New Behavioral Signs of Consciousness in Patients with Severe Brain Injuries

Diagnostic and prognostic assessment of patients with disorders of consciousness (DoC) presents ethical and clinical implications as they may affect the course of medical treatment and the decision to withdraw life-sustaining therapy. There has been increasing research in this field to lower misdiagnosis rates by developing standardized and consensual tools to detect consciousness. In this article, we summarize recent evidence regarding behavioral signs that are not yet included in the current clinical guidelines but could detect consciousness. The new potential behavioral signs of consciousness described here are as follows: resistance to eye opening, spontaneous eye blink rate, auditory localization, habituation of auditory startle reflex, olfactory sniffing, efficacy of swallowing/oral feeding, leg crossing, facial expressions to noxious stimulation, and subtle motor behaviors. All of these signs show promising results in discriminating patients' level of consciousness. Multimodal studies with large sample sizes in different centers are needed to further evaluate whether these behaviors reliably indicate the presence of consciousness. Future translation of these research findings into clinical practice has potential to improve the accuracy of diagnosis and prognostication for patients with DoC.

state/unresponsive wakefulness syndrome (VS/UWS), minimally conscious state minus (MCS À ), minimally conscious state plus (MCS þ ), and emergence from MCS (eMCS). The acute period of DoC is defined as the first 28 days after the brain injury, and subacute-to-chronic (or "prolonged" DoC) as longer than 28 days. 7 Coma is defined as the complete absence of arousal and awareness. 8 VS/UWS is defined as preserved arousal (eye opening spontaneously or upon stimulation) without awareness, the patient showing only reflexive behaviors. [9][10][11] MCS is defined as minimal, reproducible but inconsistent behavioral signs of awareness. 12 MCSÀ patients show nonreflex movements such as localization of noxious stimuli, visual pursuit or fixation, localization of objects, and movement or affective behaviors in a contextual manner to relevant environmental stimuli. MCSþ patients display behaviors related to language expression and comprehension, including command-following, intelligible verbalization, and intentional communication. 13 When patients demonstrate functional object use or functional communication, they are considered eMCS. 12 Cognitive motor dissociation (CMD), 14 functional locked-in syndrome, 13 MCS Ã , 15,16 and covert cognition 17 are terms suggested by different research teams to define behaviorally unresponsive patients who show brain activity compatible with (minimal) consciousness detected by functional magnetic resonance imaging (fMRI), electroencephalography (EEG), or positron emission tomography (PET). CMD is specifically used for patients who show no (VS/UWS) or little (MCS À ) behavioral evidence of consciousness at the bedside but have cortical responses related to language processing in fMRI or EEG active paradigms. 14 MCS Ã encompasses VS/UWS patients with CMD as well as VS/UWS patients who have residual brain activation in neuroimaging compatible with diagnosis of MCS even in the absence of active paradigms. 15,16 Accurate diagnosis of DoC patients is highly challenging. Indeed, over one-third of DoC patients previously diagnosed with VS/UWS by clinical consensus (based on behavioral observations and clinical experience) had evidence of consciousness when they were later evaluated based on standardized behavioral assessments using the Coma Recovery Scale-Revised (CRS-R). [18][19][20] However, it is important to note that, if performed only once, standardized behavioral assessment with the CRS-R can also lead to a 35% rate of misdiagnosis (compared with five CRS-R assessments). 21 Thus, repeated assessment (at least five times in a short period, e.g., 10 days) is of critical importance. The high misdiagnosis rates in DoC patients may be related to the lack of a proper gold standard to assess the presence of consciousness, and the need to integrate neuroimaging and new potential behavioral signs of consciousness into diagnostic guidelines. All these warrant the urgent need for improvement in diagnostic methods.
Currently available standardized behavioral scales to assess patients with DoC include, among others, the Glasgow Coma Scale (GCS), 8 the Full Outline of UnResponsiveness, 22 the Simplified Evaluation of CONsciousness Disorders (SECONDs), 23,24 and the CRS-R. 25 The American Congress of Rehabilitation Medicine Task Force and Euro-pean Academy of Neurology recommends the use of repeated CRS-R in the assessment of patients with subacute-tochronic DoC. 26,27 The CRS-R consists of 23 items composed of six subscales assessing auditory, visual, motor, oromotor/verbal functions, communication, and arousal. Among these items, 11 indicate an MCS diagnosis (six for MCSÀ and five for MCS þ ). From these items, visual pursuit, reproducible command-following, and automatic motor response (e.g., nose scratching, grasping bedrail, grabbing tubes) were found to be the first three most common signs of MCS to reemerge after brain injury. 28 The five most frequently observed items detecting 99% of chronic MCS patients were visual fixation, visual pursuit, reproducible movement to command, automatic motor response, and localization to noxious stimulation. 29 When transitioning into MCS, chronic VS/UWS patients were found to show mostly only one behavioral sign (73%): visual fixation, visual pursuit, localization to noxious stimulation, reproducible movement to command, or functional communication. 30 Similarly, chronic MCS patients were also found to show mostly only one behavioral sign (64%) while transitioning into eMCS, either functional communication or functional object use 30 (►Fig. 1).
In addition to the already available items that denote MCS in the CRS-R, recent studies suggest that other behaviors may be considered as signs of consciousness in DoC patients. The objective of this article is to summarize and review these new behavioral findings: resistance to eye opening, spontaneous eye blink rate, auditory localization, habituation of auditory startle reflex, olfactory sniffing, swallowing/oral feeding, facial expressions to noxious stimulation, subtle motor behavior assessed by Motor Behavioral Tool-revised (MBT-r), and leg crossing (►Table 1 and ►Fig. 2).

New Potential Behavioral Signs of Consciousness in DoC Patients
Resistance to Eye Opening Resistance to eye opening, a firm closure of already closed eyelids when an examiner touches or tries to open the eyes, is evident in multiple neurological disorders. [31][32][33][34] The presence of resistance to eye opening and its correlation with different levels of consciousness were assessed in 79 prolonged DoC patients (TBI and non-TBI). 35 The diagnosis of patients was based on repeated CRS-R assessments. The examiners considered resistance to eye opening present when there was forceful closure of one or both eyes upon manually opening the patients' upper eyelids bilaterally. Resistance to eye opening was present in 24% of patients (19/79): 26% of VS/UWS (6/23), 53% of MCSÀ (8/15), and 12% of MCSþ (5/41). Although MCSþ patients had the lowest rate of resistance to eye opening, a statistically significant relationship was present between resistance to eye opening and the level of consciousness. In addition, the repeatability of resistance to eye opening was the highest in patients with MCS þ , suggestive of a correlation between the level of consciousness and the number of times resistance to eye opening was seen. MCSþ patients having the lowest rate but the highest repeatability seem contradictory. One possible explanation might be that as patients recover their consciousness, they might be able to understand the instructions of the examiner and inhibit their resistance to eye opening. To replicate, validate, and better understand the relationship between resistance to eye opening and the level of consciousness, future studies could include eMCS patients and healthy controls.
Furthermore, atypical neuroimaging findings (brain activity consistent with MCS diagnosis) were more likely to be seen in VS/UWS patients with resistance to eye opening (83%) than without (29%). Indeed, five out of six patients diagnosed with VS/UWS and with resistance to eye opening had neuroimaging results more compatible with MCS. Four showed relatively preserved PET metabolism in the frontoparietal network (similar to MCS patients), and one showed response to command during a motor imagery task assessed with fMRI, suggesting that these patients were MCS Ã / CMD. 14,16 After 6 months of follow-up, only one of these patients showing resistance to eye opening recovered from VS/UWS, two passed away, and three remained in VS/UWS. Thus, there was no correlation between this behavior and the prognosis of these patients. Collectively, these results suggest that assessing resistance to eye opening repeatedly in prolonged DoC patients can help clinicians gain insight into patients' levels of consciousness. However, this study population had heterogeneity of etiologies and locations of brain injury. Since there might be voluntary and reflexive presentations of resistance to eye opening, future studies localizing brain lesions and correlating this with resistance to eye opening might provide more information regarding the cortical mediation of this behavior.

Spontaneous Eye Blink Rate
Several research teams have shown that eye blink rate is modulated by fatigue, vigilance, task demand, and cognitive load. [36][37][38] To test whether there was a difference in the spontaneous eye blink rate between MCS and VS/UWS patients, 24 chronic DoC patients (TBI and non-TBI) were enrolled in a recent study. 39 Ten patients were diagnosed as VS/UWS and 14 as MCS according to the CRS-R. There were two experimental sessions for each patient, at least 24 hours apart, where patients' eye blink rate was observed at rest for 3 minutes. The examiners stood next to the patients' bed, out of patients' visual field, where they could observe and count the eye blinks. All patients were encouraged to stay relaxed with their eyes open, and not move (they were not informed about the eye blink counting to avoid potential bias). Spontaneous eye blink rate at rest (the final agreement on the rate was reached after also taking into account EEG and EOG recordings) was found to be significantly higher in MCS compared with VS/UWS patients (first session: mean of 8 AE 3 blinks for UWS and 18 AE 3 for MCS; second session: mean of 6 AE 2 blinks for UWS and 26 AE 4 for MCS patients). CRS-R index (a modified linear score taking into account the highest item in each subscale) 40 was calculated and found to significantly correlate with the mean eye blink rate at rest. Due to small sample size and fluctuations in the arousal of DoC patients in this study, more studies with larger sample sizes and more standardized assessment protocols (blinding  examiners to patients' diagnosis and applying consistent timing of experimental sessions across patients) are recommended to further test the spontaneous eye blink rate as a potential indicator for the level of consciousness.

Auditory Localization
In the CRS-R, spatial localization in visual and motor domains (visual pursuit and localization to noxious stimulation, respectively) is considered signs of MCS, whereas in the auditory subscale this is not the case, and localization to auditory stimulus is considered a reflex. In the CRS-R, auditory localization is evaluated by presenting auditory stimuli (patients' name, voice, noise, etc.) behind the patient, out of view, for 5 seconds twice on each side (right and left). When there is a clear head or eye movement toward the auditory stimuli on both trials in at least one direction within 10 seconds of stimulus presentation, the patient is considered to have auditory localization. 25 In a multimodal study, 186 patients with prolonged DoC (TBI and non-TBI) were assessed to examine whether auditory localization could be considered as a sign of MCS. 41 The probability of auditory localization increased with the level of consciousness: 13% of VS/UWS, 46% of MCS À , 62% of MCS þ , and 78% of eMCS patients had auditory localization (►Fig. 3A). Notably, regardless of the diagnosis, patients with auditory localization had higher survival rates after 2 years of follow-up (despite no significant differences in clinical improvement). According to the results obtained with PET, there were no significant differences in brain metabolism between VS/UWS patients with and without auditory localization. However, fMRI analysis showed higher functional connectivity between frontoparietal network and secondary occipital regions during rest in VS/UWS patients with localization compared with those without localization. High-density EEG results showed that VS/UWS patients with localization also had a higher participation coefficient in the α-band compared with VS/UWS patients without localization. The participant coefficient is a connectivity measure that has been shown to correlate with the level of consciousness in previous studies on DoC patients. [42][43][44] Taking all these results into consideration, auditory localization might be a more complex behavior than a reflex, and with the need for additional confirmation of further studies, it could be reconsidered as a potential sign of MCS.

Habituation of Auditory Startle Reflex
In the auditory subscale of the CRS-R, auditory startle reflex is the lowest score item above zero (no response). To test its validity as a sign of MCS, habituation of auditory startle reflex was examined in 98 patients with prolonged DoC (TBI and non-TBI). 45 Habituation was assessed by presenting a loud handclap noise directly above the patients' head (out of view) 10 times consecutively ($120 bpm), administering four trials. If patients had eyelid flutter or blink immediately after the stimulus in at least two trials, this was considered as auditory startle being present. If patients had eyelid flutter or blink after each and every clap, this was considered as inextinguishable auditory startle reflex; otherwise, patients     [65][66][67][68][69][70][71][72][73][74][75][76][77][78][79][80][81][82][83]). The mechanism of habituation of auditory startle reflex could be of cortical origin, as shown with FDG-PET and high-density EEG data. PET metabolic activity in multiple networks including the salience network and the default mode network correlated with the presence of habituation of auditory startle reflex (►Fig. 3B). Higher uand α-power, together with higher values of cortico-cortical functional connectivity (especially, higher prefrontal-temporal connectivity), were observed in patients with habituation compared with patients without habituation. The recovery of command-following at 6 months was significantly higher in VS/UWS patients who showed habituation compared with those who did not show habituation. This new behavioral item could be considered as another sign of consciousness for MCS diagnosis and could be implemented in the CRS-R auditory scale.

Olfactory Sniffing
The sense of smell in DoC patients has been scarcely studied; yet, accumulating findings suggest a link between olfactory abilities and the level of consciousness. A pilot study in 11 DoC patients (nine VS/UWS and two MCS) examined olfactory discrimination abilities based on patients' behavior. 46 A discriminatory olfactory response was defined by a behavioral response (eyes closure, grimace, avoiding head movement, or vocalization) to an unpleasant odor (stuffy sockslike or rancid-like) and trigeminal-irritating odor (ammonia) but not to a pleasant odor (rose-like). All MCS patients and three VS/UWS patients (33.3%) showed a discriminatory olfactory response. Notably, the three patients diagnosed with VS/UWS who showed a discriminatory olfactory response had olfactory-related activity in olfactory cortices when assessed with fMRI, suggesting that these patients might be MCS Ã . The six VS/UWS patients (66.7%) with no discriminatory olfactory response had no olfactory-related activity in either the primary or secondary olfactory cortices. These findings align with an fMRI study that examined olfactory neural processing in 26 VS/UWS and 7 MCS patients. 47 Odor-induced activity in primary olfactory areas was evident in 58% of VS/UWS patients (15/26) and all MCS patients (100%, 7/7). Brain activation also varied with the etiology of the lesion, where most patients with anoxic brain injury had no activation in primary olfactory areas.
A recent study investigated olfactory sniffing in 43 patients with chronic DoC (TBI and non-TBI). 48 Sensorydriven sniffing (level 1, odorant detection; level 2, odorant discrimination) reflected automatic odorant-driven response, whereas cognitive-driven sniffing reflected situational understanding and/or learning (when subjects were told they will be presented with an odorant, but they were instead presented with an empty jar; nevertheless they modified their nasal inflow). Patients were presented with pleasant (shampoo) and unpleasant (rotten fish) odors, or clean air (empty jar); their nasal inhalation volume in response to the stimuli was examined via a nasal cannula directly connected to a spirometer, and an instrumentation amplifier. Patients' level of consciousness was assessed with CRS-R and/or Coma/Near Coma Scale after each session (in addition, the Loewenstein Communication Scale was used in some cases). In total, 73 sessions were conducted in 31 MCS patients, and 73 sessions were conducted in 24 VS/UWS patients. At the group level analysis (for sessions), they observed a 10% reduction of nasal airflow from baseline in response to the odorants' presentation (regardless of pleasant or unpleasant, indicating sensory-driven level 1 sniffing response) in MCS sessions, but not in VS/UWS sessions. A similar difference was also recorded for cognitive-driven sniffing, with a 5% nasal airflow reduction in response to clean air presentation (empty jar) compared with baseline among MCS patients only. When reanalyzing the data to reflect individual level differences (of clinical importance at the patient level), sniff response had a sensitivity of 64.5% to determine MCS. Surprisingly, 10 out of 24 VS/UWS patients showed sniff response in at least one session, and all 10 patients later transitioned into MCS (during the study, 16 out of 24 VS/UWS patients transitioned to MCS). In this sample, the sniff response in VS/UWS patients therefore demonstrated 100% specificity and 62.5% sensitivity (10 out of 16 VS/UWS patients who transitioned to MCS) to predict a transition to MCS. All the patients were followed up to see how their sniff response related to long-term outcome (►Fig. 3C). A sensitivity of 91.7% was measured for the sniff response in predicting survival at 3.1 AE 1.2 years after brain injury.
A more recent study evaluated DoC patients' behavioral responses to different olfactory stimuli, with a more qualitative approach compared with the aforementioned olfactory sniffing studies. 49 Twenty-three DoC patients (TBI and non-TBI) were enrolled in this study. Eight patients were diagnosed as VS/UWS and 15 as MCS according to repeated CRS-R. Videos were recorded while patients were being presented with one of three different olfactory stimuli: 1-Octen-3-ol (familiar neutral odor), pyridine (unpleasant fishlike smell), and water (odorless). Each odor was presented once for 3 seconds, with 15 seconds between different stimuli. The behavioral responses such as pouting, shaking head, pushing things away with hands, frowning, and twisting head in avoidance were scored by two independent and blinded (to the stimuli and diagnosis) raters. Among all patients, the behavioral responses to olfactory stimuli (1-Octen-3-ol and pyridine) were higher than nonolfactory (water) stimuli. During the familiar neutral odor session, 93% of MCS and none of the VS/UWS patients showed behavioral responses, and this difference was significant. During unpleasant fish-like smell session, 60% of MCS and 13% of VS/UWS patients showed behavioral responses, although this difference was not significant. During the odorless session, 13% of MCS and none of the VS/UWS patients showed behavioral responses. The patients were followed up after 1, 3, and 6 months with CRS-R evaluations. There was no significant correlation between behavioral response to olfactory stimuli and the prognosis.
Altogether, these findings suggest that olfactory stimuli are valuable additions to the current assessment protocols and that olfactory sniff response is a powerful and easily accessible tool which can be used in the assessment, diagnosis, and prognosis of DoC patients, further decreasing misdiagnosis rates.

Swallowing/Oral Feeding
Previous neuropathological studies suggest that a correlation may exist between the level of consciousness and swallowing function. [50][51][52] The presence of oral feeding was investigated in 68 chronic VS/UWS patients (TBI and non-TBI) by reviewing their clinical information. These patients also underwent multimodal assessments. 53 Only 3% (2/68) of these VS/UWS patients could be fed orally. The first patient received liquid and semi liquid oral feeding in combination with gastrostomy feeding. Otorhinolaryngological exam and fiberoptic endoscopic evaluation demonstrated intact laryngeal mobility and cough reflex, and no salivary or secretions stasis. Even though no inhalation occurred, the initiation of the swallowing reflex was delayed. Clinical evaluation and neuroimaging assessments were suggestive of VS/UWS diagnosis. The second patient received full oral feeding, with solid food. Behavioral evaluations were suggestive of VS/UWS diagnosis as well, but neuroimaging and electrophysiologic assessments showed atypical findings (relative preservation of metabolism within frontal and occipital cortices, relatively preserved white matter integrity on diffusion tensor imaging, and theta activity on EEG, despite the absence of restingstate fMRI networks). Due to dissociation between behavioral and neuroimaging findings, this patient could be considered as MCS Ã rather than VS/UWS. This study suggests that full oral feeding and a complex oral phase of swallowing might be considered as a sign of consciousness.
A more recent study collected information regarding respiratory status, nutritional status, and otolaryngological swallowing examination from 92 patients with prolonged DoC (TBI and non-TBI). 54 Ten criteria were established: respiratory status (tracheostomy), nutritional status (feeding type), oral phase of swallowing (hypertonia of the jaw muscles, oral phase, efficacy of the oral phase), and pharyngeal phase of swallowing (pharyngo-laryngeal secretions, saliva aspiration, cough reflex, cream aspiration, liquid aspiration). The presence of a tracheostomy, cough reflex, and oral phase efficacy were found to be related to consciousness. None of the VS/UWS patients (diagnosed with multiple CRS-Rs and confirmed with hypometabolism in the frontoparietal network bilaterally using PET) had an efficient oral phase, and none could be fed orally. Additionally, none of the MCS patients received ordinary oral food. Since VS/UWS patients more frequently had a tracheostomy at the time of assessment than MCS patients, their ability to correctly manage saliva differed significantly. Taken together, these results suggested that objective and systematic assessment of swallowing should also be performed in all DoC patients, which could provide additional clinical data on the level of consciousness.

Facial Expressions to Nociception
Pain assessment and management in DoC patients have long been an important ethical issue, since these patients cannot communicate their needs explicitly. Whether the level of responsiveness to painful stimuli could reflect the level of consciousness was investigated in a study enrolling 85 acute and prolonged DoC patients (TBI and non-TBI). 55 The levels of consciousness assessed by CRS-R total scores correlated with responses to the Nociception Coma Scale-Revised (NCS-R). Specifically, MCS patients had higher NCS-R scores compared with VS/UWS patients. CRS-R oromotor/verbal and motor subscores after noxious stimulation correlated with total NCS-R scores during noxious stimulation, and the NCS-R was not found to be more sensitive than CRS-R in assessing nociception. However, the importance of observing facial expressions to nociception was emphasized, with the results showing that grimace was observed more frequently in all patients during painful stimulation compared with nonpainful stimulation. Furthermore, there was a difference in grimacing frequency between MCS and VS/UWS patients during noxious stimulation, less frequent in the latter group. This difference may be due to the presence of tracheostomy (more frequently present in VS/UWS patients) having a possible effect on decreased lower face expression of patients. Nonetheless, observation of the facial expressions in the DoC patient population could add valuable information to the assessment.
Another study including 147 brain-injured patients (TBI and non-TBI) assessed behaviors related to standard ICU care procedures (nociceptive and non-nociceptive) in patients with different levels of consciousness. 56 Patients were classified as unconscious (GCS: 3-8), altered (GCS: 9-12), or conscious (GCS: [13][14][15]  by the CRS-R. MBT-r includes seven positive signs and two negative signs (►Table 2). Patients were considered to have residual cognition if at least one positive item was present. The presence of a negative item suggested brainstem dysfunction and potentially abnormal automatic responses, in which case patients were not scored with MBT-r. The tool also took into consideration the inter-rater agreement of each item, eliminating reliance on an isolated item with a low inter-rater agreement. MBT-r was administered to a cohort of 30 patients with acute DoC (TBI and non-TBI) as a complementary tool to the CRS-R. 58 The authors followed up patients at discharge, after 3 months, and after 6 months, with the Glasgow Outcome Scale score, grouping them into favorable and unfavorable outcome according to their consciousness recovery (unfavorable: remaining in VS/UWS or death). Out of 24 patients classified as unconscious (coma, VS/UWS) by the CRS-R (best score out of three assessments), 18 (75%) showed signs of residual cognition with MBT-r. Also, 66.7% of patients showing residual cognition by the MBT-r had a favorable outcome. Thus, MBT-r could be a useful clinical tool to detect signs of residual cognition (subtle motor behavior) underestimated by the CRS-R, and predict recovery in acute DoC patients.
This tool was also used in a study of 140 patients (TBI and non-TBI), where the patients were grouped into DoC (coma, VS/UWS, and MCS), non-DoC (patients who were able to interact adequately), and potential clinical CMD (patients who have residual cognition according to MBT-r assessment). 59 The latter group showed a strong improvement trajectory of functional/cognitive recovery from admission to discharge, where outcomes were measured by GOS and other outcome scales (e.g., Disability Rating Scale). 59 Collectively, these results emphasize that the combination of MBTr and CRS-R in DoC patients could help detect covert consciousness in a substantial fraction of patients.

Leg Crossing
Crossing legs is considered an automatic motor response in the CRS-R, one of the eleven signs that denotes MCS À . In one study, 34 patients with severe stroke who crossed their legs during their hospitalization ("crossers") were matched with 34 severe stroke patients who did not cross their legs ("noncrossers"). 60 Patients were evaluated at admission, upon discharge, and 1 year after discharge, with GCS, NIH Stroke Scale (NIHSS), modified Rankin scale (mRS), and Barthel Index (BI). No significant differences were observed between the two groups at the time of admission, but upon discharge NIHSS and mRS were lower and BI was higher for crossers, indicating less severe neurologic deficits, less disability, and higher functional independence, respectively. After 1-year follow-up, these differences were even larger. Also, mortality was significantly lower in the "crossers" group. Leg crossing within the first 15 days after severe stroke favored better outcome in patients, and could be used as a prognostic tool. It is a sign that any healthcare provider could easily assess and needs further attention and validation by larger studies. This behavior has not been assessed in patients with DoC, but further attention is warranted in clinical practice, and further studies should assess the validity of this sign in patients with DoC.

Discussion
In this review, we summarized recent findings regarding newly proposed behaviors denoting consciousness in patients with DoC, which could help improve the accuracy of detecting and predicting recovery of consciousness. A summary of the findings is presented in ►Table 1. While they may not all reliably reflect the presence of conscious processing, the use of these behaviors can be justified based on their safe and affordable evaluation. We therefore advocate a careful observation of these new behavioral signs (resistance to eye opening, spontaneous eye blink rate, auditory localization, habituation to auditory startle reflex, olfactory sniffing, swallowing/oral feeding, facial expressions to noxious stimulation, subtle motor behavior, and leg crossing) among DoC patients when clinically appropriate (►Fig. 2). Raising awareness about these behaviors among caregivers, families, and all the responsible healthcare personnel might drive the development of validation studies and encourage a thorough multimodal assessment of patients presenting these clinical signs.
Beyond repeated assessments with CRS-R, we encourage the use of additional standardized tools in patients with DoC to test specific functions, such as the MBT-r or the SWADOC (SWallowing Assessment in Disorders of Consciousness-a standardized swallowing tool under validation 61 ). However, we acknowledge the challenges of time management when assessing patients in ICU settings. In this context, a new validated scale (SECONDs) has been developed to provide a faster and more practical tool to administer than the CRS-R in time-constrained clinical settings. 23,24 We also encourage continuous observation and reporting by healthcare personnel of other spontaneous motor behaviors such as tube pulling, nose scratching, and grabbing sheets. As more potential new behaviors of consciousness emerge from these observations, we recommend evaluating these behaviors in future studies in three ways: by comparing these behaviors between different states of consciousness (VS/UWS, MCS, MCS Ã , CMD, eMCS), by correlating each behavior with longterm outcome measures, and finally supporting these studies with neuroimaging techniques (►Fig. 3).
Despite established criteria for the diagnosis of MCS in standardized behavioral assessments, some of the items in the CRS-R denoting MCS are still controversial, such as visual fixation. In a study comparing the cerebral metabolism of five patients with chronic anoxic VS/UWS (therefore without visual fixation) with five patients with chronic anoxic MCS in whom the only sign of consciousness was visual fixation, no significant difference was found in cortical metabolism or cortico-cortical connectivity between the two groups. 62 Although this study used a small sample size limited to anoxic brain injury, it stresses the need to further validate some of the controversial signs of consciousness, such as visual fixation, with multimodal neuroimaging studies and larger sample sizes to elucidate the neural mechanisms supporting behavior and relate it to conscious processing.
Recent studies, beyond the scope of this article, suggest that electrophysiological findings, 63-73 neuroimaging findings, 16,43,74 pupil responses, 75,76 and autonomic nervous system correlates 77,78 might also reflect the level of consciousness in patients with DoC. Thus, we emphasize the need to further investigate and validate these parameters within the framework of a multimodal assessment (behavioral and neuroimaging) in DoC patients. There have been studies also suggesting that using salient stimuli, assessing behaviors within a context relevant to the patients, or performing individually tailored motivational assessments might further enhance arousal and patient participation when diagnosing patients with DoC. 59,79-84 Thus, we encourage further validation studies about the effects of implementing emotional context and salient stimuli in the assessment of DoC patients.
Regarding the diagnosis of DoC patients, some of the recommendations by the 2020 European Academy of Neurology guideline 27 are as follows: (1) passively opening patients' eyes who have no spontaneous or stimulationtriggered eye opening, and assess for both horizontal and vertical eye movements (patients with locked-in syndrome have preserved vertical eye movements) (strong recommendation); (2) using a mirror for visual pursuit, and if not elicited by a mirror, the use of pictures showing the patient's or relatives' faces or personal objects (strong recommendation); and (3) using repeated CRS-R (at least five times) assessments in the subacute-chronic setting and the Full Outline of UnResponsiveness scale in the acute setting instead of the GCS (strong recommendation). Also, the need for multicenter collaborations is highly stressed in this guideline, as well as the need for more studies investigating resistance to eye opening, 35 pupillary dilation assessment following mental arithmetic with automated pupillometry, 75,85 quantitative assessment of visual track-ing, 86,87 standardized rating of spontaneous motor behavior, 58 the possibility of oral feeding, 53 evidence of circadian rhythms, 88 vegetative responses to salient stimuli, 89 and modulations of cardiac cycle (heart rate, heart rate variability, cardiac cycle phase shifts). 69,70 The diagnostic assessment of DoC patients is associated with major clinical and ethical implications. Feasible solutions to improve diagnostic accuracy are urgently needed in clinical practice, which should be addressed in a multidisciplinary approach. Further formal validation studies for the proposed new behavioral signs of consciousness summarized in this article are needed for implementation in clinical practice, although we acknowledge the challenge of lacking a gold standard reference for consciousness. Currently, very few specialized centers perform an up-to-date multimodal assessment of these patients. As a result, there is a distinct gap between scientific research and clinical practice. We encourage ICU and rehabilitation healthcare workers to collaborate with translational research teams worldwide and adopt a multidisciplinary approach in their assessment of patients with DoC, which may help bridge the gap between research and clinical practice in this field.

Funding
The study was supported by the University and University Hospital of Liège, the 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 European Space Agency (ESA) and the Belgian Federal Science Policy Office (BELSPO) in the framework of the PRODEX Programme, the BIAL Foundation, the Mind Science Foundation, the fund Generet of the King Baudouin Foundation, the Léon Fredericq Foundation, the Mind-Care foundation, the AstraZeneca Foundation, the Télévie and the Fondation Contre le Cancer, and a Marie Curie Individual Fellowship (840711) awarded to A.A. L.R.D.S. is research fellow, O.G. is research associate, and S.L. is research director at FRS-FNRS.

Conflict of Interest
None declared.