Table 1. Summary of results from studies reviewed in this article.

Author (year)	Behavior	No. of patients and sex	Age (mean±SD or median, IQR or range)	Onset (mean±SD or median, IQR or range)	Etiology	Behavioral tool	Diagnosis	Brain imaging	Main findings	Sign of C	Prognosis	Bias/Limitations
van Ommen et al 35 (2018)	Resistance to eye opening	79 36 F 43 M	37, IQR: 28–52	15 mo, IQR: 1–41 mo	42 TBI 25 post-anoxic encephalopathy due to cardiac arrest 6 stroke 5 SAH 2 encephalitis 1 hypoglycemia	CRS-R	23 VS/UWS 15 MCS− 41 MCS+	PET MRI	Significant relationship between REO and the level of consciousness. Higher REO repeatability in MCS. Atypical neuroimaging findings similar to MCS in VS/UWS patients with REO.	Yes	No correlation	Heterogeneity in time since onset, etiologies, and brain damage.
Magliacano et al 39 (2021)	Spontaneous eye blink rate	24 6 F 18 M	50 ± 18 VS/UWS 53 ± 18 MCS	6 ± 16 mo	7 TBI 9 ABI 8 vascular	CRS-R	10 VS/UWS 14 MCS	EEG	Eye blink rate was significantly higher in MCS patients compared with VS/UWS patients. A significant positive correlation was found between CRS-R index and eye blink rate.	Yes	n/a	Researcher who evaluated eye blink rates was not blinded to patients' diagnosis. Only investigating eye blink rate, but not other blink characteristics. Not recording all sessions for all patients at the same time of the day. Some patients were assessed less than 5 times by CRS-R for diagnosis. Small sample size.
Carrière et al 41 (2020)	Auditory localization	186 66 F 120 M	39 ± 16	9 mo, range: 1mo–29y	100 TBI 86 non-TBI	CRS-R	64 VS/UWS 28 MCS− 71 MCS+	PET fMRI hdEEG	Auditory localization increased with level of consciousness. Higher survival rates after 2-y follow-up in patients with auditory localization compared with patients without auditory localization. Higher fMRI functional connectivity between frontoparietal network and secondary occipital regions in VS/UWS patients with auditory localization. Higher participant coefficient in α band in VS/UWS patients with auditory localization.	Yes	Increased survival rate	Small sample size of VS/UWS with localization (auditory localization as the only “sign of consciousness” is rare). Subgroups not matched for age and time since injury. Missing clinical outcome data. Some patients had light sedation for fMRI. Lack of auditory evoked potentials or otoacoustic emissions to rule out deafness in the absence of auditory response.
Hermann et al 45 (2020)	Habituation to auditory startle reflex	96 M/F ratio: 1.8	44 ± 16	58 d, IQR: 31–236d	39 ABI 27 TBI 12 vascular 18 other	CRS-R	48 VS/UWS 48 MCS	PET DTI MRI hdEEG	More habituation in MCS compared with VS/UWS. Higher CRS-R scores in all subscales except communication in patients with habituation. Highest prevalence and sensitivity for habituation compared with the performance of all MCS items of the CRS-R to discriminate MCS. PET activity in salience and default mode networks correlated with habituation. Higher θ and α power, with higher prefrontal-temporal connectivity in patients with habituation. Higher recovery of command-following after 6-mo follow-up in VS/UWS with habituation.	Yes	Yes (recovery of command-following after six-mo follow up)	None reported.
Sattin et al 46 (2019)	Olfactory discrimination	11 6 F 5 M	57, IQR: 14	3–146 mo	6 ABI 3 hemorrhagic brain injury 1 ischemic and hemorrhagic brain injury 1 TBI + ABI	CRS-R	9 VS/UWS 2 MCS	fMRI	All MCS and 33% of VS/UWS had a discriminatory olfactory response. All VS/UWS with discriminatory olfactory response had olfactory-related activity in olfactory cortices in fMRI.	Yes	n/a	Small sample size. Lack of quantitative analysis of nasal airflow. Using only 4 odors could underestimate the real olfactory functions due to chance of specific anosmia. 4 patients excluded due to head movements in fMRI.
Nigri et al 47 (2016)	Olfactory processing	42 19 F 23 M *	57, IQR: 23–77 for VS/UWS; 44, IQR: 20–71 for MCS	26mo, IQR: 3–146 for VS/UWS; 41mo, IQR: 11–170 for MCS	17 ABI 7 hemorrhagic brain injury 6 TBI 2 ischemic and hemorrhagic brain injury 1 TBI + ABI	CRS-R	26 VS/UWS 7 MCS	fMRI	58% of VS/UWS, 100% of MCS showed odor-induced activity in primary olfactory areas. 39% of VS/UWS and 71% of MCS showed activation within a higher-order olfactory processing area. Most patients with anoxic brain injury had no activation in primary olfactory areas.	n/a	n/a	Small sample size. The excluded populations (due to movement in fMRI) were majorly MCS, and thus might have skewed the results.
Arzi et al 48 (2020)	Olfactory sniffing	43 8 F 35 M	43 ± 17	1mo–10mo	27 TBI 5 ABI 10 cerebrovascular accident 1 infection	CRS-R CNC	21 VS/UWS 22 MCS	n/a	Reduction of nasal airflow in response to odorants and empty jar presentation in MCS sessions, but not in VS/UWS sessions. Sniff response had 64.5% sensitivity to determine MCS. On individual level, VS/UWS patients who showed sniff response in at least one session later transitioned to MCS. Sniff response had 100% specificity and 62.5% sensitivity in predicting transition from VS/UWS to MCS. Sniff response had sensitivity of 91.7% in predicting survival after >3 y.	Yes	Recovery of consciousness (transition to MCS) and predicting survival after >3 y	Data using two different behavioral assessment protocols (this was alleviated by having a third tool applied equally to all participants). Not being able to test more frequently due to clinical schedule (possible observation of advance detection with more frequent testing). This method requires uninflated tracheostomy balloons and would not work with inflated tracheostomy balloons.
Wang et al 49 (2022)	Behavioral response to olfactory stimuli	23 7 F 16 M	22–69	1–11 mo	10 TBI 13 non-TBI	CRS-R	8 VS/UWS 15 MCS	n/a	Behavioral response to odorant stimuli compared with nonodorant stimuli (water) was higher among all patients. In response to the neutral odor presentation (1-Octen-3-ol), MCS patients had higher behavioral response compared with VS/UWS patients.	Yes	No correlation	No neuroimaging. Small sample size.
Mélotte et al 54 (2020)	Oral feeding	92 39 F 53 M	41 ± 12 VS/UWS 38 ± 12 MCS	30mo±22 for VS/UWS; 4mo±34 for MCS	60 focal 32 global	CRS-R PET	26 VS/UWS 66 MCS	PET	Presence of tracheostomy, cough reflex, and oral phase efficacy related to consciousness. 0 VS/UWS with oral feeding or efficient oral phase. 0 MCS with complete feeding.	Yes	n/a	Missing data for cough reflex criterion. Limited number of available criteria due to retrospective nature of the study.
Chatelle et al 55 (2018)	Facial expression to noxious stimuli	85 28 F 57 M	48 ± 17 VS/UWS 43 ± 17 MCS	142d, IQR: 88–396 for VS/UWS; 133d, IQR: 78–350 for MCS	35 TBI 25 ABI 25 other	CRS-R	28 VS/UWS 57 MCS	n/a	MCS had higher NCS-R scores compared with VS/UWS. Grimace observed more frequently in painful stimulation compared with nonpainful stimulation in all patients. Grimacing frequency more frequent in MCS compared with VS/UWS.	Yes	n/a	High overlap of the items in NCS-R and CRS-R. In group analyses, single CRS-R assessment was used for diagnosis. Lack of blinding, same raters carrying out both assessments.
Gélinas et al 56 (2019)	Behavioral response to noxious stimulation	147 51 F 96 M	56 ± 20	> 4 wk after brain injury	94 TBI 33 aneurysm 13 stroke 1 brain abscess	GCS	26 not conscious 56 altered 65 conscious	n/a	Higher number of active behaviors during nociceptive procedures in conscious patients. Grimace was a strong indicator for pain intensity in conscious patients.	Yes	n/a	Raters not blinded. Checklist ratings included both bedside observation and videos which may have led to differences. Only 35 patients were able to self-report their pain (limiting power of analyses).
Pincherle et al 58 (2019)	Subtle motor behavior	30 13 F 17 M	64 ± 16	10±5d	16 hemorrhage 3 metabolic 7 trauma 1 stroke 3 ABI	CRS-R MBT-r	13 coma 10 VS/UWS 5 MCS 2 eMCS	n/a	75% of coma and VS/UWS diagnosed by CRS-R showed signs of residual cognition with MBT-r. 66.7% of patients showing residual cognition by MBT-r had favorable outcome.	Yes	Yes, favorable outcome (discharge, 3 and 6 mo)	Small and heterogenous cohort.
Jöhr et al 59 (2020)	Subtle motor behavior	141 54 F 87 M	53 ± 17	35±107d for clinical CMD; 55±19d for DoC; 25±20d for non-DoC	55 severe traumatic 63 vascular 12 ABI 7 encephalopathy 4 neoplasm	CRS-R MBT-r	105 clinical CMD 19 DoC 17 non-DoC	n/a	Strong improvement trajectory of functional/cognitive recovery from admission to discharge in patients with residual cognition based on MBT-r assessment.	Yes	Yes, functional, and cognitive recovery from admission to discharge.	Categorizing patients into clinical CMD solely based on MBT-r, and not performing active mental-imagery tasks stated in the definition of CMD. No differentiation of subtypes of CMD. Potential measurement error and reliability issues due to nonstandardized assessment of outcome measures retrospectively. Medical complications and their possible impact on outcomes not taken into consideration. Data collection only at two time points, which limited information about recovery course during hospitalization.

Abbreviations: ABI, anoxic brain injury; CNC, Coma/Near Coma Scale; CMD, cognitive motor dissociation; CRS-R, Coma Recovery Scale-Revised; DoC, disorders of consciousness; DTI, diffusion tensor imaging; eMCS, emerging from minimally conscious state; F, female; fMRI, functional magnetic resonance imaging; hdEEG, high-density electroencephalography; IQR, interquartile range; M, male; MBT-r, Motor Behavioral Tool-revised; MCS, minimally conscious state; n/a, not available; NCR-R, Nociception Coma Scale-Revised; REO, resistance to eye opening; SAH, subarachnoid hemorrhage; SD, standard deviation; TBI, traumatic brain injury; VS/UWS, vegetative state/unresponsive wakefulness syndrome.

^*9 patients were discarded in data analysis due to excessive movements. The columns of etiology, diagnosis and main findings take into account only analyzed patients.