The feasibility of assessing cognitive and motor function in multiple sclerosis patients using robotics

Relationship between somatosensory and visuo-perceptual impairments and motor functions in adults with hemiparetic cerebral palsy

Introduction

Children with cerebral palsy (CP) exhibit a variety of sensory impairments that can interfere with motor performance, but how these impairments persist into adulthood needs further investigation. The objective of this study was to describe the sensory impairments in adults having CP and how they relate to motor impairments.

Methods

Nineteen adults having CP performed a set of robotic and clinical assessments. These assessments were targeting different sensory functions and motor functions (bilateral and unilateral tasks). Frequency of each type of impairments was determined by comparing individual results to normative data. Association between the sensory and motor impairments was assessed with Spearman correlation coefficient.

Results

Impairment in stereognosis was the most frequent, affecting 57.9% of participants. Although less frequently impaired (26.3%), tactile discrimination was associated with all the motor tasks (unilateral and bilateral, either robotic or clinical). Performance in robotic motor assessments was more frequently associated with sensory impairments than with clinical assessments. Finally, sensory impairments were not more closely associated with bilateral tasks than with unilateral tasks.

Discussion

Somatosensory and visuo-perceptual impairments are frequent among adults with CP, with 84.2% showing impairments in at least one sensory function. These sensory impairments show a moderate association with motor impairments.

Effects of transcranial direct current stimulation over human motor cortex on cognitive-motor and sensory-motor functions

The primary motor cortex (M1) is broadly acknowledged for its crucial role in executing voluntary movements. Yet, its contributions to cognitive and sensory functions remain largely unexplored. Transcranial direct current stimulation (tDCS) is a noninvasive neurostimulation method that can modify brain activity, thereby enabling the establishment of a causal link between M1 activity and behavior. This study aimed to investigate the online effects of tDCS over M1 on cognitive-motor and sensory-motor functions. Sixty-four healthy participants underwent either anodal or sham tDCS while concurrently performing a set of standardized robotic tasks. These tasks provided sensitive and objective assessments of brain functions, including action selection, inhibitory control, cognitive control of visuomotor skills, proprioceptive sense, and bimanual coordination. Our results revealed that anodal tDCS applied to M1 enhances decision-making capacity in selecting appropriate motor actions and avoiding distractors compared to sham stimulation, suggesting improved action selection and inhibitory control capabilities. Furthermore, anodal tDCS reduces the movement time required to accomplish bimanual movements, suggesting enhanced bimanual performance. However, we found no impact of anodal tDCS on cognitive control of visuomotor skills and proprioceptive sense. This study suggests that augmenting M1 activity via anodal tDCS influences cognitive-motor and sensory-motor functions in a task-dependent manner.

Examining the Influence of Cognition on the Relationship Between Backward Walking and Falls in Persons With Multiple Sclerosis

BACKGROUND

Multiple sclerosis (MS) causes motor, cognitive, and sensory impairments that result in injurious falls. Current fall risk measures in MS (ie, forward walking [FW] speed and balance) are limited in their sensitivity. Backward walking (BW) velocity is a sensitive marker of fall risk and correlates with information processing speed (IPS) and visuospatial memory (VSM) in persons with MS. Backward walking is a complex motor task that requires increased cognitive demands, which are negatively affected by MS; however, whether cognitive function modifies the sensitivity of BW as a fall risk assessment in MS remains unknown. This study examines the influence of cognition on the relationship between BW and falls in persons with MS.

METHODS

Measures of BW, FW, IPS, VSM, and retrospective falls were collected. Hierarchical regression tested moderation and included an interaction term predicting number of falls. Covariates for all analyses included age and disease severity.

RESULTS

Thirty-eight persons with MS participated. Although BW, IPS, and covariates significantly predicted the number of falls (R 2 = 0.301; P = .016), there was no evidence of moderation. Backward walking, VSM, and covariates also significantly predicted number of falls (R 2 = 0.332, P = .008), but there was no evidence of moderation. The FW models generated comparable results.

CONCLUSIONS

The relationship between BW velocity and falls was not conditional on IPS or VSM in this sample. Larger-scale studies examining additional cognitive domains commonly affected by MS and prospective falls are needed to characterize neurobiological processes relevant to BW and its clinical application in the assessment of fall risk.

Reliability, validity and clinical usability of a robotic assessment of finger proprioception in persons with multiple sclerosis

BACKGROUND

Multiple sclerosis often leads to proprioceptive impairments of the hand. However, it is challenging to objectively assess such deficits using clinical methods, thereby also impeding accurate tracking of disease progression and hence the application of personalized rehabilitation approaches.

OBJECTIVE

We aimed to evaluate test-retest reliability, validity, and clinical usability of a novel robotic assessment of hand proprioceptive impairments in persons with multiple sclerosis (pwMS).

METHODS

The assessment was implemented in an existing one-degree of freedom end-effector robot (ETH MIKE) acting on the index finger metacarpophalangeal joint. It was performed by 45 pwMS and 59 neurologically intact controls. Additionally, clinical assessments of somatosensation, somatosensory evoked potentials and usability scores were collected in a subset of pwMS.

RESULTS

The test-retest reliability of robotic task metrics in pwMS was good (ICC=0.69-0.87). The task could identify individuals with impaired proprioception, as indicated by the significant difference between pwMS and controls, as well as a high impairment classification agreement with a clinical measure of proprioception (85.00-86.67%). Proprioceptive impairments were not correlated with other modalities of somatosensation. The usability of the assessment system was satisfactory (System Usability Scale ≥73.10).

CONCLUSION

The proposed assessment is a promising alternative to commonly used clinical methods and will likely contribute to a better understanding of proprioceptive impairments in pwMS.

Assessment of Neurological Impairment and Recovery Using Statistical Models of Neurologically Healthy Behavior

While many areas of medicine have benefited from the development of objective assessment tools and biomarkers, there have been comparatively few improvements in techniques used to assess brain function and dysfunction. Brain functions such as perception, cognition, and motor control are commonly measured using criteria-based, ordinal scales which can be coarse, have floor/ceiling effects, and often lack the precision to detect change. There is growing recognition that kinematic and kinetic-based measures are needed to quantify impairments following neurological injury such as stroke, in particular for clinical research and clinical trials. This paper will first consider the challenges with using criteria-based ordinal scales to quantify impairment and recovery. We then describe how kinematic-based measures can overcome many of these challenges and highlight a statistical approach to quantify kinematic measures of behavior based on performance of neurologically healthy individuals. We illustrate this approach with a visually-guided reaching task to highlight measures of impairment for individuals following stroke. Finally, there has been considerable controversy about the calculation of motor recovery following stroke. Here, we highlight how our statistical-based approach can provide an effective estimate of impairment and recovery.

Upper Limb Sensory-Motor Control During Exposure to Different Mechanical Environments in Multiple Sclerosis Subjects With No Clinical Disability

Multiple sclerosis (MS) is an autoimmune and neurodegenerative disease resulting in motor impairments associated with muscle weakness and lack of movement coordination. The goal of this work was to quantify upper limb motor deficits in asymptomatic MS subjects with a robot-based assessment including performance and muscle synergies analysis. A total of 7 subjects (MS: 3 M−4 F; 42 ± 10 years) with clinically definite MS according to McDonald criteria, but with no clinical disability, and 7 age- and sex-matched subjects without a history of neurological disorders participated in the study. All subjects controlled a cursor on the computer screen by moving their hand or applying forces in 8 coplanar directions at their self-selected speed. They grasped the handle of a robotic planar manipulandum that generated four different environments: null, assistive or resistive forces, and rigid constraint. Simultaneously, the activity of 15 upper body muscles was recorded. Asymptomatic MS subjects generated less smooth and less accurate cursor trajectories than control subjects in controlling a force profile, while the end-point error was significantly different also in the other environments. The EMG analysis revealed different muscle activation patterns in MS subjects when exerting isometric forces or when moving in presence of external forces generated by a robot. While the two populations had the same number and similar structure of muscle synergies, they had different activation profiles. These results suggested that a task requiring to control forces against a rigid environment allows better than movement tasks to detect early sensory-motor signs related to the onset of symptoms of multiple sclerosis and to differentiate between stages of the disease.

Assessing proprioceptive acuity in people with multiple sclerosis

Background

Proprioceptive acuity and impairments in proprioceptively guided reaches have not been comprehensively examined in people with multiple sclerosis (MS).

Objective

To examine proprioceptive acuity in people with MS who self-report and who do not self-report upper limb (UL) impairment, and to determine how people with MS reach proprioceptive targets.

Methods

Twenty-four participants with MS were recruited into two groups based on self-reported UL impairment: MS-R (i.e. report UL impairment; n  =  12) vs. MS-NR (i.e. do not report UL impairment; n  =  12). Proprioception was assessed using ipsilateral and contralateral robotic proprioceptive matching tasks.

Results

Participants in the MS-R group demonstrated worse proprioceptive acuity compared to the MS-NR group on the ipsilateral and contralateral robotic matching tasks. Analyses of reaches to proprioceptive targets further revealed that participants in the MS-R group exhibited deficits in movement planning, as demonstrated by greater errors at peak velocity in the contralateral matching task in comparison to the MS-NR group.

Conclusion

Our findings suggest that people with MS who self-report UL impairment demonstrate worse proprioceptive acuity, as well as poorer movement planning in comparison to people with MS who do not report UL impairment.

Impairments in Cognitive Control Using a Reverse Visually Guided Reaching Task Following Stroke

BACKGROUND

Cognitive and motor function must work together quickly and seamlessly to allow us to interact with a complex world, but their integration is difficult to assess directly. Interactive technology provides opportunities to assess motor actions requiring cognitive control.

OBJECTIVE

To adapt a reverse reaching task to an interactive robotic platform to quantify impairments in cognitive-motor integration following stroke.

METHODS

Participants with subacute stroke (N=59) performed two tasks using the Kinarm: Reverse Visually Guided Reaching (RVGR) and Visually Guided Reaching (VGR). Tasks required subjects move a cursor "quickly and accurately" to virtual targets. In RVGR, cursor motion was reversed compared to finger motion (i.e., hand moves left, cursor moves right). Task parameters and Task Scores were calculated based on models developed from healthy controls, and accounted for the influence of age, sex, and handedness.

RESULTS

Many stroke participants (86%) were impaired in RVGR with their affected arm (Task Score > 95% of controls). The most common impairment was increased movement time. Seventy-three percent were also impaired with their less affected arm. The most common impairment was larger initial direction angles of reach. Impairments in RVGR improved over time, but 71% of participants tested longitudinally were still impaired with the affected arm ∼6 months post-stroke. Importantly, although 57% were impaired with the less affected arm at 6 months, these individuals were not impaired in VGR.

CONCLUSIONS

Individuals with stroke were impaired in a reverse reaching task but many did not show similar impairments in a standard reaching task, highlighting selective impairment in cognitive-motor integration.

Personalized prediction of rehabilitation outcomes in multiple sclerosis: a proof-of-concept using clinical data, digital health metrics, and machine learning

Predicting upper limb neurorehabilitation outcomes in persons with multiple sclerosis (pwMS) is essential to optimize therapy allocation. Previous research identified population-level predictors through linear models and clinical data. This work explores the feasibility of predicting individual neurorehabilitation outcomes using machine learning, clinical data, and digital health metrics. Machine learning models were trained on clinical data and digital health metrics recorded pre-intervention in 11 pwMS. The dependent variables indicated whether pwMS considerably improved across the intervention, as defined by the Action Research Arm Test (ARAT), Box and Block Test (BBT), or Nine Hole Peg Test (NHPT). Improvements in ARAT or BBT could be accurately predicted (88% and 83% accuracy) using only patient master data. Improvements in NHPT could be predicted with moderate accuracy (73%) and required knowledge about sensorimotor impairments. Assessing these with digital health metrics over clinical scales increased accuracy by 10%. Non-linear models improved accuracy for the BBT (+ 9%), but not for the ARAT (-1%) and NHPT (-2%). This work demonstrates the feasibility of predicting upper limb neurorehabilitation outcomes in pwMS, which justifies the development of more representative prediction models in the future. Digital health metrics improved the prediction of changes in hand control, thereby underlining their advanced sensitivity.

Graphical Abstract

Principal Components Analysis Using Data Collected From Healthy Individuals on Two Robotic Assessment Platforms Yields Similar Behavioral Patterns

BACKGROUND

Kinarm Standard Tests (KSTs) is a suite of upper limb tasks to assess sensory, motor, and cognitive functions, which produces granular performance data that reflect spatial and temporal aspects of behavior (>100 variables per individual). We have previously used principal component analysis (PCA) to reduce the dimensionality of multivariate data using the Kinarm End-Point Lab (EP). Here, we performed PCA using data from the Kinarm Exoskeleton Lab (EXO), and determined agreement of PCA results across EP and EXO platforms in healthy participants. We additionally examined whether further dimensionality reduction was possible by using PCA across behavioral tasks.

METHODS

Healthy participants were assessed using the Kinarm EXO (N = 469) and EP (N = 170-200). Four behavioral tasks (six assessments in total) were performed that quantified arm sensory and motor function, including position sense [Arm Position Matching (APM)] and three motor tasks [Visually Guided Reaching (VGR), Object Hit (OH), and Object Hit and Avoid (OHA)]. The number of components to include per task was determined from scree plots and parallel analysis, and rotation type (orthogonal vs. oblique) was decided on a per-task basis. To assess agreement, we compared principal components (PCs) across platforms using distance correlation. We additionally considered inter-task interactions in EXO data by performing PCA across all six behavioral assessments.

RESULTS

By applying PCA on a per task basis to data collected using the EXO, the number of behavioral parameters were substantially reduced by 58-75% while accounting for 76-87% of the variance. These results compared well to the EP analysis, and we found good-to-excellent agreement values (0.75-0.99) between PCs from the EXO and those from the EP. Finally, we were able to reduce the dimensionality of the EXO data across tasks down to 16 components out of a total of 76 behavioral parameters, which represents a reduction of 79% while accounting for 73% of the total variance.

CONCLUSION

PCA of Kinarm robotic assessment appears to capture similar relationships between kinematic features in healthy individuals and is agnostic to the robotic platform used for collection. Further work is needed to investigate the use of PCA-based data reduction for the characterization of neurological deficits in clinical populations.