pathological mechanisms in DoC etiology: the cerebral hemisphere axonal injury was predominant in anoxic injury, while the central myelin injury was chiefly present in TBI (van der Eerden et al., 2014). Another cause is that the post-anoxic DoC patients tended to experience a severe burden of clinical complexity; most of them had worse premorbid clinical comorbidity (e.g., older age and high-risk factors for cerebral diseases) (Estraneo et al., 2021). Our results also add insights into neuromodulation decision-making: the younger age, shorter time post-injury, and higher baseline CRS-R scores were associated with better consciousness recovery (Estraneo et al., 2020). Interestingly, a higher baseline CRS-R score does not mean a higher likelihood of clinical improvement, and the effectiveness of neuromodulation may not be affected by the course of DoC. In addition, DoC patients may present with delayed consciousness recovery, which could take from a year and last for several years or even decades (Estraneo et al., 2010).

An exciting result of this study was that invasive surgical intervention is more effective in clinical improvement than non-invasive brain stimulation. However, there was a circulatory stimulation mode (daytime-on and night-off) in the DBS and SCS intervention that corresponded to the regular sleep-wake cycles, and there was apparent intermittent consciousness behavior before surgical intervention (Yamamoto et al., 2013). Therefore, it is vital to interpret such a finding cautiously, and how to carry out invasive intervention for patients with DoC is still in the exploratory stage; developing more consensus regarding the standardization of surgical procedures is desirable. Significantly, recent evidence indicated that the invasive intervention is better for behavior improvement for patients with UWS/VS; it is at least a reasonable measure when non-invasive stimulation is ineffective. Regarding the two most widely used non-invasive interventions (tDCS and TMS), the role of methodological factors in effectiveness discrepancy is relatively modest compared to the above-mentioned implicated risk factors. For example, for patient selection, the application of tDCS is related to higher behavior scores in MCS compared with UWS/VS; for target selection, applying TMS over TPJ was associated with higher behavior scores than other targets.

Further, the ascending reticular activating system (ARAS) of the brainstem and its connections with the thalamus and cortex are involved in the pathophysiology of coma. The subcortical areas (e.g., basal forebrain, striatum, globus pallidus, and thalamocortical or corticothalamic projections) (Weng et al., 2017; Wu et al., 2018) and the cortical regions located in large-scale brain networks [e.g., default mode network (DMN), salience network (SN)] (Qin, Wu, Huang, et al., 2015; Zheng et al., 2023) are probably involved in the pathophysiology of MCS and UWS/VS. However, there is ongoing scientific debate about whether the neural correlates of consciousness emerge from the anterior or posterior cortex regions (Edlow, 2021). A recent study highlights the importance of brainstem-cortical interplay for consciousness perturbation, that is, the ventral tegmental area (VTA) in the brainstem to the precuneus and posterior cingulate (PCu/PCC) in the cortical DMN (Spindler et al., 2021). These findings indicated that the “VTA-DMN” appears to be a prime target of neuromodulation for consciousness-promoting therapies (Xu et al., 2023). However, whether the top-down (i.e., cortical), bottom-up (i.e., subcortical), or multi-directional (i.e., concurrent cortical and subcortical) is the most effective way is unclear (Edlow, 2021). Meanwhile, whether a Hebbian spike-timing-dependent plasticity theory will affect the consciousness-related interregional brain coupling during the neuromodulation process is unclear (Sel et al., 2021). Our pooled analysis results show these conundrums graphically; we cannot yet provide a “first-choice” neuromodulation target for consciousness promotiong. Despite all this, the good news is that the Connectome-based Clinical Trial Platform (CCTP), a mechanistic clinical trial paradigm, was proposed in 2020, which uses personalized brain connector maps to guide targeted neuromodulation therapy (Edlow et al., 2020).

A positive correlation between the duration of stimulation and behavioral improvement revealed a time-dependent effectiveness of neuromodulation in DoC individuals. Based on our previous findings, GABA_A receptor binding level is associated with consciousness recovery after three months of onset DoC, and this correlation could be determined by the effect of neurotransmitter action (Qin, Wu, Duncan, et al., 2015). Neuromodulation could change brain function by inducing or/and expressing the synaptic plasticity, of which neurotransmitters (e.g., dopamine, serotonin, acetylcholine, and noradrenaline) are released in response to stimulation. Specifically, neurotransmitters can function from short-term adjustments of neuron and/or synapse function to the persistence of long-term regulation of brain state (Nadim & Bucher, 2014). In addition, the effectiveness of neuromodulation could result from the restorative neuroplasticity of DoC patients in different dimensions (Bagnato, 2022). At the macroscopic dimension, it could enhance the expression of brain-derived neurotrophic factors and the density of dendritic spines in the motor cortex (Longo et al., 2022). At the mesoscopic dimension, it could promote novo myelination tunes activated circuits (Steadman et al., 2020). At the microscopic dimension, it could lead to specific neuronal reorganization and trigger the reorganization of functional brain modules (Pallanti et al., 2021).

5 Limitations and suggestions

Several limitations of this study warrant consideration. Firstly, the primary defect of this pooled analysis is that treatment effects, meaning active versus sham, were not considered. Only active treatment could have biased the results (>positive), especially in open-label trials. Secondly, we did not study the different effects of individually detailed neuromodulation protocol. Although a strict inclusion criterion could help achieve this goal, the variation in these studies (e.g., the different parameters, duration, and frequency for intervention, the different evaluation times, or the environmental factor) limits the generalizability. Thirdly, single-arm and case report studies are inevitably included in the analysis to increase sample size, as a rigorous RCT trial has been published on unified neuromodulation protocol for DoC. In addition, we failed to investigate the long-term effects of neuromodulation and repetition of behavioral assessments to avoid misdiagnosis (Wannez et al., 2017). These limitations may need further analysis to stratify different strategies, participants, and environments based on risk.

Even so, a significant strength of this pooled analysis was the inclusion of DoC patients who undergo neuromodulation intervention to improve consciousness and have a complete record of behavioral scores from a large sample. By focusing on the implication of neuroregulation in DoC patients rather than detailed protocols, we preliminarily evaluated the factors that affect behavioral effectiveness. The included participants come from countries worldwide, making this a generalizable suggestion to clinicians. The literature search was rigorous and comprehensive, allowing for more convincing results.

6 Conclusion

This study underscores the effectiveness of neuromodulation as a consciousness-promoting therapy in patients with DoC. The current neuromodulation trend was that patients with relatively younger age, traumatic (vs. non-traumatic) etiology, in MCS (vs. UWS/VS), and received long-term stimulation tend to have better behavioral improvement following neuromodulation. Non-invasive neuromodulation could be carried out as much as possible if the environment and conditions permit, while a treatment protocol for invasive neuromodulation needs to be established. Considering the heterogeneity and limitations of existing studies, prospective studies with a uniform protocol are needed to confirm our findings and determine the preliminary regularity in different settings.

Footnotes

Acknowledgments

This work was supported by the following funding sources: The National Key Research and Development Program of China (No. 2022YFE0141300); the National Natural Science Foundation of China (No.82271224); the Shanghai Municipal Science and Technology Major Project (No.2018SHZDZX01); the ZhangJiang Lab, Shanghai Center for Brain Science and Brain-Inspired Technology; the Lingang Laboratory Shanghai, China (No. LG202105-02-03); the National major pre-research project, China (No. IDF151042) (Pilot project); the University and University Hospital of Liège, Belgian National Funds for Scientific Research (FRS-FNRS); the European Union’s Horizon 2020 Framework Programme for Research and Innovation under the Specific Grant Agreement (No. 945539) (Human Brain Project SGA3); the National Natural Science Foundation of China (Joint Research Project 81471100); the European Space Agency (ESA) and the Belgian Federal Science Policy Office (BELSPO) in the framework of the PRODEX Programme, “Fondazione Europea di Ricerca Biomedica”; the Mind Science Foundation; the European Foundation of Biomedical Research FERB Onlus, BIAL Foundation; the Fund Generet of King Baudouin Foundation; the Mind Care International Foundation; and the AstraZeneca Foundation. A.T. is a research associate, and S.L. is research director at the F.R.S.-FNRS.

Author contributions

Xh.W., Y.M., and S.L. supervised this study. Xh.W. and Rz.Z. contributed to the conception; Rz.Z. design of the work and drafted this manuscript; Hb.D., Z.W., and Zy.X. provided help for data collection; Zx.Q. and Xh.W. provided with original data from our institution; A.T. and S.L. are responsible for the manuscript revisions. All authors read and agreed to the final version of the manuscript. Xh.W. and Y.M. are co-correspondence authors.

Conflict of interest

The authors declare that they have no conflicts of interest.

The supplementary material is available in the electronic version of this article: .

References

Annen,

, Filippini,

M. M

, Bonin,

, Cassol,

, Aubinet,

, Carriere,

, Gosseries,

, Thibaut,

, Barra,

, Wolff,

, Sanz,

L. R. D

, Martial,

, Laureys,

, Chatelle,

(2019). Diagnostic accuracy of the CRS-R index in patients with disorders of consciousness. Brain Inj 33(11), 1409–1412. https://doi.org/10.1080/02699052.2019.1644376

Bagnato,

(2022). The role of plasticity in the recovery of consciousness. Handb Clin Neurol 184, 375–395. https://doi.org/10.1016/B978-0-12-819410-2.00020-5

Cullen,

N. K

, Crescini,

, Bayley,

M. T.

(2009). Rehabilitation outcomes after anoxic brain injury: A case-controlled comparison with traumatic brain injury. PM R 1(12), 1069–1076. https://doi.org/10.1016/j.pmrj.2009.09.013

Delahaye,

, Landrivon,

, Ecochard,

, Colin,

(1991). Meta-analysis. Health Policy 19(2-3), 185–196. https://doi.org/10.1016/0168-8510(91)90007-k

Edlow,

B. L.

(2021). Dopaminergic modulation of human consciousness via default mode network connectivity. Proc Natl Acad Sci U S A, 118(31). https://doi.org/10.1073/pnas.2111268118

Edlow,

B. L

, Barra,

M. E

, Zhou,

D. W

, Foulkes,

A. S

, Snider,

S. B

, Threlkeld,

Z. D

, Chakravarty,

, Kirsch,

J. E

, Chan,

S. T

, Meisler,

S. L

, Bleck,

T. P

, Fins,

J. J

, Giacino,

J. T

, Hochberg,

L. R

, Solt,

, Brown,

E. N

, Bodien,

Y. G.

(2020). Personalized connectome mapping to guide targeted therapy and promote recovery of consciousness in the intensive care unit. Neurocrit Care 33(2), 364–375. https://doi.org/10.1007/s12028-020-01062-7

Edlow,

B. L

, Sanz,

L. R. D

, Polizzotto,

, Pouratian,

, Rolston,

J. D

, Snider,

S. B

, Thibaut,

, Stevens,

R. D

, Gosseries,

, Curing Coma,

, its contributing, m. (2021). Therapies to restore consciousness in patients with severe brain injuries: A gap analysis and future directions. Neurocrit Care 35(Suppl 1), 68–85. https://doi.org/10.1007/s12028-021-01227-y

Estraneo,

, Fiorenza,

, Magliacano,

, Formisano,

, Mattia,

, Grippo,

, Romoli,

A. M

, Angelakis,

, Cassol,

, Thibaut,

, Gosseries,

, Lamberti,

, Noe,

, Bagnato,

, Edlow,

B. L

, Chatelle,

, Lejeune,

, Veeramuthu,

, Bartolo,

, DoC-SIG,

(2020). Multicenter prospective study on predictors of short-term outcome in disorders of consciousness. Neurology 95(11), e1488–e1499. https://doi.org/10.1212/WNL.0000000000010254

Estraneo,

, Masotta,

, Bartolo,

, Pistoia,

, Perin,

, Marino,

, Lucca,

, Pingue,

, Casanova,

, Romoli,

, Gentile,

, Formisano,

, Salvi,

G. P

, Scarponi,

, De Tanti,

, Bongioanni,

, Rossato,

, Santangelo,

, Diana,

A. R

, De Bellis,

(2021). Multi-center study on overall clinical complexity of patients with prolonged disorders of consciousness of different etiologies. Brain Inj 35(1), 1–7. https://doi.org/10.1080/02699052.2020.1861652

10.

Estraneo,

, Moretta,

, Loreto,

, Lanzillo,

, Santoro,

, Trojano,

(2010). Late recovery after traumatic, anoxic, or hemorrhagic long-lasting vegetative state. Neurology 75(3), 239–245. https://doi.org/10.1212/WNL.0b013e3181e8e8cc

11.

Giacino,

J. T

, Fins,

J. J

, Laureys,

, Schiff,

N. D.

(2014). Disorders of consciousness after acquired brain injury: The state of the science. Nat Rev Neurol 10(2), 99–114. https://doi.org/10.1038/nrneurol.2013.279

12.

Giacino,

J. T

, Kalmar,

, Whyte,

(2004). The JFK coma recovery scale-revised: Measurement characteristics and diagnostic utility. Arch Phys Med Rehabil 85(12), 2020–2029. https://doi.org/10.1016/j.apmr.2004.02.033

13.

Giacino,

J. T

, Katz,

D. I

, Schiff,

N. D

, Whyte,

, Ashman,

E. J

, Ashwal,

, Barbano,

, Hammond,

F. M

, Laureys,

, Ling,

G. S. F

, Nakase-Richardson,

, Seel,

R. T

, Yablon,

, Getchius,

T. S. D

, Gronseth,

G. S

, Armstrong,

M. J.

(2018). Practice guideline update recommendations summary: Disorders of consciousness: Report of the Guideline Development, Dissemination, and Implementation Subcommittee of the American Academy of Neurology; the American Congress of Rehabilitation Medicine; and the National Institute on Disability, Independent Living, and Rehabilitation Research. Neurology 91(10), 450–460. https://doi.org/10.1212/WNL.0000000000005926

14.

Giacino,

J. T

, Kezmarsky,

M. A

, DeLuca,

, Cicerone,

K. D.

(1991). Monitoring rate of recovery to predict outcome in minimally responsive patients. Arch Phys Med Rehabil 72(11), 897–901. https://doi.org/10.1016/0003-9993(91)90008-7

15.

Harbinson,

, Zarshenas,

, Cullen,

N. K.

(2017). Long-term functional and psychosocial outcomes after hypoxic-ischemic brain injury: A case-controlled comparison to traumatic brain injury. PM R 9(12), 1200–1207. https://doi.org/10.1016/j.pmrj.2017.04.015

16.

Hirschberg,

, Giacino,

J. T.

(2011). The vegetative and minimally conscious states: Diagnosis, prognosis and treatment. Neurol Clin 29(4), 773–786. https://doi.org/10.1016/j.ncl.2011.07.009.

17.

Kondziella,

, Bender,

, Diserens,

, van Erp,

, Estraneo,

, Formisano,

, Laureys,

, Naccache,

, Ozturk,

, Rohaut,

, Sitt,

J. D

, Stender,

, Tiainen,

, Rossetti,

A. O

, Gosseries,

, Chatelle,

, Ean Panel on Coma, D. o. C. (2020). European Academy of Neurology guideline on the diagnosis of coma and other disorders of consciousness. Eur J Neurol 27(5), 741–756. https://doi.org/10.1111/ene.14151

18.

La Porta,

, Caselli,

, Ianes,

A. B

, Cameli,

, Lino,

, Piperno,

, Sighinolfi,

, Lombardi,

, Tennant,

(2013). Can we scientifically and reliably measure the level of consciousness in vegetative and minimally conscious States? Rasch analysis of the coma recovery scale-revised. Arch Phys Med Rehabil 94(3), 527–535e521. https://doi.org/10.1016/j.apmr.2012.09.035

19.

Longo,

, Barbati,

S. A

, Re,

, Paciello,

, Bolla,

, Rinaudo,

, Miraglia,

, Alu,

, Di Donna,

M. G

, Vecchio,

, Rossini,

P. M

, Podda,

M. V

, Grassi,

(2022). Transcranial direct current stimulation enhances neuroplasticity and accelerates motor recovery in a stroke mouse model. Stroke 53(5), 1746–1758. https://doi.org/10.1161/STROKEAHA.121.034200

20.

Nadim,

, Bucher,

(2014). Neuromodulation of neurons and synapses. Curr Opin Neurobiol 29, 48–56. https://doi.org/10.1016/j.conb.2014.05.003

21.

Pallanti,

, Marras,

, Dickson,

S. L

, Adan,

R. A

, Vieta,

, Dell Osso,

, Arango,

, Fusar-Poli,

, Soriano-Mas,

, Carmi,

, Meyer Lindenberg,

, Zohar,

(2021). Manifesto for an ECNP Neuromodulation Thematic Working Group (TWG): Non-invasive brain stimulation as a new Super-subspecialty. Eur Neuropsychopharmacol 52, 72–83. https://doi.org/10.1016/j.euroneuro.2021.07.002

22.

Qin,

, Wu,

, Duncan,

N. W

, Bao,

, Tang,

, Zhang,

, Hu,

, Jin,

, Wu,

, Gao,

, Lu,

, Guan,

, Lane,

, Huang,

, Bodien,

Y. G

, Giacino,

J. T

, Mao,

, Northoff,

(2015). GABAA receptor deficits predict recovery in patients with disorders of consciousness: A preliminary multimodal [(11) C]Flumazenil PET and fMRI study. Hum Brain Mapp 36(10), 3867–3877. https://doi.org/10.1002/hbm.22883

23.

Qin,

, Wu,

, Huang,

, Duncan,

N. W

, Tang,

, Wolff,

, Hu,

, Gao,

, Jin,

, Wu,

, Zhang,

, Lu,

, Wu,

, Qu,

, Mao,

, Weng,

, Zhang,

, Northoff,

(2015). How are different neural networks related to consciousness? Ann Neurol 78(4), 594–605. https://doi.org/10.1002/ana.24479

24.

Sel,

, Verhagen,

, Angerer,

, David,

, Klein-Flugge,

M. C

, Rushworth,

M. F. S.

(2021). Increasing and decreasing interregional brain coupling increases and decreases oscillatory activity in the human brain. Proc Natl Acad Sci U S A, 118(37). https://doi.org/10.1073/pnas.2100652118

25.

Shou,

, Li,

, Wang,

, Chen,

, Bai,

, Di,

(2021). Non-invasive brain intervention techniques used in patients with disorders of consciousness. Int J Neurosci 131(4), 390–404. https://doi.org/10.1080/00207454.2020.1744598

26.

Spindler,

L. R. B

, Luppi,

A. I

, Adapa,

R. M

, Craig,

M. M

, Coppola,

, Peattie,

A. R. D

, Manktelow,

A. E

, Finoia,

, Sahakian,

B. J

, Williams,

G. B

, Allanson,

, Pickard,

J. D

, Menon,

D. K

, Stamatakis,

E. A.

(2021). Dopaminergic brainstem disconnection is common to pharmacological and pathological consciousness perturbation. Proc Natl Acad Sci U S A, 118(30). https://doi.org/10.1073/pnas.2026289118

27.

Steadman,

P. E

, Xia,

, Ahmed,

, Mocle,

A. J

, Penning,

A. R. A

, Geraghty,

A. C

, Steenland,

H. W

, Monje,

, Josselyn,

S. A

, Frankland,

P. W.

(2020). Disruption of oligodendrogenesis impairs memory consolidation in adult mice. Neuron, 105(1), 150–164e156. https://doi.org/10.1016/j.neuron.2019.10.013

28.

Stewart,

L. A

, Clarke,

, Rovers,

, Riley,

R. D

, Simmonds,

, Stewart,

, Tierney,

J. F

, Group,

P. -I. D.

(2015). Preferred reporting items for systematic reviewand meta-analyses of individual participant data: The PRISMA-IPD state-ment. JAMA, 313(16), 1657–1665. https://doi.org/10.1001/jama.2015.3656

29.

Taioli,

, Bonassi,

(2002). Methodological issues in pooled analysis of biomarker studies. Mutat Res 512(1), 85–92. https://doi.org/10.1016/s1383-5742(02)00027-

30.

Thibaut,

, Bruno,

M. A

, Ledoux,

, Demertzi,

, Laureys,

(2014). tDCS in patients with disorders of consciousness: Sham-controlled randomized double-blind study. Neurology 82(13), 1112–1118. https://doi.org/10.1212/WNL.0000000000000260

31.

Thibaut,

, Schiff,

, Giacino,

, Laureys,

, Gosseries,

(2019). Therapeutic interventions in patients with prolonged disorders of consciousness. Lancet Neurol 18(6), 600–614. https://doi.org/10.1016/S1474-4422(19)30031-6

32.

van der Eerden,

A. W

, Khalilzadeh,

, Perlbarg,

, Dinkel,

, Sanchez,

, Vos,

P. E

, Luyt,

C. E

, Stevens,

R. D

, Menjot de Champfleur,

, Delmaire,

, Tollard,

, Gupta,

, Dormont,

, Laureys,

, Benali,

, Vanhaudenhuyse,

, Galanaud,

, Puybasset,

, Consortium,

(2014). White matter changes in comatose survivors of anoxic ischemic encephalopathy and traumatic brain injury: Comparative diffusion-tensor imaging study. Radiology 270(2), 506–516. https://doi.org/10.1148/radiol.13122720

33.

Wannez,

, Heine,

, Thonnard,

, Gosseries,

, Laureys,

, Coma Science Group, C. (2017). The repetition of behavioral assessments in diagnosis of disorders of consciousness. Ann Neurol 81(6), 883–889. https://doi.org/10.1002/ana.24962

34.

Weng,

, Xie,

, Zhao,

, Zhang,

, Ma,

, Wang,

, Jiang,

, He,

, Chen,

, Li,

, Ni,

, Xu,

, Yu,

, Huang,

(2017). Abnormal structural connectivity between the basal ganglia, thalamus, and frontal cortex in patients with disorders of consciousness. Cortex 90, 71–87. https://doi.org/10.1016/j.cortex.2017.02.011

35.

Wu,

, Zhang,

, Cui,

, Tang,

, Shao,

, Hu,

, Zhu,

, Zhao,

, Lu,

, Chen,

, Northoff,

, Gong,

, Mao,

, He,

(2018). White matter deficits underlying the impaired consciousness level in patients with disorders of consciousness. Neurosci Bull 34(4), 668–678. https://doi.org/10.1007/s12264-018-0253-3

36.

Xu,

, Zheng,

, Zhou,

, Feng,

(2023). Alteration in ventral tegmental area and default mode network interplay and prediction of coma recovery in patients with sTBI. Heliyon 9(4), e15279. https://doi.org/10.1016/j.heliyon.2023.e15279.

37.

Yamamoto,

, Katayama,

, Obuchi,

, Kobayashi,

, Oshima,

, Fukaya,

(2013). Deep brain stimulation and spinal cord stimulation for vegetative state and minimally conscious state. World Neurosurg, 80(3-4), S30e31–39. https://doi.org/10.1016/j.wneu.2012.04.010

38.

Zhang,

, Song,

(2018). Transcranial direct current stimulation in disorders of consciousness: A review. Int J Neurosci 128(3), 255–261. https://doi.org/10.1080/00207454.2017.1381094

39.

Zheng,

R. Z

, Qi,

Z. X

, Wang,

, Xu,

Z. Y

, Wu,

X. H

, Mao,

(2022). Clinical decision on disorders of consciousness after acquired brain injury: Stepping forward. Neurosci Bull. https://doi.org/10.1007/s12264-022-00909-7

40.

Zheng,

R. Z

, Xu,

C. X

, Zhou,

L. Y

, Feng,

D. F.

(2023). Default mode network overshadow executive control network in coma emergence and awakening prediction of patients with sTBI. Neuroimage Clin 37, 103361. https://doi.org/10.1016/j.nicl.2023.103361