Severe brain injury can significantly disrupt the structural and functional connectivity (FC) within brain networks, leading to disorders of consciousness (DoC) (Giacino et al., 2014). DoC are marked by altered states of arousal and/or awareness. Patients with severely diminished arousal typically enter a coma—a state of profound brain function inhibition during the acute phase of injury. While coma often resolves as patients regain wakefulness within 2 to 4 weeks (Giacino et al., 2014), some patients (e.g., those with traumatic brain injury or stroke) do not progress to wakefulness due to extensive damage to the forebrain, thalamus, and brainstem, resulting in a prolonged DoC (pDoC) (Kondziella et al., 2020). Patients with pDoC are clinically classified based on the manifestations of conscious awareness into two primary categories: vegetative state/unresponsive wakefulness syndrome (VS/UWS) and minimally conscious state (MCS). Patients in MCS exhibit some level of conscious behavior (Naccache, 2018), whereas those in VS/UWS show no behavioral responses to external stimuli (Monti et al., 2010). MCS can further be divided into MCS minus (MCS-) and MCS plus (MCS+) (Bruno et al., 2011), depending on the presence of language-related behaviors (Thibaut et al., 2020). MCS may act as a transitional state for VS/UWS patients, potentially allowing a gradual return to wakefulness. The study concluded that patients with VS/UWS typically exhibit a lack of response to treatment due to extensive neural pathway disconnections in the forebrain (Edlow et al., 2021). Consequently, the focus of subsequent research was shifted to patients with MCS (Dong et al., 2023; Hu et al., 2024). However, premature palliative care for patients with VS/UWS may lead to a shift in focus from treatment to care during recovery, ultimately leading to a poor prognosis.

With advancements in central intervention techniques, including transcranial magnetic stimulation (TMS) and transcranial direct current stimulation (tDCS), new therapeutic hope has emerged for pDoC patients (Shou et al., 2021). Nevertheless, TMS is contraindicated in patients with intracranial metallic foreign bodies, aneurysms, and integration into the routine treatment of acute and chronic phases is challenging. Furthermore, TMS and tDCS directly stimulate specific cortical areas, including the dorsolateral prefrontal cortex (DLPFC), precuneus, and other brain regions associated with consciousness (Shou et al., 2021; Peng et al., 2023), as opposed to the broader activation of the forebrain. Invasive VNS (iVNS) has been demonstrated to promote the release of intracranial neurotransmitters, to have a widespread effect on the forebrain, to improve frontal-parietal FC, and to facilitate the restoration of consciousness (Thibaut et al., 2019). These findings offer a promising new avenue for treating patients with DoC. Nevertheless, implantation can result in complications such as peri-incisional haematoma and dyspnoea. Furthermore, the high cost of the necessary equipment and surgical procedures represents a significant barrier to its adoption in clinical practice. The advent of a novel stimulation paradigm, non-invasive VNS (nVNS), has rendered its utilization in a clinical setting more feasible, particularly in terms of portability and the incidence of adverse effects.

Neuroimaging evidence indicates that nVNS may have similar effects to iVNS (Badran et al., 2018). The nVNS is contingent upon the distribution of vagal afferent nerve fibers in the skin and is categorized into two distinct stimulation modalities: transcutaneous auricular VNS (taVNS) and transcutaneous cervical VNS (tcVNS) (Silberstein et al., 2016; Wang et al., 2023). TcVNS is predominantly left-sided stimulation, and this left-sided bias is largely due to the early FDA approval of left-sided VNS for the treatment of epilepsy, as well as the belief that stimulation of the right side of the tcVNS induces direct electrical input to the sinus node, which may increase the risk of adverse effects, such as arrhythmias (Morris and Mueller, 1999; Peng et al., 2023). In contrast, taVNS—whether applied on the left or right side—stimulates the auricular branch of the vagus nerve (ABVN), which sends signals to the nucleus tractus solitarius (NTS) via afferent fibers, thereby supporting integrative processing of information (Chen et al., 2015). The NTS subsequently transmits processed signals to the heart via bilateral cervical vagus nerves, thus avoiding direct stimulation of cardiac pacing points (Chen et al., 2015). Brain network configuration changes and altered FC in patients with DoC following severe brain injury may, however, complicate the bottom-up effects of VNS (Vitello et al., 2023). Additionally, complex neural reorganization post-injury may disrupt both afferent and efferent VNS pathways.

Bilateral taVNS has demonstrated initial efficacy in conditions such as major depression (Tu et al., 2018) and epilepsy (He et al., 2015), with minimal side effects and straightforward localization. Thus, in this study, we employed bilateral taVNS to promote arousal in patients with DoC, aiming to explore its potential mechanisms in facilitating consciousness recovery.

Recent studies have demonstrated that the neuroprotective effects of taVNS can be achieved through anti-inflammatory mechanisms (Jiang et al., 2014; Kaczmarczyk et al., 2017), regulation of cerebral blood flow, reduction of blood-brain barrier permeability (Lopez et al., 2012; Yang et al., 2018), and modulation of the release of neurotrophic factors and neurotransmitters (Dong and Feng, 2018). In the treatment of DoC, Briand proposed the Vagal Cortical Pathways model, suggesting that taVNS activates the locus coeruleus and the raphe nuclei to release neurotransmitters such as norepinephrine and pentraxin, and extensively activates the lateral frontoparietal network (Briand et al., 2020). A study by Cavinato et al. (2015) demonstrated that simple sensory stimulation could enhance FC within the lateral frontoparietal network in patients with MCS, a response not observed in patients with VS/UWS. Interestingly, in this case, we observed that with the addition of taVNS treatment to conventional therapy, the patient showed enhanced FC of the frontoparietal network and showed clinical improvement. As Edlow noted in their review, the external frontoparietal network plays a pivotal role in consciousness (Edlow et al., 2021), which consists of the DLPFC and the posterior parietal cortices (PPC) (Zhou et al., 2016). This network is predominantly associated with external awareness, encompassing somatic sensory, visual and auditory perception. Given its foundational role in external awareness, the lateral frontoparietal network appears to be a crucial target in the rehabilitation of patients with DoC, potentially facilitating recovery of consciousness through its engagement with taVNS (Briand et al., 2020).

Currently, taVNS has been applied in the clinical management of DoC; however, its efficacy remains uncertain due to a limited understanding of the underlying neural mechanisms and substantial variability in patient responses. We reviewed nine existing trials that explored taVNS treatment for DoC patients (Yu et al., 2017, 2021; Hakon et al., 2020; Noé et al., 2020; Osińska et al., 2022; Yifei et al., 2022; Chen et al., 2023; Zhou et al., 2023; Wang et al., 2024). Our analysis revealed that patients receiving prolonged taVNS therapy demonstrated more substantial improvements in Coma Recovery Scale-Revised (CRS-R) scores and showed better prognoses (see Table 1). For example, in a case study by Osińska and colleagues, patients diagnosed with UWS for 2 years were treated with taVNS for 6 months. As a result, CRS-R scores improved from an initial 4–6 to 13, and the prognosis was favorable based on electroencephalography (EEG) results (Osińska et al., 2022). In contrast, a brief course of taVNS treatment, as demonstrated by Yifei, who treated 12 patients with DOC using taVNS for 4 weeks, revealed no improvement in CRS-R scores. However, the patients' EEGs exhibited an enhancement in their state of consciousness (Yifei et al., 2022). Accordingly, the authors posit that in practice, treatment-induced behavioral changes may be cumulative effects over time that are not immediately observable (Yifei et al., 2022). Other studies have demonstrated that a four-week course of taVNS is an effective method of accelerating the recovery of consciousness in patients with DoC, with a more pronounced effect observed in patients with MCS and a minimal effect in those with UWS (Yu et al., 2017, 2021; Noé et al., 2020; Zhou et al., 2023). Most of the current studies have treated patients with 4 weeks of taVNS (Yu et al., 2017, 2021; Noé et al., 2020; Yifei et al., 2022; Zhou et al., 2023; Wang et al., 2024), with fewer studies of long-term wake-promoting therapy. In response, we administered taVNS over 3 months to patients with a 50-day history of impaired consciousness. Following this intervention, patients transitioned from VS/UWS on admission to MCS+, with CRS-R scores improving from 2 to 14.

In this study, taVNS was administered for 1 h daily over 3 months to a patient diagnosed with VS/UWS. We observed that after approximately 90 days of stimulation, the patient's CRS-R score improved from an initial score of 2 to 14 and the diagnosis changed from VS/UWS to MCS+. Notably, taVNS was associated with minimal or no adverse effects on the patient. To investigate the potential modulatory effects of taVNS on neural activity in patients with DoC, we analyzed seven brain networks, examined changes in resting-state FC across 12 ROI, and compared mean HbO and HbR levels in each brain region during task states before and after 3 months of taVNS treatment. The results showed enhanced FC between the frontal and parietal cortices after treatment compared to before treatment. These findings suggest that taVNS may aid in consciousness recovery by enhancing FC between the frontal and parietal cortices.

The human brain is a complex, interconnected network system with a variety of important topological properties. Among them, the brain functional network is a typical small-world network (Li et al., 2019). The characteristics of high Cp and short Lp are conducive to the efficient transmission of information in the network (Suo et al., 2015). Functional segregation and integration are two major organizational principles of the human brain, the coexistence of which ensures the effective integration of multiple segregated sources of information in the brain (Sporns, 2011). Despite the fact that the functional brain networks demonstrated small-world properties both before and after treatment, Cp and Eloca exhibited smaller pre-treatment relative to post-treatment values across a wide range of sparsity levels. Lower pre-treatment Cp and Eloca values are indicative of patients exhibiting reduced functional segregation and less efficient and more unstable local messaging in the functional network (Luo et al., 2015). Eloc is thought to correlate with information processing and network fault tolerance between neighboring regions, and low Eloc usually corresponds to reduced fault tolerance and diminished immunity to interference in brain networks (Bullmore and Sporns, 2009). In this study, Cp, σ, Eg, and Eloc increased and Lp decreased as patient's level of consciousness improved after treatment, which is consistent with previous studies (Liu et al., 2023a).

The PFC has been identified as a pivotal center in higher cognitive functions in humans, including emotion regulation, attention and the formation of working memory (Farzan et al., 2009). In this study, we observed extensive activation in the PFC of patients after treatment by fNIRS. Several bedside assessments conducted before treatment showed no observable response to auditory stimuli, including family members calling or startle responses. However, when assessing consciousness using task-state fNIRS, we observed a sustained increase in mean HbO concentration in Wernicke's area (CH 13) during auditory tasks. Additional activations were noted in the RPFC (CH 5) and the RM1 (CH 18, 31). After treatment, activation in the RPFC (CH 19) became more extensive and intense, with significant activation also observed in the RS1 (CH 2, 16, 17), RM1(CH 1, 15), Broca's area (CH 19) and left Wernicke's area (CH 13). In addition, the present study found that patients could receive the physician's commands and activate the appropriate motor areas for feedback, possibly due to an impaired motor output pathway, which resulted in a diagnosis of unresponsiveness as the physician could not clinically observe the patient's motor output. Yifei showed that patients retaining hearing are more favorable for disease recovery (Yifei et al., 2022). In this study, the patient's response to instructions was detected using fNIRS and it can be surmised that the patient retained some auditory function and likewise the patient had a better prognosis.

pDoC are complex and refractory conditions, with an unclear pathogenesis that limits the optimization of therapeutic strategies. The precise mechanisms by which taVNS influences central nervous system (CNS) circuits are also not yet fully understood (Thibaut et al., 2019). This gap in understanding hinders the development of targeted therapeutic options. Notably, FC between the frontal and parietal brain regions has been shown to correlate significantly with the level of consciousness in DoC patients (Giacino et al., 2014). Studies further suggest that patients with VS/UWS can be distinguished from those in MCS based on interhemispheric connectivity within the frontoparietal cortex (Chennu et al., 2017). According to the mesocircuit hypothesis (Schiff, 2010), the frontal and parietal cortices act as critical hubs, supporting consciousness-related information processing within the cortical level of the consciousness circuit through frontoparietal connections. The frontal cortex plays a central role in orchestrating goal-directed behavior, regulating arousal across varying states, and activating the central thalamus through descending projections (Nee, 2021). This activation enables the brain to respond to higher cognitive demands. Additionally, parietal involvement has shown a significant correlation with levels of consciousness, indicating its role in supporting conscious states (Chennu et al., 2017). The FC between the frontal and parietal cortices enables these two cortices to regulate not only the mesocircuit directly through feedback but also indirectly through the frontal cortical-striatopallidal-thalamocortical loop systems (Schiff, 2010). This allows for the maintenance of normal conscious pathways in the brain (Liu et al., 2023b).

In clinical practice, the gold standard for assessing levels of consciousness in DoC patients is the CRS-R, which aids in differentiating between VS/UWS and MCS (Giacino et al., 2004). However, the CRS-R relies on the clinician's subjective interpretation of the patient's reflexive and non-reflexive behaviors in response to external stimuli. This subjectivity introduces variability, as different raters may assign different scores, and even the same rater might produce varying scores due to fluctuations in the patient's level of consciousness. Research indicates that up to 40% of patients with uncommunicative DoC may be misdiagnosed as being in VS/UWS, and inaccurate assessments of VS/UWS permanence can lead to a focus on management rather than rehabilitation, resulting in poorer prognoses (Wannez et al., 2017; Bodien et al., 2024). Furthermore, the evaluation of non-reflexive behaviors may not directly reflect the patient's true level of consciousness, as fluctuating consciousness levels, cognitive impairments (e.g., aphasia, dysarthria), sensory deficits (e.g., blindness, deafness), motor dysfunction, low activity levels, fatigue, and pain can all impact assessment accuracy (van Ommen et al., 2018). Consequently, DoC patients with motor or sensory impairments are at risk of misdiagnosis. Additionally, because CRS-R is based on behavioral assessment, it does not account for changes in brain function and thus cannot detect intrinsic improvements in consciousness. In a study, a large number of active consciousness paradigms in patients with DoC were investigated using EEG (Yifei et al., 2022). The study found that, despite an absence of visible signs of consciousness, 15% of VS/UWS patients were able to comply with instructions by modulating their brain activity. This highlights an urgent need for new methods to objectively and quantitatively assess consciousness from the perspective of brain function to enhance clinical diagnostic accuracy. fNIRS shows promise in providing insights into cognitive processes that may not be discernible through traditional behavioral assessments (Li et al., 2021; Lu et al., 2023). fNIRS is a non-invasive technique that monitors conscious brain activity by measuring near-infrared light absorption in cerebral tissue (Eastmond et al., 2022). Due to the differing absorption spectra of HbO and HbR at wavelengths between 650 and 950 nm, fNIRS can detect changes in their relative concentrations through diffuse light scattering, thereby inferring brain activity (Almajidy et al., 2020). Compared to EEG/MEG, fNIRS offers higher spatial resolution (1–2 cm), and compared to fMRI, it provides superior temporal resolution (milliseconds). Most importantly, fNIRS is portable, quiet, cost-effective, easy to administer, and tolerates movement artifacts and varied environments, making it ideal for prolonged or repetitive bedside measurements in DoC patients (Song et al., 2018). Therefore, fNIRS-related metrics offer a viable method for assessing the effects of taVNS on brain function in DoC patients.

When analyzing subdomains of the CRS-R scale, we observed that improvements in patient scores were primarily centered on visual and auditory function. Using an auditory paradigm, we examined cortical responses to auditory stimuli in patients with DoC to explore potential correlations between these responses and patient outcomes. Our study found that musical stimulation activates the frontal and parietal cortices, which subsequently enhances FC within the cerebral cortex, supporting the hypothesis proposed by Wojtecki that simple somatosensory, auditory, and visual stimuli typically activate primary cortical areas in VS/UWS patients (Wojtecki et al., 2014). Beyond speech, music is one of the most versatile and complex auditory stimuli, engaging auditory, visual, and somatosensory processing systems (Rollnik and Altenmüller, 2014). In this study, different types of music were played for patients, revealing variations in cortical connectivity edges based on the type of music. This suggests that the brain may process different music types in a basic manner that elicits arousal and emotional responses, with emotional improvement potentially linked to functional recovery, as observed in stroke patients. Several previous studies have reported that in healthy subjects, listening to music induces arousal and mood changes in the listener, which improves the listener's cognitive abilities (e.g., verbal processing, visual attention, etc.) (Soto et al., 2009; Gómez-Gallego et al., 2021), and there is neuroimaging evidence to support this notion, whereby listening to pleasing music activates limbic and paralimbic structures (e.g., the hippocampus), as well as structures involved in reward circuits (e.g., nucleus of the arcus accumbens), leading to a person's feeling of comfort and calm (Särkämö et al., 2008; Bleibel et al., 2023). Similarly, musical stimulation has been reported to induce broad cortical and subcortical activation in DoC patients, particularly in the bilateral frontal, temporal, parietal, and cerebellar regions, with strong responses in the frontal cortex, cingulate gyrus, amygdala, and hippocampus—areas closely associated with emotion (Rollnik and Altenmüller, 2014). In this study, musical stimulation in DoC patients enhanced bilateral FC between the frontal and parietal cortices, with the most pronounced effect seen when patients listened to their dislike music. It is hypothesized that music influences neural networks, enhances synaptic plasticity, and improves perceptual processing (Sihvonen et al., 2017). Therefore, music therapy-especially with dislike music-could potentially be employed to promote arousal during the early neurological rehabilitation of DoC patients. Further research is needed to establish the therapeutic potential of music therapy for early rehabilitation in patients with coma, VS/UWS, or MCS.

This study attempted to explore the evidence for the benefit of bilateral taVNS for VS/UWS treatment lasting 90 days. We observed a gradual increase in patient's CRS-R scores during taVNS therapy. The frontal-parietal FC of the patients was enhanced after treatment compared with before treatment, and we speculate that this may be one of the reasons for the improvement in the patient's consciousness. Therefore, we hypothesized that long-term taVNS treatment may be an adjunctive therapy for impaired consciousness due to hemorrhagic stroke. Further studies in larger patient populations are recommended to confirm these findings and refine the taVNS protocol in clinical practice.

This study has several limitations. First, the small sample size limits our ability to conclusively attribute improvements in consciousness to taVNS alone, as we cannot fully rule out the possibility of spontaneous recovery. Additionally, fNIRS has its own limitations; for instance, it cannot assess subcortical brain function recovery or activation levels in the thalamus and brainstem. Future research should incorporate multimodal assessments of DoC patients and further explore the effects of taVNS on subcortical FC in this population. To establish the efficacy of taVNS, larger-scale randomized controlled trials are necessary. Such studies should aim to optimize the stimulation protocol, evaluate whether binaural taVNS is more effective than monaural stimulation, and determine whether therapeutic effects vary according to etiology and lesion type.