Introduction: The Coma Recovery Scale – Revised (CRS-R) is the reference standard for evaluating patients in disorders of consciousness after both traumatic and non-traumatic brain injury. Previous Rasch analyses with a sample of participants in DoC suggested that the CRS-R items, rating scale categories, and person abilities could be calibrated along an equal-interval 0-100 scale using the Rasch model. However, little research has examined whether the scale can be applied to people with disorders of consciousness stemming from a non-traumatic brain injury (nTBI). A previous study showed the Visual item of the CRS-R is more difficult for people with anoxic brain injuries compared to traumatic brain injuries (TBI); however, the study did not determine which rating scale categories within the Visual item were more challenging. In this study, we conducted a Rasch analysis of the 4576 CRS-R records from 403 participants who had disorders of consciousness from nTBI.

Methods: We applied a four-step analytical approach to generate and validate Rasch item and rating scale category calibrations for the six CRS-R items (Communication, Auditory, Visual, Motor, Oromotor/Verbal, and Arousal) on an equal-interval 0-100 scale. We compared the hierarchy of rating scale categories generated in this analysis with previous anchors generated for a sample of participants with traumatic brain injuries (TBI).

Results: All items fit the Rasch model (mean-square information-weighted fit statistics ranged from .70 – 1.14). A principal components analysis of standardized Rasch residuals provided evidence for unidimensionality (eigenvalue of the largest contrast = 1.72, accounting for 10.5% of variance in the model). The average person measure was 54.27 (SD = 12.79). Fewer than 2.5% of participants had a measure at the ceiling or floor of the instrument. Reliability indices were moderate, with an unadjusted person separation reliability of 0.78; however, the sample-independent Wright’s person separation reliability was 0.94. We found that the three lowest categories on the Visual item (0: No visual function, 1: Visual startle, 2: Visual fixation) were associated with much higher average person abilities for the nTBI participants compared to the previous anchors generated with the TBI participants.

Conclusions: This study provides key insights into applying the CRS-R for individuals with nTBI. Using Rasch analysis, we transformed CRS-R scores into a 0-100 scale and compared rating hierarchies with previous results for TBI. Notably, nTBI participants showed significantly higher average person abilities when they achieved the lower Visual categories, indicating that more basic elements of visual function are more challenging after nTBI than TBI. By examining individual rating scale categories rather than items, we highlight important differences within CRS-R items. These findings suggest that clinicians should consider item-specific challenges, especially in visual function, when assessing recovery in nTBI.

Introduction: While the Coma Recovery Scale – Revised (CRS-R) is routinely used for pediatric patients, its psychometric properties have primarily been established in adult populations. Specifically, the CRS-R is used for assessing neurobehavioral function in patients with disorders of consciousness >5 years old, while the Coma Recovery Scale-Pediatric (CRS-P) adapts this instrument for children who are <5 years old. The unidimensionality and reliability of the scale have been evaluated using the Rasch model in adults; however, this has not been replicated in the pediatric population. For this study, we combined data from both CRS-R and CRS-P to investigate preliminary evidence for the structural validity and reliability of the assessments administered to pediatric patients with disorders of consciousness. A secondary objective was to compare the hierarchy of rating scale categories generated using this study with a previously-established hierarchy for the adult population.

Methods: We conducted a retrospective analysis of data from children 1-18 years old with disorders of consciousness in inpatient rehabilitation. We used the partial credit Rasch measurement model for repeated measures.

Results: The average age of the 47 participants was 103 months (SD: 60). Each participant had >1 CRS-R/P record for a total of 1249 CRS-R records. After randomizing first and last records to a calibration/validation sample, we used the Rasch model to generate anchored item and rating scale measures along an equal-interval scale ranging from -7.10 to 6.65 logits. Items ranged from -1.29 logits (Arousal) to 2.71 logits (Communication). All items showed acceptable model fit. Principal components analysis revealed eigenvalue 1.59. The separation and person reliability indices for the CRS-R were 2.11 and 0.85, respectively. For all items, the most difficult rating scale categories were used < 5% of all observations. 81% of all CRS-R records fit the model and mean person measure was -1.39 logits. Several rating scale categories were decidedly harder (i.e., exceeded the CRS-R/P’s standard error) when comparing them with the calibrations established for adults, including a 3 on Arousal (Attention) and a 3 on Oromotor/Verbal (Intelligible Verbalization).

Conclusions: Preliminary evidence suggests that the items comprise a unidimensional hierarchy of neurobehavioral function and are sufficiently reliable for group-level decisions. Because the more challenging rating scale categories were used less frequently, they may be too difficult for young children to achieve. This is supported by the comparison with the adult hierarchy; some rating scale categories were much more difficult. However, because some categories were used fewer than 10 times in the calibration dataset, more robust analyses are necessary to determine structural validity and reliability of the CRS-R/P.

Background: The Coma Recovery Scale-Revised (CRS-R) is a widely used assessment for patients with disorders of consciousness (DoC). The rating scale categories on the CRS-R have been aligned to each state of consciousness to diagnose patients to either the unresponsive wakefulness syndrome, minimally conscious state minus, minimally conscious state plus, or emerged from the minimally conscious state. The CRS-R also provided an ordinal total raw score that ranged from 0 to 23. However, an ordinal score only indicates the direction of change between two assessments (i.e., improvement or decline). Prior work used the Rasch Measurement Model to transform the CRS-R’s ordinal, total raw score that ranges from 0 to 23 to an equal interval 0 to 100 Rasch measure. The transformed measure enables the ability to evaluate the direction and magnitude of change between two assessments.

Objectives: To evaluate change during rehabilitation for patients with DoC after severe brain injury.

Methods: We conducted a secondary analysis of data from two DoC programs. The CRS-R was conducted at baseline (i.e., admission to rehabilitation) and week 4. We applied a previously published minimal detectable change value with a 95% confidence interval (11 units) to the CRS-R change measure to determine whether participants improved or declined beyond measurement error.

Results: 323 patients with DoC had a baseline and four-week CRS-R record. The mean days post-injury at baseline were 49 (SD: 47 days). The mean age at admission to rehabilitation was 47 years (SD: 20 years). 135 (42%) patients with DoC were male, 214 (66%) were white, and 165 (51%) had a traumatic injury. 37% of patients (119/323) improved beyond measurement error and 5% of patients (17/323) declined beyond measurement error. Patients who declined or had no change beyond measurement error (n=145) had a baseline CRS-R administered an average of 53 days post-injury (SD: 32 days). Patients who improved beyond measurement error (n=91) had a baseline CRS-R administered an average of 45 days post-injury (SD: 17).

Conclusions: Patients who were admitted to post-acute care rehabilitation earlier may be a proxy for those who had fewer complications or co-occurring conditions. Identifying patients who have improved indicates the current treatment is having a positive impact. Whereas patients who are declining beyond measurement error may require additional days in rehabilitation to progress and additional evaluations to rule out co-occurring conditions such as hydrocephalus or subclinical seizures. Incorporating the minimal detectable change into clinical practice may support data-driven treatment decisions.

Introduction: Patients with severe brain injuries and disorders of consciousness (DoC) are unable to communicate and often suffer from severe dysphagia. To address these challenges, the DoC Special Interest Group of the International Brain Injury Association (IBIA) developed a survey to identify the tools currently used by speech-language therapists, assess their needs, and explore new practices for improving the rehabilitation of language and swallowing in post-comatose patients.

Methods: The survey investigated (1) socio-demographic data, (2) general questions about speech and language therapy, (3) swallowing assessment and management, (4) language and communication assessment and management. The questionnaire was initially developed in English and subsequently translated into six other languages. It was distributed widely via the Alchemer platform. The survey was open to any therapist involved in language and/or swallowing rehabilitation of patients with DoC.

Results: Responses from 137 clinicians have been collected so far. Of these, the majority were speech therapists (74%). Most respondents work in chronic rehabilitation centers (52%) or acute/subacute care settings (22%). During their academic studies, 27% of clinicians reported being exposed to post-comatose states, while 37% received formal training in the assessment and management of DoC during their professional careers. A significant proportion of therapists had experience assessing (74%) and managing (76%) swallowing disorders. Regular monitoring of swallowing function (93%) and providing guidance to patients' families on swallowing management (91%) were common practices, and approximately one third (36%) reported creating or adapting their own tools for swallowing assessment. Regarding language, most therapists assessed (80%) and managed (76%) communication in DoC patients. They frequently evaluated patients' ability to use augmentative and alternative communication (76%) and gathered information about patients' interests from family and friends (96%). Observing communication abilities across different contexts was also common (89%). Fewer therapists created or adapted tools for language assessment (31%) and management (53%). They also reported the tools they preferentially use in their practice for each clinical domain. After completing the survey, over 40% of therapists reported gaining new insights.

Conclusions: The survey highlights the lack of tools and training for speech-therapy practices in patients with severe brain injury and DoC. Early and long-term assessment and management of both language and swallowing abilities should be improved, notably by providing and/or adapting new clinical tools.

Objective: The purpose of this study was to identify, compare, and contrast the informational needs of potential user groups to develop a data visualization tool that tracks progress in adults with acquired disorders of consciousness (DoC).

Background: Data visualization is the representation of information in a chart, diagram, picture, or other pictorial format. By definition, patients with DoC may not show consistent behavioral responses to stimuli or commands and cannot engage in basic self-care and mobility activities that are frequently targeted in rehabilitation interventions. Behavioral signs of consciousness are often difficult to detect and reflexive responses can be mistaken for purposeful actions. A dashboard showing key metrics that reflect current level of consciousness and track recovery trajectory in patients with DoC can aid prognostication inform treatment decisions, and enhance communication among clinicians, patients’ families, and third-party payors.

Methods: We conducted a qualitative descriptive study informed by a learning health systems approach to understand the needs of rehabilitation therapists (n=2), family members of patients with DoC (n=2), and liaisons for third-party payors (n=3). A common set of guiding questions was used for all interviews to determine stakeholder preferences for clinical information. Analysis was descriptive and utilized a matrix approach. In collaboration with the informatics team, we identified and extracted key data elements from the electronic health record to construct a prototype dashboard.

Results: All participants desired similar content for a dashboard; however, each user group needed different contextual information and had varying design preferences. Clinicians preferred to see all key assessment tools displayed longitudinally on a single screen to support clinical reasoning. Family members preferred simple, easy to understand displays with context for score interpretation (i.e., descriptors of DoC level). Payor liaisons prioritized visualizations that would provide context for prognosis, such as time elapsed since initial injury and normative recovery trajectories. The final design includes simple graphs of key metrics as well as “drill down” views with detail on specific assessment results.

Discussion: Data visualization has great potential to support communication and shared decision-making among different stakeholders about rehabilitation progress. An advantage of engaging multiple user groups early in the design process of a data visualization tool is that a variety of needs can be anticipated and addressed, thereby increasing the likelihood of adoption and satisfaction with the resulting tool. The iterative design process in which we are engaging multiple user groups aims to bridge these needs.

Introduction: The minimally conscious state (MCS) is a prolonged disorder of consciousness (PDOC) and one of the most severe outcomes of acquired brain injury. Prevalence data are scarce. In the Netherlands, since 2019, almost all new patients with a PDOC after acquired brain injury are admitted in a so called “chain of care” embedded within a nationwide network of academic expertise. After hospital discharge, early intensive neurorehabilitation (EINP) is offered for up to 14 weeks. EINP patients who do not recover consciousness can continue their rehabilitation in a prolonged intensive neurorehabilitation program for up to 24 months post event, in one of three specialized nursing homes. The aim of this study was to establish the nationwide point prevalence of institutionalized patients in MCS in the Netherlands.

Methods: This was a descriptive cross-sectional study in which all 86 Dutch hospitals, all five specialized PDOC rehabilitation facilities, and all 274 nursing homes were asked whether they were treating patients with a PDOC on the point prevalence date of September 15, 2021. Patients were eligible for inclusion when they had a PDOC after acute acquired brain injury with a duration ≥ 4 weeks and were 16 years or older. Each patient’s legal representative provided informed consent for their inclusion. Patient level of consciousness was verified using the Coma Recovery Scale-Revised (CRS-R) in a single assessment session performed in the facility of residence by an experienced physician. Data on patient demographics, etiology, level of consciousness, facility of residence, and clinical status were collected from a questionnaire by the treating physician.

Results: Seventy patients were reported to have a PDOC, of whom six were excluded. The level of consciousness was verified for 49 patients, while for 15 it could not be verified. Of the patients verified, 38 had a PDOC, of whom 32 were in MCS (mean age 44.8 years, 68.8% male). The prevalence of institutionalized patients in MCS is 0.2–0.3/100,000 Dutch inhabitants. Traumatic brain injury was present in 21/32 (65.6%) of these patients. Specialized PDOC rehabilitation was received by 17/32 (53.1%) of patients, with the rest admitted to nursing homes. The most frequent signs of consciousness on the CRS-R were visual pursuit 15/32 (46.9%), reproducible movement to command 13/32 (40.6%), and an automatic motor response 10/32 (31.3%).

Conclusions: This study shows that a nationwide prevalence study of patients in MCS can be successfully conducted within a specialized chain of expertise in which PDOC rehabilitation and chronic care are connected. The prevalence of institutionalized patients in MCS in the Netherlands is low. These findings are used for further organization of PDOC care in the Netherlands.

Introduction: Prolonged disorders of consciousness (pDOC) remain a significant diagnostic challenge. Assessment based on behavioral responses and symptoms is often ambiguous and prone to error. Analysis of brain activity offers a way to extend the scope of diagnostic tools, providing better insight into the neurocognitive capacity of pDOC patients. In the present study, we introduce an approach based on perturbation of brain activity using auditory steady-state responses (ASSR), which are frequency-domain responses to auditory periodic stimulation. Through a series of experiments, we show evidence that these responses may help to differentiate between pDOC patients with and without awareness.

Methods: We conducted three experiments using different types of periodic auditory stimulation: Constant frequency 40 Hz stimulation (CFS), Variable frequency stimulation in the narrow band of 25-55 Hz (NBS), Variable stimulation in the broad band of 30-100 Hz (WBS). We measured averaged intertrial phase clustering/phase-locking index (ITPC/PLI) from frontocentral channels. The study included pDOC patients with diagnoses of Unresponsive Wakefulness Syndrome (UWS), Minimally Conscious State (MCS-/+), and emergent from the Minimally Conscious State (eMCS). PDOC diagnosis was assessed through repeated Coma Recovery Scale-Revised (CRS-R) administration, with the most frequent diagnosis treated as final. The integrity of the auditory pathway was tested using otoacoustic emissions and auditory brainstem responses.

Results: To maintain balanced group sizes, we compared responses between two groups: Unaware pDOC group (UWS diagnosis), and aware pDOC group (MCS-/+ and eMCS diagnoses). For the CFS condition, we observed significantly stronger responses in the aware pDOC group. A similar pattern, although less systematic, was observed for NBS and WBS stimulation. In all cases, the response around 40 Hz frequency was most sensitive to group differences.

Conclusion: Our results provide evidence that the ASSR protocol shows promise as a tool to support clinical diagnostic methods based on behavioral responses for patients with prolonged disorders of consciousness.

Background: While individuals with Disorders of Consciousness (DoC) arising from a severe traumatic brain injury (TBI) benefit from comprehensive multidisciplinary inpatient rehabilitation, only limited evidence is available to guide optimal interventions to improve consciousness (Murtaugh et al, 2024). Robotic devices offer a novel approach for efficiently integrating ingredients often emphasized for neurologic recovery including mass repetition, cardiovascular exercise, and sensory stimulation. However integration of robotics is often delayed until later stages of recovery (Burnfield et al, 2021). While feasible (William et al, 2020), there is a gap in research examining robotics’ impact on changes in consciousness during inpatient post-TBI rehabilitation. The purpose of our retrospective study was to examine and compare changes in DoC of adults undergoing inpatient rehabilitation following a severe TBI who did and did not engage in a robotic gait intervention.

Methods: Following Institutional Review Board approval, electronic health records of inpatients cared for between 7/1/2017 and 7/15/2024 were reviewed to identify English-speaking adults with a primary diagnosis of TBI resulting in DoC who underwent serial assessment using the Coma Recovery Scale-Revised (CRS-R). Individuals with SCI were excluded. Relevant data were extracted manually and validated. Descriptive and comparative statistics, including independent and paired t-tests (or non-parametric equivalents) were performed (SigmaPlot 11.0; significance, p< 0.05; average mean values reported).

Results: Use of robotic gait (i.e., Lokomat) training was determined on an individual basis by each patient’s clinical team. The Robotic group (n= 32), on average initiated device training 43.6 days (SD= 26.9) post-admission and completed an average of 12.4 (SD= 12.6) Robotic gait sessions, lasting 23.2 minutes (SD= 12.6), for a total usage time of 293.2 minutes (SD= 335.4). The No-Robotic group (n=65) received conventional multidisciplinary therapy without gait robotics. Robotic and No-Robotic cohorts did not vary significantly in age (39.9 vs. 41.0 years, respectively), gender (72% vs 77%), or initial CRS-R score (7.6 vs. 9.3 points). Length of stay was longer for the Robotic vs. No-Robotic group (127.7 vs. 81.4 days; p= 0.008). Both groups significantly increased CRS-R scores from initial to final assessment (p< 0.001); however, the overall improvement in CRS-R scores from initial to final assessment was greater for the Robotic vs. No-Robotic group (9.5 vs. 6.7; p= 0.022). The Robotic and No-Robotic groups did not differ significantly in the number of returns to acute care (0.7 vs. 0.4 events).

Discussion: Not unexpectedly, both groups evidenced significant improvements in CRS-R scores across their stay. However, improvements were greater for those who engaged in mobilization on the Robotic and risk (i.e., return to acute care) was not significantly heightened. Future, randomized controlled trials are needed to discern whether these differences persist when factors such as medical co-morbidities, medications, injury severity and time since injury are controlled.

Introduction: Electroencephalogram (EEG) features have been widely used to assess depth of anesthesia (DOA) and levels of consciousness in adults, as well as in pediatric populations. EEG features, such as permutation entropy, Lempel-Ziv complexity, Higuchi fractal dimension, Hurst exponent measured by detrended fluctuation analysis, and spectral properties have been used to assess levels of consciousness in severely brain-injured adults. However, it remains unclear whether these EEG features, which are age-dependent, can reliably track DOA and levels of consciousness in brain-injured pediatric patients. This study hypothesized that features of EEG, collected in the intensive care unit and observed during sedation would predict outcomes in severely brain-injured children.

Methods: EEG data were collected at the bedside from (n = 31) severely brain-injured children, ages 5 and older, in the pediatric intensive care units at McMaster Children’s Hospital and Alberta Children’s Hospital. All patients received anesthetic exposure to Propofol, Midazolam, or Dexmedetomidine. Outcome was measured using the Glasgow Outcome Scale – Extended (GOS-E) at 3 months post-injury, with a score of ≥4 indicating recovery and <4 indicating non-recovery. EEG preprocessing was performed using MNE Python, including filtering (0.5-45 Hz), re-referencing to mastoids, ICA for artifact removal from EOG and ECG, segmentation of 10-second epochs, and autorejection of bad epochs. Neurokit and FOOOF were used to extract 23 EEG features. Mann-Whitney U tests assessed group differences in features between recovered and non-recovered patients. A logistic regression model was trained with 23 features, age, and etiology, using stratified 5-fold cross-validation. RESULTS: The analysis included 116 files for the recovered group (n=21) and 58 files for the non-recovered group (n=10). Of the 23 EEG features, 20 showed statistically significant differences between the recovered and non-recovered groups. The logistic regression model achieved an F1-score of 80%, with precision and recall of 80% and 95% for the recovered group, and 83% and 52% for the non-recovered group. The area under the curve (AUC) for the model was 0.88, demonstrating strong performance in predicting patient outcomes.

Conclusions: This study demonstrates that EEG features collected during sedation have significant prognostic value in predicting outcomes for severely brain-injured children. Machine learning models utilizing EEG features were able to predict outcomes with an accuracy of 80%, which supports the use of temporal pattern-based EEG features for outcome assessment in pediatric neurocritical care settings.

Introduction: Upright position recovery (i.e. verticalization) is a crucial step in the rehabilitation of patients with severe acute brain injury (sABI) to prevent complications of prolonged immobilization, regain hemodynamic stability and improve muscular cutaneous trophism and cognitive performance. However, verticalization with traditional tilt table (TV) is frequently complicated by orthostatic hypotension due to blood pooling in the lower limbs and lack of leg movements. Recently, tilt-table equipped with robot-assisted lower limbs cyclic mobilization has been proposed as a safe and suitable device for accelerating the adaptation to vertical posture in bedridden patients with sABI since the acute phase. Verticalization Equipped with Mobilization (VEM) improves venous return and leads to better cardiac output. To date, contrasting and inconclusive effects of VEM on functionality, gait and level of consciousness in patients with sABI have been found.
The present multicentric study (ID: NCT06469983) aims to evaluating the effectiveness of VEM versus TV in a large cohort of patients with sABI.


Methods: According to an a priori power analysis, 198 patients with sABI including both prolonged disorder of consciousness (pDoC) and emerged from pDoC (EpDoC) will be enrolled in 6 post-acute Neurorehabilitation Units and will be randomized to receive either VEM or TV. The protocol includes 5 VEM or TV sessions per week for 4 weeks (20 sessions total). Each session will last 30 minutes. The VEM protocol will involve gradual verticalization on the Erigo® tilt table, which incorporates rhythmic passive movement of the lower limbs. The TV protocol will follow the same gradual verticalization approach using a traditional standing device. Patients’ clinical-motor-functional findings will be assessed by the following measures at study entry, after 5-week treatment and after 1 month. Additional neurophysiological data and blood biomarkers will be recorded at the same time points. Statistical analyses will be conducted blind, according to the intention-to-treat principle.

Outcomes and Measures: 1. Primary Outcome: level of consciousness in pDoC by Coma Recovery-Scale revised, cognitive functioning in EpDoC by Level of Cognitive Functioning. Secondary outcomes: spasticity by Modified Ashworth Scale, limbs movements by Medical Research Council scale in EpDoC and reflex movement responses to nociceptive stimulation in pDoC, disability by Disability Rating Scale and modified Barthel Index, professionals’ devices’ usability by System Usability Scale. Daily sessions will be monitored using the Adverse Events Report, the Agitated Behavior Scale for EpDoC, and the Nociception Coma Scale for pDoC. This work was supported by the Italian Ministry of Research, under the complementary actions to the NRRP “Fit4MedRob - Fit for Medical Robotics” Grant #NC0000007).

Conclusions: Data collection will start in November 2024. The preliminary data will be analyzed and presented at the Congress. This trial aims to provide new insights into the effectiveness of VEM in sABI rehabilitation.