Blink reflex as a prognostic indicator for patients with prolonged disorders of consciousness: a prospective case-control study

Abstract

Objective

This study aimed to investigate the characteristics of the blink reflex and its prognostic value for consciousness improvement in patients with prolonged disorders of consciousness (DOC).

Design

Prospective case-control study.

Setting

This study was conducted in a local hospital between March 2022 and March 2023.

Participants

Patients in a vegetative state/unresponsive wakefulness syndrome or in a minimally conscious state were enrolled within 3 months from their brain injury.

Primary and secondary outcome measures

The early component (R1), ipsilateral late component (iR2) and contralateral late component (cR2) responses at baseline were recorded using electromyography. The patients’ clinical diagnosis and the best Coma Recovery Scale-Revised (CRS-R) total score were assessed based on Chinese CRS-R evaluations.

Outcome definition

At the 6-month follow-up, patients were categorised as improved or non-improved based on CRS-R score changes (improved (transition to a higher consciousness state); non-improved (worsened condition, static or death)).

Results

A total of 58 DOC patients were included in this study. Of the 58 DOC patients, 32 were classified as the improved group and 26 as the non-improved group. In the improved group, R2 responses were elicited in 30 patients, while only 16 patients in the non-improved group had elicited R2 responses. The non-improved group exhibited significantly lower R2 mean amplitudes (iR2 (105.08 µV vs 173.25 µV, p=0.01); cR2 (55.15 µV vs 114.03 µV, p=0.01)) and longer mean latencies (iR2 (41.08±6.72 ms vs 37.77±3.94 ms, p=0.03); cR2 (41.32±6.28 ms vs 37.48±4.07 ms, p=0.01)) compared with the improved group. The result demonstrated that the iR2 mean amplitude (OR=1.01, area under the curve (AUC)=0.78 (95% CI 0.63 to 0.93), sensitivity=78.12%, specificity=83.33%, p=0.02) and cR2 mean amplitude (OR=1.02, AUC=0.76 (95% CI 0.62 to 0.90), sensitivity=81.25%, specificity=72.22%, p=0.02) were significant predictors of consciousness improvement. Meanwhile, Pearson correlation analysis revealed that iR2 mean amplitude (r=0.42, p=0.003) and cR2 mean amplitude (r=0.53, p=0.001) significantly correlated with CRS-R score at baseline.

Conclusion

The R2 amplitude in patients with prolonged DOC may serve as a prognostic indicator for consciousness improvement.

Strengths and limitations of this study.

• The prospective case-control design enables comparison between improved and non-improved disorders of consciousness (DOC) patients, offering valuable insights into blink reflex responses and recovery patterns.

• The use of objective electrophysiological assessments provides reliable biomarkers for predicting recovery in DOC patients.

• Blinded evaluations and a 6-month follow-up period minimise bias and track patient recovery, enhancing the reliability of the findings.

• The single-centre design, sample size and participant heterogeneity may limit the generalisability and broader applicability of the results.

Introduction

Disorders of consciousness (DOC) encompass a spectrum of conditions characterised by impaired consciousness to varying degrees, primarily resulting from severe brain injuries. These conditions range from minimally conscious states (MCS) to complete unresponsiveness.¹ Recent epidemiological studies in the USA indicate a rising prevalence of DOC, which can be attributed to advancements in emergency and critical care interventions. Approximately 2.9 million individuals present to emergency departments annually for head trauma, with 288 000 of these patients requiring hospitalisation. This trend underscores an increasing incidence of brain injuries along with improved survival rates.² However, the long-term outcomes for these patients are often suboptimal, resulting in considerable burdens on healthcare systems, families and caregivers. This situation is exacerbated by the extensive care required and the accompanying psychosocial, economic, ethical and legal challenges.³

The blink reflex (BR), an involuntary, rapid eyelid closure induced by external stimuli, is mediated through brainstem pathways and potentially influenced by higher cortical functions. It serves as a neurophysiological test for assessing the integrity of the trigeminal and facial nerves and their brainstem interconnections. Neurocognitive impairments in the subcortical regions, including the brainstem, significantly influence BR. These abnormalities may serve as indicators of dysfunction within the central nervous system.⁴ Recent studies suggest the utility of BR in evaluating brainstem function in DOC patients. Deviations in BR parameters from normal values may predict neurological recovery or deterioration, emphasising its potential as a prognostic tool in a clinical setting.^{5 6}

Despite this potential, the application of BR for DOC prognosis remains underexplored. Previous studies have predominantly focused on structural imaging7,9 and behavioural assessments,10,12 such as the Coma Recovery Scale-Revised (CRS-R). However, CRS-R assessments are constrained by a high misdiagnosis rate (43%) due to factors including motor impairments, aphasia and fluctuating arousal levels, which complicate the evaluation of consciousness.¹³ Electrophysiological assessments, such as BR, can provide vital complementary information, enabling continuous bedside monitoring unaffected by these complicating factors. Acute brain injuries could result in pronounced suppression of both iR2 and cR2 components.¹⁴ Previous studies indicate that multimodal evoked potentials and BRs increase the specificity of the diagnosis of brainstem injury compared with clinical observation only and improve prognostic reliability.¹⁵ Similar studies have found that early recovery of the BR after primary brain stem injury was a further favourable sign. Consistent absence of R2 in prolonged coma indicated a bad outcome.¹⁶ While the traditional binary classification (presence or absence of BR responses) maintains its clinical utility in the acute stage of coma, it remains uncertain whether incorporating quantitative analysis of BR parameters could improve prognostic accuracy in patients with PDOC.

This study aimed to evaluate whether BR parameters can offer a more refined prediction of recovery potential in patients with prolonged DOC.

Methods

Patient and public involvement in research

Patients or the public were not involved in the design, conduct, reporting or dissemination plans of our research.

Patients and study design

 This prospective case-control study enrolled patients with prolonged DOC admitted to the Critical Care Rehabilitation Department of Jiangning Hospital, Nanjing Medical University from March 2022 to March 2023. The sample size was determined a priori using an independent samples t-test based on pilot data and previous studies.17,19 We hypothesised a large effect size (Cohen’s d=0.8) based on prior studies reporting similar effects for R2 amplitude differences (Cohen’s d 0.8²⁰ ～1.3²¹). Power analysis was performed using G*Power 3.1 software with the following parameters: two-tailed t-test, effect size d=0.8, α=0.05, power (1-β)=80%, anticipated improvement rate of 50%^{22 23} and an estimated 10% dropout rate. The calculation yielded a minimum required sample size of 58 completers. To ensure adequate power, we enrolled 66 participants, of whom 58 completed all follow-up assessments and were included in the final analysis. Inclusion criteria for patients were: (1) aged 18–75 years; (2) diagnosed with traumatic brain injury,²⁴ intracerebral haemorrhage²⁵ or hypoxic-ischaemic encephalopathy;²⁶ (3) clinical diagnosis of vegetative state (vs) or MCS confirmed by a neuropsychologist using five consecutive Chinese CRS-R assessments per week;^{27 28} (4) 1–3 months post-injury; and (5) stable clinical condition. Exclusion criteria for both groups were: (1) pre-existing neurological diseases; (2) unstable clinical conditions (eg, severe cardiac stress or respiratory failure); (3) use of sedatives, anti-epileptic drugs or neurological stimulants; (4) incomplete clinical data; (5) declined participation; and (6) refusal to be followed up. The study flowchart is presented in figure 1.

Figure 1. Flow diagram illustrating the patient screening and exclusion criteria. BR, blink reflex; CRS-R score, Coma Recovery Scale-Revised score; R1, early component; R2, late component.

All participants’ legal caregivers provided written informed consent. The study adhered to the Declaration of Helsinki II and good clinical practice guidelines and was approved by the Ethics Committee of Jiangning Hospital, Nanjing Medical University (2022-03-047-k01).

Data collection and outcomes

Clinical assessment

At study entry, demographic, clinical and neurophysiological data were systematically collected and analysed. Age (years), intensive care unit (ICU) stay (days) and disease duration (days post-injury) were calculated from hospitalisation records and recorded as continuous variables. Aetiology was classified as stroke, traumatic or anoxic through comprehensive medical record review and neuroimaging confirmation (CT/MRI). Within 1 week from study entry, all patients were assessed by skilled investigators by means of repeated (at least five in a week) CRS-R evaluations in order to confirm the patients’clinical diagnosis and to determine the best CRS-R total score (range 0–23). Patients were clinically followed up at 6 months after study entry by repeated CRS-R conducted by the same clinical team from their hospital stay or, if after discharge, in an outpatient follow-up clinic, at home or in chronic care facilities.

Neurophysiological examination

For BR examination via electromyography, active electrodes were bilaterally placed on the orbicularis oculi muscle, with reference electrodes positioned laterally to the lateral canthus. Right and left supraorbital nerves were stimulated over the supraorbital foramen. Electrical stimuli parameters were set at 20–40 mA intensity (approximately three times the subjective sensory threshold),²⁹ with a duration of 0.1 ms and interstimulus intervals varying arrhythmically between 5 and 10 s. Responses were elicited eight times consecutively on both sides.

Early component (R1), ipsilateral late component (iR2) and contralateral late component (cR2) responses were recorded. Latency and amplitude were measured, and averages were calculated.³⁰ Given the potential asymmetric effects of brain injury on the hemispheres, we calculated mean BR latency and BR amplitudes separately for left and right sides. Signal processing included offline analysis with a band-pass filter of 10–5000 Hz, sensitivity of 200 µV/Div and sweep speed of 10 ms/Div. Two blinded authors independently performed electrophysiological evaluations. Discrepancies between evaluations were resolved through discussion.

Outcome definition

The primary outcome was a patient’s clinical diagnosis at 6 months following their enrolment. Based on the CRS-R scores, patients were categorised into unresponsive wakefulness syndrome/vegetative state (vs/UWS), MCS, or emergence from MCS (EMCS).³¹ Patients in the vs/UWS category exhibited sleep-wake cycles, autonomic nerve activity and motor reflexes but lacked self-awareness and awareness of their surroundings.³² In contrast, patients classified as MCS demonstrated clear evidence of conscious behaviours, such as following simple commands or responding to noxious stimuli.³³ Those reaching EMCS were capable of functional communication or purposeful object use, indicating a more advanced level of recovery.¹ For statistical analysis, outcomes were classified as ‘Improved’ for patients whose clinical diagnoses advanced during the study period (eg, transitions from vs/UWS to MCS, MCS to EMCS or EMCS to full consciousness). Conversely, outcomes were classified as ‘Non-improved’ for patients whose condition either worsened (eg, transitions from MCS to vs/UWS), remained static or resulted in death during the study period.³⁴

Statistical analysis

Statistical analyses were performed using IBM SPSS Statistics V.22.0 (IBM Corp., Armonk, NY, USA). Descriptive statistics were employed to summarise demographic and clinical data. The normality of data distributions was assessed using the Shapiro–Wilk test. Continuous variables with normal distributions were expressed as means with SD, while variables with non-normal distributions were presented as medians with interquartile ranges (25th–75th percentiles). Categorical variables were summarised as frequencies and percentages. BR latencies/amplitudes were compared using independent samples t-tests (normally distributed) or Mann-Whitney U-tests (non-normal). To determine the independent predictive value of BR parameters, we performed multivariate logistic regression analysis following a systematic approach: first, we identified a priori covariates (age, sex, disease duration) based on established clinical relevance in DOC prognosis. These were retained in all models regardless of statistical significance to ensure adjustment for fundamental demographic and temporal factors. Additional potential confounders were evaluated through univariate screening (p>0.2 threshold for exclusion) and assessed for clinical plausibility. Given our sample size constraints (n=58) and population characteristics (72.4% stroke aetiology), we excluded variables showing: (1) non-significant univariate associations (ICU stay duration, p=0.68), (2) minimal between-group differences at baseline (comorbidities, table 1) or (3) potential to introduce collinearity or model instability (aetiology). The final model thus included iR2/cR2 amplitudes with adjustment for age, sex and disease duration. The prognostic validity of BR parameters for differentiating between improved and non-improved patients was determined by receiver operating characteristic (ROC) curves analysis. Finally, correlation analyses were conducted in two stages: initial Pearson correlation coefficients (r) assessed unadjusted linear relationships between BR parameters and baseline CRS-R scores, followed by partial correlation analysis (adj.r) adjusting for potential confounding factors including age, sex, disease duration and aetiology. R value was used to represent the strength of correlation. According to Cohen’s classification, r=0 was no correlation, r≥0.20 was a weak correlation, r≥0.50 was a moderate correlation and r≥0.80 was a strong correlation.³⁵ Statistical significance was defined as p<0.05 for two-sided tests.

Table 1. Baseline characteristics of all participants.

Variable	Non-improved group (n=26)	Improved group (n=32)	P value
Sex
Male	17 (65.4%)	19 (59.4%)	0.42
Female	9 (34.6%)	13 (40.6%)
Mean age in years	58.27±15.11	58.56±13.41	0.94
Disease duration (days)	46.32±24.33	46.81±28.26	0.18
ICU stay (days)	24.19±11.85	23.78±16.91	0.91
Aetiology
Stroke	17 (65.4%)	25 (78.1%)	0.11
Traumatic	4 (15.4%)	6 (18.8%)
Anoxic	5 (19.2)	1 (0.1%)
Craniectomy
No	14 (53.8%)	19 (59.4%)	0.44
Yes	12 (46.2%)	13 (40.6%)
Smoking history
No	13 (50.0%)	18 (56.3%)	0.42
Yes	13 (50.0%)	14 (43.7%)
Alcohol drinking history
No	14 (53.8%)	16 (50.0%)	0.48
Yes	12 (46.2%)	16 (50.0%)
Hypertension
No	8 (30.8%)	4 (12.5%)	0.12
Yes	18 (69.2%)	28 (87.5%)
Diabetes
No	22 (84.6%)	30 (93.8%)	0.24
Yes	4 (15.4%)	2 (6.3%)
Coronary heart disease
No	23 (83.5%)	30 (93.8%)	0.41
Yes	3 (11.5%)	2 (6.3%)
Diagnose
vs/UWS	19 (73.1%)	5 (15.6%)	<0.001
MCS	7 (26.9%)	27 (84.4%)
CRS-R score
baseline	6.96±2.14	11.19±3.73	<0.001

CRS-R, Coma Recovery Scale-Revised.ICU, intensive care unit; MCS, minimally conscious states; vs/UWS, vegetative state/unresponsive wakefulness syndrome.

Results

All 58 patients completed the 6 months follow-up (figure 1). The study included 58 patients with prolonged DOC; the non-improved group comprised 26 patients (65.4% male, mean age 58.27±15.11 years; duration 46.32±24.33 days), while the improved group included 32 patients (59.4% male, mean age 58.56±13.41 years; duration 46.81±28.26 days). Baseline clinical characteristics, including disease duration, ICU stay duration, aetiology, alcohol use and history of hypertension, diabetes or coronary heart disease, did not differ significantly between outcome groups. At baseline assessment, the non-improved group contained a significantly higher proportion of vs/UWS patients and showed lower total baseline CRS-R scores compared with the improved group (p<0.05) (table 1).

Of the 58 DOC patients, 46 elicited both R1 and R2 responses and were included in the final analysis. This group included 16 non-improved patients and 30 improved patients. Compared with the improved group, the non-improved group exhibited a lower R1 elicitation rate (24/26 vs 32/32, p=0.19) and a significantly lower R2 elicitation rate (16/26 vs 30/32, p=0.003) (figure 1). Meanwhile, the non-improved group exhibited significantly longer mean R2 latencies and lower mean R2 amplitudes, compared with the improved group (iR2 mean latency (41.08±6.72 ms vs 37.77±3.94 ms, p=0.03); cR2 mean latency (41.32±6.28 ms vs 37.48±4.07 ms, p=0.01); iR2 mean amplitude (105.08 (68.65, 135.75) µV vs 173.25 (147.88, 232.74) µV, p=0.01); cR2 mean amplitude (55.15 (43.78, 84.9) µV vs 114.03 (81.01, 207.18) µV, p=0.01)). No significant differences were observed in R1 mean latency or mean amplitude between the two groups (p>0.05) (table 2).

Table 2. BR parameters comparing the non-improved group and the improved group at baseline.

Variable	Non-improved group	Improved group	P value
R1mean latency, ms	11.61±1.45	11.43±1.25	0.64
iR2 mean latency, ms	41.08±6.72	37.77±3.94	0.03
cR2 mean latency, ms	41.32±6.28	37.48±4.07	0.01
R1 mean amplitude, µV	159.6 (102.76, 325.25)	208.75 (143.88, 299.38)	0.84
iR2 mean amplitude, µV	105.08 (68.65, 135.75)	173.25 (147.88, 232.74)	0.01
cR2 mean amplitude, µV	55.15 (43.78, 84.9)	114.03 (81.01, 207.18)	0.01

cR2, contralateral late component; iR2, ipsilateral late component; R1, early component.

To assess the role of BR parameters in predicting conscious improvement, we performed a univariate and multivariate logistic regression analysis, with non-improved conscious or improved conscious based on the CRS-R scores as the dependent variable. After adjusting for age, sex and disease duration, the results demonstrated that the iR2 mean amplitude (OR=1.01,p=0.02) and cR2 mean amplitude (OR=1.02,p=0.02) were significant predictors of conscious improvement after 6-month follow-up (table 3).

Table 3. Logistic regression analysis for conscious improvement prediction in PDOC patients.

Variable	Univariate logistic regression		Multivariate logistic regression
	OR (95%CI)	P value	OR (95%CI)	P value
Gender: 1 for males, 2 for female	1.06 (0.36,3.18)	0.91	1.87 (0.47,8.26)	0.38
Age	0.99 (0.95,1.03)	0.63	0.97 (0.91,1.02)	0.26
Disease duration	1.00 (0.99,1.02)	0.17	0.96 (0.89,1.02)	0.23
ICU stay	0.99 (0.96,1.03)	0.68	–	–
Aetiology: 1 for stroke, 2 for not stroke	0.79 (0.23,2.86)	0.72	–	–
R1 mean latency	0.89 (0.59,1.33)	0.58	–	–
iR2 mean latency	0.92 (0.82,1.02)	0.13	–	–
cR2 mean latency	0.98 (0.91,1.05)	0.52	–	–
R1 mean amplitude	1.00 (0.99,1.01)	0.53	–	–
iR2 mean amplitude	1.01 (1.00,1.02)	0.02	1.01 (1.00,1.03)	0.02
cR2 mean amplitude	1.02 (1.02,1.03)	0.01	1.02 (1.00,1.03)	0.02

Note: Multivariate model adjusted for age, sex and disease duration. Other variables were excluded due to non-significant univariate associations (ICU stay), population homogeneity (aetiology) or sample size limitations.

CRS-R, coma recovery scale-revised.cR2, contralateral late component; ICU, intensive care unit; iR2, ipsilateral late component; R1, early component.

ROC analysis demonstrated the predictive utility of R2 mean amplitudes at the 6-month follow-up. The optimal threshold of the iR2 mean amplitude was 147.25 µV, yielding an area under the curve (AUC)-ROC value of 0.78 (95% CI 0.63 to 0.93), with a sensitivity of 78.12% and specificity of 83.33%. For the cR2 mean amplitude, the threshold was 64.65 µV, demonstrating an AUC-ROC of 0.76 (95% CI 0.62 to 0.90), with 81.25% sensitivity and 72.22% specificity (figure 2).

Figure 2. ROC curves for R2 mean amplitudes in predicting conscious improvement at 6 months post-enrolment. The area under the ROC curve (AU-ROC) for iR2 mean amplitude and cR2 mean amplitude was 0.78 (95% CI 0.63 to 0.93) and 0.76 (95% CI 0.62 to 0.90), respectively. cR2, contralateral late component; iR2, ipsilateral late component.

To further clarify the relationship between R2 amplitudes and CSR-R scores, we first conducted Pearson correlation analysis, which revealed that the mean amplitudes of iR2 and cR2 showed positive correlations with baseline CRS-R scores (r=0.42, p=0.003; r=0.53, p=0.001) (figure 3). After adjusting for age, sex, disease duration and aetiology, partial correlation analysis revealed that the mean amplitudes of iR2 and cR2 showed positive correlations with baseline CRS-R scores (adj.r=0.42, p=0.004; adj. r=0.48, p=0.001).

Figure 3. Correlation analysis between CRS-R scores and R2 mean amplitudes. cR2, contralateral R2; CRS-R score, Coma Recovery Scale-Revised score; iR2, ipsilateral R2; R2, late component.

Discussion

Prolonged DOC, which are often caused by severe brain injury, always lead to severe complications, poor clinical outcomes and high mortality. Predicting clinical improvement in patients with PDOC is very helpful in supporting clinicians to offer suitable clinical care and aiding in the decision making of patients and their family members.³⁶ Although several previous studies aimed at identifying the risk factors that affect poor clinical outcomes, severe neurological deficit or awakening from coma, quantitative assessments of BR parameters in patients with PDOC remain conspicuously lacking in existing neurophysiological studies. Thus, our present study aimed to build and validate BR characteristics for predicting improved clinical outcomes in patients with PDOC.

Consciousness is regulated by a complex network of neural mechanisms involving multiple brain structures.^{37 38} The ascending reticular activating system (ARAS) has long been recognised as the primary neural substrate for maintaining consciousness. In patients with severe brain injuries, DOC may result from transient dysfunction or structural damage to the ARAS.^{39 40} This intricate network integrates partial brainstem reticular formations with thalamic nuclei, basal forebrain structures and the cerebral cortex.³⁷ Given this neuroanatomical framework, brainstem functional assessment may provide valuable prognostic information regarding consciousness recovery in chronic DOCs. The BR, as an objective electrophysiological measure, offers a reliable means of evaluating brainstem integrity post-injury.

Recent research underscores the importance of BR responses in various neurological conditions, demonstrating their role as reliable biomarkers for assessing neurological status and predicting outcomes.^{21 41 42} In our study, no significant differences were observed in R1 mean latency or mean amplitude between the two groups (p>0.05). The non-improved group exhibited significantly lower R2 mean amplitudes and longer R2 mean latencies compared with the improved group (all p<0.05). A particularly pertinent study¹⁴ demonstrated that acute brain injuries result in pronounced suppression of both iR2 and cR2 components, indicating significant disruptions in brainstem reflex pathways immediately following injury. Previous studies have demonstrated strong correlations between BR parameters and the clinical evaluation of stages of coma.¹⁶ One is the early unilateral oligosynaptic response R1, the other the late bilateral polysynaptic response R2. Although both the early and late reflexes are altered by brain stem lesions, the late reflex is particularly vulnerable to altered states of consciousness and is either totally absent, or at best minimal in amplitude regardless of the site of the responsible lesion and the side and the intensity of stimulation. These findings align with our observations of prolonged R2 latency and reduced R2 amplitude in PDOC patients, thereby reinforcing the BR abnormalities in PDOC patients.

Neuroanatomical and neurofunctional studies^{43 44} have emphasised the crucial role of brainstem-cortical interactions in modulating BR, providing a structural and functional basis for understanding BR abnormalities in our patients. Additionally, studies focusing on neurochemical dynamics,45,47 particularly within the dopaminergic and GABAergic systems, have explored how neurotransmitter dysregulation contributes to BR abnormalities in DOC patients. Ageing and dopamine depletion have also been linked to changes in reflex latency and amplitude, with older individuals and those with cognitive decline showing altered BR responses.^{46 48} Patients with mild cognitive impairment exhibit an increased blink rate, attributed to dopamine activity.⁴⁹

Unlike previous studies that primarily focused on the acute stage of coma,^{50 51} our study highlights the importance of baseline BR parameters as potential prognostic indicators for long-term outcomes in PDOC patients. Many factors are known to influence the prognosis of patients with PDOC.⁵² Prior studies suggest that age, female sex and⁵³ traumatic aetiology⁵⁴ are related to better clinical outcomes, but our multivariate analysis showed age and sex were not an independent predictor for consciousness recovery after adjusting for BR parameters (p>0.05). This discrepancy may reflect our cohort’s predominance of stroke-related DOC (72.4%), where age effects could be masked by primary injury severity. Future studies with larger, aetiology-stratified samples are needed to clarify these relationships. Remarkably, R2 mean amplitudes retain their strong prognostic value (AUC 0.76–0.82) across varying age and sex groups, underscoring the clinical utility of BR as an objective biomarker. While both pupillary light reflex (PLR) and brainstem auditory evoked potentials remain valuable for evaluating brainstem integrity, their predictive capacity for functional outcomes appears more limited. Notably, Lee et al reported PLR achieved only moderate accuracy (AUC=0.744) in predicting severe neurological deficits in comatose patients,⁵⁵ lower than our R2 mean amplitudes findings. Pearson and partial correlation analysis revealed that the mean amplitudes of R2 showed positive correlations with CRS-R scores (p<0.05). Notably, R2’s polysynaptic circuitry-integrating brainstem, thalamic and cortical pathways may³⁰ explain its sensitivity to axonal injury, a key determinant of recovery independent of demographic factors. A prior study⁵⁶ investigated the role of R1 in neurological prognosis and concluded that while R1 was typically maintained across varying severities of brain injury, it was not a sufficient standalone predictor of recovery outcomes. This conclusion is consistent with our findings, which revealed no significant differences in R1 responses between the two groups, and R1 did not demonstrate prognostic value. Comparative analyses with similar studies⁵⁷ support our results, particularly regarding the association of prolonged latencies and reduced amplitudes in iR2 and cR2 with poorer recovery in brain injury patients.

Recent studies have enhanced our understanding of sensory processing in DOC, showing that altered pain perception in severe brain injury patients may indicate broader sensory integration dysfunctions.⁵⁸ Similarly, research has highlighted the complexity of sensory abnormalities in DOC, emphasising their value in assessing and managing sensory functions.⁵⁷ Our findings align with these insights, as abnormalities in the R2 response of the BR—indicative of impaired nociceptive A-delta fibre conduction—may correlate with patient outcomes. These correlations support the nociceptive sensory pathway as a valuable prognostic marker in neurological disorders. Our study contributes to this by showing that baseline BR amplitude is strongly correlated with prognosis in PDOC patients. These findings are supported by evidence linking BR patterns to recovery likelihood.⁵⁹ This underscores the utility of BR as a critical tool for clinicians in evaluating sensory integration dysfunctions associated with severe brain injuries.

Several limitations of this study warrant consideration. First, the single-centre design and relatively small sample size (n=58), combined with the lack of stratification by aetiology of DOC patients, may limit the external validity and generalisability of our findings. Second, the interpretation of the CRS-R) results requires particular caution. The scale’s documented limitations in precisely differentiating awareness levels, coupled with reported misdiagnosis rates approaching 43% based on total CRS-R scores, suggest that the observed moderate correlations with BR amplitudes may reflect behavioural assessment constraints rather than genuine neurophysiological relationships. These methodological challenges highlight critical directions for future research, including: multicentre collaborations to recruit larger, more aetiologically diverse patient cohorts; advanced phenotyping of structural and functional brain injury characteristics using multimodal imaging; and systematic validation of specific BR patterns against standardised behavioural and neurophysiological measures of consciousness.

In conclusion, R2 amplitude shows promise as prognostic markers for consciousness recovery in prolonged DOC, but their clinical utility must be further validated in studies with rigorous adjustment for confounders. The findings should be interpreted as exploratory, warranting confirmation in larger multicentre cohorts.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.