Recovery from severe brain injury often follows a trajectory¹ that includes: (1) emergence from coma into a vegetative state (VS;² also known as unresponsive wakefulness syndrome [UWS]³), followed by (2) a minimally conscious state minus (MCS − ), (3) MCS plus (MCS + ), and finally, (4) emergence from MCS (eMCS). VS/UWS is characterized by spontaneous eye-opening and no signs of conscious awareness. Patients in MCS demonstrate inconsistent but convincing evidence of purposeful behaviors, either without (MCS − ) or with (MCS + )^4,5 some preservation of language function. Finally, eMCS, which is generally consistent with an acute confusional state,⁶ and has been termed post-traumatic confusional state (PTCS)⁷ in patients with traumatic brain injury (TBI), is marked by functional use of common objects or reliable, basic yes/no communication.⁴ The behaviors distinguishing VS/UWS from MCS and MCS from eMCS were established in 2002,⁴ although revisions to the criteria for eMCS have been recently proposed.^8–10,90

Historically, assessment of consciousness, and thus the diagnostic characterization of disorders of consciousness (DoC), has relied almost exclusively on bedside neurological examinations. However, due to biases introduced by examiner error (e.g., failing to assess a broad range of behaviors, over-/under-interpretation of observations) and patient-related factors (e.g., fluctuating arousal, motor deficits, language deficits, pain, sensory impairments, sedating medications), the bedside clinical examination can be an unreliable marker of level of consciousness. Indeed, the approximate rate of misdiagnosing a patient with at least minimal consciousness (MCS) as unconscious (VS/UWS) on routine clinical examination is 40%.^11–15 This alarming rate of misdiagnosis may affect critical clinical decisions, lead to premature withdrawal of life-sustaining treatment, and restrict access to rehabilitation services.

In the past 3 years, clinical guidelines published in the United States,¹⁶ Europe,¹⁷ and United Kingdom¹⁸ have all strongly recommended the use of specialized, standardized measures for assessment of DoC. Developed through a collaboration of the American Academy of Neurology (AAN), the American Congress of Rehabilitation Medicine (ACRM), and the National Institute on Disability, Independent Living, and Rehabilitation Research (NIDILRR), the 2018 practice guideline recommendations for DoC state that clinicians should use serial, standardized neurobehavioral assessment measures that have been shown to be valid and reliable that, to improve diagnostic accuracy, clinicians should use serial, standerdized.¹⁶ The European Academy of Neurology (EAN) guidelines published in 2020 alsosupport useofa standardized behavioral assessment of DoC,¹⁷ as do the 2020 Royal College of Physicians guidelines.¹⁸ An ACRM and NIDILRR position statement on the minimum competency recommendations for rehabilitation programs that provide services for individuals diagnosed with DoC includes repeated diagnostic assessment with a standardized and validated tool.¹⁹ However, even standardized behavioral evaluation is susceptible to biases and confounders that may affect diagnostic assessment, and examiners should ensure that testing conditions are optimized to promote maximal responsiveness. Table 1 lists strategies that can increase the reliability and validity of clinical examinations of persons with DoC and improve the accuracy of diagnosis.

Use of a standardized assessment of DoC may reduce diagnostic error, improve the accuracy of prognostication, and enable monitoring of recovery and response to interventions. Systematic and standardized evaluation of level of consciousness ensures consistency within and across examiners in how subtle behaviors are assessed and documented. When compared with qualitative bedside examination, the increased reliability associated with the use of standardized tools adds confidence that behavioral observations represent the patient’s true level of function rather than artifacts caused by variations in clinical examination techniques. Standardized assessment also facilitates clear communication across clinical teams and with caregivers. Although multiple tools for assessing DoC are available, not all have equally robust psychometric properties. Among 13 behavioral scales reviewed by ACRM in 2010, 7 were recommended, with some variation in level of confidence, for patients with DoC. However, only the Coma Recovery Scale-Revised (CRS-R)²⁰ was recommended with minor reservations for both clinical and research applications (Table 2).²¹ The National Institutes of Neurological Disorders and Stroke (NINDS) includes the CRS-R as a Common Data Element (CDE), strongly encouraging its use in studies enrolling persons with DoC.²² Importantly, the CRS-R was the “reference-standard” behavioral assessment tool that revealed evidence of consciousness (i.e., MCS) in the approximately 40% of patients who were misdiagnosed as VS/UWS in the aforementioned study.¹¹ All three recently published clinical guidelines recommend the CRS-R above or in addition to other behavioral assessments.^16–18

The CRS-R (Table 3) is a 23-item rating scale composed of six subscales that assess behaviors mediated by language, visuoperceptual, and motor networks.⁶ The items are hierarchically arranged reflecting brainstem, subcortical, and cortically mediated functions.^23,24 Importantly, the diagnostic criteria for VS/UWS, MCS − , MCS + , and eMCS are embedded into the CRS-R profile, enabling differentiation of these states. The CRS-R total score ranges from 0 to 23 and, although scores of 10 or greater indicate MCS or eMCS, individual subscale scores provide a more precise diagnostic classification.²⁵ Serial CRS-R assessment has high sensitivity for detecting signs of consciousness,²⁶ and the CRS-R diagnosis, total score, and rate of change may assist with prediction of subsequent functional outcome.^27–29 The CRS-R meets minimum standards for measurement and evaluation tools as well as criteria as an interval scale, which allows for both intra-individual and between-patient comparisons.²³

The CRS-R assessment manual, recommended training, and ancillary materials were updated in 2020 and are available online at: https://www.sralab.org/rehabilitation-measures/coma-recovery-scale-revised.

Extensions and adaptations of the CRS-R have expanded its use globally and to special populations and clinical questions. The CRS-R has been translated and validated in Spanish,³⁰ Italian,^31–33 French,³⁴ Portuguese,³⁵ Norwegian,³⁶ Russian,³⁷ German,³⁸ Polish,³⁹ Korean,⁴⁰ and Chinese^41,42 with validated versions forthcoming in Japanese, Hebrew and Brazilian Portuguese. The CRS-R is also available in Dutch, Swedish, Danish, and Greek, but has not been validated in these languages.⁴³ A pediatric version, the Coma Recovery Scale for Pediatrics, was validated in healthy subjects to determine which behaviors should be assessed in children with DoC who are at different developmental milestones.⁴⁴ The Motor Behavior Tool, which was developed to complement the CRS-R, identifies subtle motor behaviors that may identify residual cognition,⁴⁵ signaling the potential for recovery in persons with DoC. The Nociception Coma Scale and its revised version were specifically developed to assess pain perception in patients with VS/UWS and MCS.⁴⁶ Additional scales that complement the CRS-R but focus specifically on language assessment⁴⁷ and swallowing⁴⁸ have also been recently developed. Alternate CRS-R scoring criteria have been proposed to improve the utility of the total score for differentiating VS/UWS from MCS,^49,50 and a new measure, the Brain Injury Functional Outcome Measure,⁵¹ which expands the floor and ceiling of the Functional Independence Measure instrument (FIM),⁵² incorporates multiple CRS-R subscales.

Individuals who transition to eMCS typically demonstrate clinical features consistent with a confusional state or PTCS, including disorientation, impairment in attention and memory, behavioral dysregulation, disturbance in sleep–wake cycles, and fluctuation in the severity of these symptoms.^6,7 Although no currently available behavioral scales assess all symptoms of the confusional state, the Confusion Assessment Protocol (CAP), a composite measure of cognitive assessments and clinical signs, captures most aspects of the recently published PTCS case definition.⁵³ The CAP may therefore be used to fill the assessment gap for patients who are able to respond to basic yes/no questions, indicating eMCS, but are still in a confusional state and unable to participate in a full neuropsychological assessment.

The use of standardized scales for assessment of patients with DoC is most common in the rehabilitation setting and in clinical trials. In the intensive care unit (ICU), the preferred approach to assessment is the Glasgow Coma Scale (GCS)⁵⁴ which, developed in the 1970s, was the first attempt to measure “depth of coma” and track recovery. The GCS is simple to administer, takes only a few minutes, and has been adopted internationally for both diagnostic and prognostic⁵⁵ applications in prehospital, emergency department, and ICU settings. The GCS is a core TBI CDE^22,56 and is used in clinical trials both as a criterion for subject inclusion and as an approach for subject stratification.⁵⁷ Despite its widespread use, the psychometric strength of the GCS is moderate,^58,59 which may be attributed to the lack of standardized procedures for administration and scoring (though new online materials provide some additional guidance; https://www.glasgowcomascale.org ). The GCS was also not designed to provide a DoC diagnosis and therefore multiple behaviors that differentiate between MCS and VS/UWS are not assessed (e.g., visual pursuit, which, along with localization of noxious stimulation, is an early indicator of emerging consciousness⁶⁰). Furthermore, most GCS total scores represent a wide range of potential DoC diagnoses, suggesting that the total score is not an adequate proxy for level of consciosuess.⁶¹

The Full Outline of UnResponsiveness (FOUR) score was developed to address some of the limitations inherent to the GCS and replaces the GCS verbal subscale, which is often untestable due to intubation, with an assessment of respiratory patterns and brainstem reflexes.^62,63 A systematic review published in 2019 found that both the GCS and FOUR score predicted in-hospital mortality and 3-month outcome with similar accuracy, but that the brainstem and respiratory subscales of the FOUR had lower accuracy as compared with the visual and motor subscales.⁶⁴ Like the GCS, the FOUR score does not assess all behaviors that differentiate between MCS and VS/UWS and is not rigorously standardized.^21,65 A comparison between the CRS-R, GCS, and FOUR score is provided in Table 4.

In contrast to the GCS, the CRS-R is standardized, has strong psychometric properties, and differentiates between DoC diagnoses, but the duration of the assessment, approximately 25 to 45 minutes, may be too lengthy for some patients in the ICU setting who can only tolerate a few minutes of sedation-free behavioral assessment. Accurate assessment of level of consciousness is arguably most critical in the ICU because the behavioral exam serves as the primary proxy for brain function and recovery potential, and is therefore a driving factor in decisions around continuing or withdrawing life-sustaining therapies.^66,67 Moreover, access to rehabilitation services hinges on the ability to participate in therapy and on the trajectory of improvement, both derived from the behavioral exam. Abbreviated, standardized, and validated assessments of consciousness that account for common acute care confounding factors are needed to improve early diagnostic and prognostic precision. The SECONDs score was developed as an abbreviated standardized assessment of consciousness⁶⁸ and a rapid version of the CRS-R (CRS-R for Accelerated Standardized Assessment [CRSR-FAST], clinicaltrails.gov NCT03549572) is being developed specifically for use in the ICU.

When questions arise regarding individual behaviors−for example, the volitional nature of low-frequency, simple movements−even standardized measures that rigorously prescribe all aspects of an assessment, including approach to eliciting responses, number of trials, inter-trial interval, and other parameters, may not be sufficient to distinguish purposeful from random, spontaneous behaviors. Individualized Quantitative Behavioral Assessments (IQBAs) are customized protocols based on principles of single-subject research design that assess the cognitive and behavioral capacities of individuals with marked limitations in responsiveness.^69,70 In this technique, focused clinical questions are probed using a personalized behavioral protocol that operationally defines stimuli and response criteria, and statistically analyzes response frequency to determine whether a behavior of interest occurs more often than chance or more often under specific conditions compared with others. For example, if the CRS-R cannot differentiate between random movement and purposeful low-frequency movement to command, an IQBA can be designed to quantify whether the movement is more likely to occur to a command rather than spontaneously. A finger movement observed on two of four trials immediately after the command to “move your fingers” raises the possibility of verbal comprehension. However, if finger movement fails to occur during the next 20 trials administered, the probability that the two initial responses represented evidence of verbal comprehension is diminished. Conversely, a patient may move his or her fingers following command on 10 consecutive trials, but if the same finger movement also precedes the command or persists well after the last command is administered, it is less likely that this response is an indication of verbal comprehension. IQBA addresses these problems by creating a standardized protocol to assess finger movement with different commands presented in random order (e.g., “give me a thumbs up,” “move your toe,” “hold still”), during multiple sessions on different days over time, to establish a statistical likelihood that the movement is under volitional control. This approach has been successfully applied to command-following, visual attention to stimuli of varying salience (e.g., a photograph vs. a white card), accuracy of yes/no responses, and other behaviors.^69,70

Standardized and individualized procedures are complementary as they serve different purposes. Standardized methods provide a broad overview of the integrity of sensory, motor, and cognitive processes. These findings can help establish diagnosis, prognosis and lesion locus, and may inform the optimal approach to treatment. Moreover, standardized measures are necessary for clinical trials that rely on a systematic and harmonized approach to assessment of all participants. Conversely, the flexibility offered by IQBA provides an opportunity to assess case-specific influences on behavior which may contribute to diagnostic inaccuracy and erroneous judgments concerning consciousness. IQBA may also supplement standardized scales in providing additional quantitative evidence of change, such as number of correct behavioral responses over serial sessions.

As outlined in Table 1, standardized assessment of patients with DoC may be influenced by a variety of factors that affect the accuracy of an evaluation. In addition to these factors, there may be overt threats to assessment validity that result in failure to complete a measure or a completed, but invalid measure. For example, if cortical blindness is suspected, an examiner may choose not to administer portions of a measure that require intact basic visual function due to the high likelihood of an inaccurate result. Alternatively, if sedation is being weaned to support a behavioral assessment, it may become apparent that items administered at the beginning of the assessment were confounded by residual effects of the sedation (e.g., if there is gradual increasing responsiveness throughout an assessment independent of exam procedures aimed at increasing arousal). Given the heavy reliance of most behavioral assessments on language function, aphasia would also compromise assessment of consciousness.^71,72 Under these circumstances, it is important to consider using a tool designed for assessing patients with suspected aphasia,⁴⁷ avoid scoring items that may be confounded by language impairment, and to record, in a systematic way, reasons an assessment could not be completed. One approach to documenting test interference is to apply a Test Completion Code (TCC) each time an assessment is planned, to indicate that the test was completed and valid, attempted but not completed, or completed but not valid. TCCs may provide additional information about a patient’s level of function and assure that missing test scores are not interpreted inaccurately as scores at the floor of the assessment. A list of suggested TCCs for the CRS-R is provided in Table 5.

In the absence of a ground-truth index of consciousness, behavioral evaluation remains the recommended approach for establishing level of consciousness after severe brain injury. However, routine clinical testing aimed at detecting preserved or recovered brain function is mired by methodological limitations. When the process for eliciting behaviors and interpreting responses varies among examiners, it is not possible to disambiguate change in level of function from inconsistencies in assessment methodology. Furthermore, there are differences in how observers of the same behaviors interpret and document responses. Standardizing assessment of level of consciousness reduces the variability of the evaluation procedure and the ambiguity in the interpretation and documentation of behavioral responses. When conducted systematically and serially,²⁶ standardized assessments minimize examiner biases and maximize the opportunity for detecting a patient’s highest level of function.

However, even measures such as the CRS-R, which provides precise instructions for item administration and scoring, are susceptible to biases and may miss or misattribute subtle signs of awareness. IQBA is a complementary approach to diagnostic assessment that provides additional insight into the volitional nature of subtle or infrequent behaviors. Studies conducted over the past 15 years have shown that task-based functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) detect covert command-following in some patients who are unable to follow commands at the bedside.^15,73–76 Termed covert consciousness or cognitive-motor dissociation,⁷⁷ the presence of this phenomenon may be a prognostic marker of further recovery.⁷³ Although not reflective of diagnosis, fMRI and EEG responses to passive stimuli (i.e., covert cortical processing)^75,78,79 or the presence of intact resting-state networks^80,81 may also suggest the potential for recovery of consciousness. The AAN-ACRM-NIDILRR guidelines recommend supplementing the behavioral examination with multimodal assessment when the behavioral examination remains ambiguous or when confounding factors may interfere with the ability to follow commands behaviorally.¹⁶ The EAN guidelines recommend task-based, stimulus-based, and resting-state multimodal assessment for all patients with DoC.¹⁸ It is important to note, however, that many patients diagnosed with MCS on standardized assessment do not show evidence of command-following on fMRI^28,74,76 or EEG,^15,74 and thus these approaches are complimentary and not intended to be used in isolation. Efforts to expand implementation of these techniques outside of major academic centers and establish reimbursement procedures to integrate advanced fMRI and EEG into clinical practice are ongoing.⁸²

In summary, repeated, standardized behavioral assessment is the most reliable and valid approach to evaluating patients with DoC. However, examiners should be cognizant of the limitations inherent in behavioral assessment and should consider strategies to mitigate biases and confounders. Developing brief standardized assessment measures to detect conscious awareness in the ICU setting is an active area of research and is needed to improve diagnostic and prognostic precision, as well as to establish a common language among clinicians, with families, and across care settings.