Improving Dysarthric Speech Segmentation With Emulated and Synthetic Augmentation.

Acoustic features extracted from speech can help with the diagnosis of neurological diseases and monitoring of symptoms over time. Temporal segmentation of audio signals into individual words is an important pre-processing step needed prior to extracting acoustic features. Machine learning techniques could be used to automate speech segmentation via automatic speech recognition (ASR) and sequence to sequence alignment. While state-of-the-art ASR models achieve good performance on healthy speech, their performance significantly drops when evaluated on dysarthric speech. Fine-tuning ASR models on impaired speech can improve performance in dysarthric individuals, but it requires representative clinical data, which is difficult to collect and may raise privacy concerns. This study explores the feasibility of using two augmentation methods to increase ASR performance on dysarthric speech: 1) healthy individuals varying their speaking rate and loudness (as is often used in assessments of pathological speech); 2) synthetic speech with variations in speaking rate and accent (to ensure more diverse vocal representations and fairness). Experimental evaluations showed that fine-tuning a pre-trained ASR model with data from these two sources outperformed a model fine-tuned only on real clinical data and matched the performance of a model fine-tuned on the combination of real clinical data and synthetic speech. When evaluated on held-out acoustic data from 24 individuals with various neurological diseases, the best performing model achieved an average word error rate of 5.7% and a mean correct count accuracy of 94.4%. In segmenting the data into individual words, a mean intersection-over-union of 89.2% was obtained against manual parsing (ground truth). It can be concluded that emulated and synthetic augmentations can significantly reduce the need for real clinical data of dysarthric speech when fine-tuning ASR models and, in turn, for speech segmentation.

Detecting bulbar amyotrophic lateral sclerosis (ALS) using automatic acoustic analysis.

Automatic speech assessments have the potential to dramatically improve ALS clinical practice and facilitate patient stratification for ALS clinical trials. Acoustic speech analysis has demonstrated the ability to capture a variety of relevant speech motor impairments, but implementation has been hindered by both the nature of lab-based assessments (requiring travel and time for patients) and also by the opacity of some acoustic feature analysis methods. These challenges and others have obscured the ability to distinguish different ALS disease stages/severities. Validation of automated acoustic analysis tools could enable detection of early signs of ALS, and these tools could be deployed to screen and monitor patients without requiring clinic visits. Here, we sought to determine whether acoustic features gathered using an automated assessment app could detect ALS as well as different levels of speech impairment severity resulting from ALS. Speech samples (readings of a standardized, 99-word passage) from 119 ALS patients with varying degrees of disease severity as well as 22 neurologically healthy participants were analyzed, and 53 acoustic features were extracted. Patients were stratified into early and late stages of disease (ALS-early/ALS-E and ALS-late/ALS-L) based on the ALS Functional Ratings Scale-Revised bulbar score (FRS-bulb) (median [interquartile range] of FRS-bulbar scores: 11[3]). The data were analyzed using a sparse Bayesian logistic regression classifier. It was determined that the current relatively small set of acoustic features could distinguish between ALS and controls well (area under receiver-operating characteristic curve/AUROC = 0.85), that the ALS-E patients could be separated well from control participants (AUROC = 0.78), and that ALS-E and ALS-L patients could be reasonably separated (AUROC = 0.70). These results highlight the potential for automated acoustic analyses to detect and stratify ALS.

Validation of automated pipeline for the assessment of a motor speech disorder in amyotrophic lateral sclerosis (ALS).

Background and objective: Amyotrophic lateral sclerosis (ALS) frequently causes speech impairments, which can be valuable early indicators of decline. Automated acoustic assessment of speech in ALS is attractive, and there is a pressing need to validate such tools in line with best practices, including analytical and clinical validation. We hypothesized that data analysis using a novel speech assessment pipeline would correspond strongly to analyses performed using lab-standard practices and that acoustic features from the novel pipeline would correspond to clinical outcomes of interest in ALS. Methods: We analyzed data from three standard speech assessment tasks (i.e., vowel phonation, passage reading, and diadochokinesis) in 122 ALS patients. Data were analyzed automatically using a pipeline developed by Winterlight Labs, which yielded 53 acoustic features. First, for analytical validation, data were analyzed using a lab-standard analysis pipeline for comparison. This was followed by univariate analysis (Spearman correlations between individual features in Winterlight and in-lab datasets) and multivariate analysis (sparse canonical correlation analysis (SCCA)). Subsequently, clinical validation was performed. This included univariate analysis (Spearman correlation between automated acoustic features and clinical measures) and multivariate analysis (interpretable autoencoder-based dimensionality reduction). Results: Analytical validity was demonstrated by substantial univariate correlations (Spearman's ρ > 0.70) between corresponding pairs of features from automated and lab-based datasets, as well as interpretable SCCA feature groups. Clinical validity was supported by strong univariate correlations between automated features and clinical measures (Spearman's ρ > 0.70), as well as associations between multivariate outputs and clinical measures. Conclusion: This novel, automated speech assessment feature set demonstrates substantial promise as a valid tool for analyzing impaired speech in ALS patients and for the further development of these technologies.

Detecting bulbar amyotrophic lateral sclerosis (ALS) using automatic acoustic analysis.

Home-based speech assessments have the potential to dramatically improve ALS clinical practice and facilitate patient stratification for ALS clinical trials. Acoustic speech analysis has demonstrated the ability to capture a variety of relevant speech motor impairments, but implementation has been hindered by both the nature of lab-based assessments (requiring travel and time for patients) and also by the opacity of some acoustic feature analysis methods. Furthermore, these challenges and others have obscured the ability to distinguish different ALS disease stages/severities. Validation of remote-capable acoustic analysis tools could enable detection of early signs of ALS, and these tools could be deployed to screen and monitor patients without requiring clinic visits. Here, we sought to determine whether acoustic features gathered using a remote-capable assessment app could detect ALS as well as different levels of speech impairment severity resulting from ALS. Speech samples (readings of a standardized, 99-word passage) from 119 ALS patients with varying degrees of disease severity as well as 22 neurologically healthy participants were analyzed, and 53 acoustic features were extracted. Patients were stratified into early and late stages of disease (ALS-early/ALS-E and ALS-late/ALS-L) based on the ALS Functional Ratings Scale - Revised bulbar score (FRS-bulb). Data were analyzed using a sparse Bayesian logistic regression classifier. It was determined that the current relatively small set of acoustic features could distinguish between ALS and controls well (area under receiver operating characteristic curve/AUROC = 0.85), that the ALS-E patients could be separated well from control participants (AUROC = 0.78), and that ALS-E and ALS-L patients could be reasonably separated (AUROC = 0.70). These results highlight the potential for remote acoustic analyses to detect and stratify ALS.

Analytical Validation of a Webcam-Based Assessment of Speech Kinematics: Digital Biomarker Evaluation following the V3 Framework.

Introduction: Kinematic analyses have recently revealed a strong potential to contribute to the assessment of neurological diseases. However, the validation of home-based kinematic assessments using consumer-grade video technology has yet to be performed. In line with best practices for digital biomarker development, we sought to validate webcam-based kinematic assessment against established, laboratory-based recording gold standards. We hypothesized that webcam-based kinematics would possess psychometric properties comparable to those obtained using the laboratory-based gold standards. Methods: We collected data from 21 healthy participants who repeated the phrase "buy Bobby a puppy" (BBP) at four different combinations of speaking rate and volume: Slow, Normal, Loud, and Fast. We recorded these samples twice back-to-back, simultaneously using (1) an electromagnetic articulography ("EMA"; NDI Wave) system, (2) a 3D camera (Intel RealSense), and (3) a 2D webcam for video recording via an in-house developed app. We focused on the extraction of kinematic features in this study, given their demonstrated value in detecting neurological impairments. We specifically extracted measures of speed/acceleration, range of motion (ROM), variability, and symmetry using the movements of the center of the lower lip during these tasks. Using these kinematic features, we derived measures of (1) agreement between recording methods, (2) test-retest reliability of each method, and (3) the validity of webcam recordings to capture expected changes in kinematics as a result of different speech conditions. Results: Kinematics measured using the webcam demonstrated good agreement with both the RealSense and EMA (ICC-A values often ≥0.70). Test-retest reliability, measured using the absolute agreement (2,1) formulation of the intraclass correlation coefficient (i.e., ICC-A), was often "moderate" to "strong" (i.e., ≥0.70) and similar between the webcam and EMA-based kinematic features. Finally, the webcam kinematics were typically as sensitive to differences in speech tasks as EMA and the 3D camera gold standards. Discussion and Conclusions: Our results suggested that webcam recordings display good psychometric properties, comparable to laboratory-based gold standards. This work paves the way for a large-scale clinical validation to continue the development of these promising technologies for the assessment of neurological diseases via home-based methods.

3D Video Tracking Technology in the Assessment of Orofacial Impairments in Neurological Disease: Clinical Validation.

PURPOSE: This study sought to determine whether clinically interpretable kinematic features extracted automatically from three-dimensional (3D) videos were correlated with corresponding perceptual clinical orofacial ratings in individuals with orofacial impairments due to neurological disorders. METHOD: 45 participants (19 diagnosed with motor neuron diseases [MNDs] and 26 poststroke) performed two nonspeech tasks (mouth opening and lip spreading) and one speech task (repetition of a sentence "Buy Bobby a Puppy") while being video-recorded in a standardized lab setting. The color video recordings of participants were assessed by an expert clinician-a speech language pathologist-on the severity of three orofacial measures: symmetry, range of motion (ROM), and speed. Clinically interpretable 3D kinematic features, linked to symmetry, ROM, and speed, were automatically extracted from video recordings, using a deep facial landmark detection and tracking algorithm for each of the three tasks. Spearman correlations were used to identify features that were significantly correlated (p value < .05) with their corresponding clinical scores. Clinically significant kinematic features were then used in the subsequent multivariate regression models to predict the overall orofacial impairment severity score. RESULTS: Several kinematic features extracted from 3D video recordings were associated with their corresponding perceptual clinical scores, indicating clinical validity of these automatically derived measures. Different patterns of significant features were observed between MND and poststroke groups; these differences were aligned with clinical expectations in both cases. CONCLUSIONS: The results show that kinematic features extracted automatically from simple clinical tasks can capture characteristics used by clinicians during assessments. These findings support the clinical validity of video-based automatic extraction of kinematic features.

Technical and clinical considerations for electroencephalography-based biomarkers for major depressive disorder.

Major depressive disorder (MDD) is a prevalent and debilitating psychiatric disease that leads to substantial loss of quality of life. There has been little progress in developing new MDD therapeutics due to a poor understanding of disease heterogeneity and individuals' responses to treatments. Electroencephalography (EEG) is poised to improve this, owing to the ease of large-scale data collection and the advancement of computational methods to address artifacts. This review summarizes the viability of EEG for developing brain-based biomarkers in MDD. We examine the properties of well-established EEG preprocessing pipelines and consider factors leading to the discovery of sensitive and reliable biomarkers.

Reliability of Automatic Computer Vision-Based Assessment of Orofacial Kinematics for Telehealth Applications.

Introduction: Telehealth/remote assessment using readily available 2D mobile cameras and deep learning-based analyses is rapidly becoming a viable option for detecting orofacial and speech impairments associated with neurological and neurodegenerative disease during telehealth practice. However, the psychometric properties (e.g., internal consistency and reliability) of kinematics obtained from these systems have not been established, which is a crucial next step before their clinical usability is established. Methods: Participants were assessed in lab using a 3 dimensional (3D)-capable camera and at home using a readily-available 2D camera in a tablet. Orofacial kinematics was estimated from videos using a deep facial landmark tracking model. Kinematic features quantified the clinically relevant constructs of velocity, range of motion, and lateralization. In lab, all participants performed the same oromotor task. At home, participants were split into two groups that each performed a variant of the in-lab task. We quantified within-assessment consistency (Cronbach's α), reliability (intraclass correlation coefficient [ICC]), and fitted linear mixed-effects models to at-home data to capture individual-/task-dependent longitudinal trajectories. Results: Both in lab and at home, Cronbach's α was typically high (>0.80) and ICCs were often good (>0.70). The linear mixed-effect models that best fit the longitudinal data were those that accounted for individual- or task-dependent effects. Discussion: Remotely gathered orofacial kinematics were as internally consistent and reliable as those gathered in a controlled laboratory setting using a high-performance 3D-capable camera and could additionally capture individual- or task-dependent changes over time. These results highlight the potential of remote assessment tools as digital biomarkers of disease status and progression and demonstrate their suitability for novel telehealth applications.

Facial Landmark Tracking in Videos of Individuals with Neurological Impairments: Is There a Trade-off Between Smoothness and Accuracyƒ.

Orofacial kinematics are valuable markers of function and progression in a variety of neurological disorders. Recent advances in facial landmark detection have been used to improve landmark tracking in video, for example by accounting for interframe optical flow. It has been demonstrated that finetuning (a type of transfer learning) can improve the performance of some facial landmark detection systems. Here, we asked whether a neural network model that is pretrained using video data (supervision by registration, SBR) can be finetuned to improve landmark detection and tracking, using data from the Toronto Neuroface Dataset (n=36), which comprises 3 different clinical populations. We finetuned the supervision by registration (SBR) model using data from 3 individuals from each of 3 clinical populations (n=9), with or without neurological impairments. The remaining individuals from our dataset (n=27) were used for evaluation. Finetuning SBR moderately improved the model's accuracy but substantially increased the smoothness of tracked landmarks. This suggests that finetuning on video-trained models, like SBR, could improve the estimation of orofacial kinematics in individuals with neurological impairments. This could be used to improve the detection and characterization of neurological diseases using video data.Clinical Relevance-This work demonstrated that transfer learning applied to video-trained facial landmark detectors could improve the measurement of orofacial kinematics in individuals with neurological impairments.

Quantifying changes over 1 year in motor and cognitive skill after transient ischemic attack (TIA) using robotics.

Recent work has highlighted that people who have had TIA may have abnormal motor and cognitive function. We aimed to quantify deficits in a cohort of individuals who had TIA and measured changes in their abilities to perform behavioural tasks over 1 year of follow-up using the Kinarm Exoskeleton robot. We additionally considered performance and change over time in an active control cohort of migraineurs. Individuals who had TIA or migraine completed 8 behavioural tasks that assessed cognition as well as motor and sensory functionality in the arm. Participants in the TIA cohort were assessed at 2, 6, 12, and 52 weeks after symptom resolution. Migraineurs were assessed at 2 and 52 weeks after symptom resolution. We measured overall performance on each task using an aggregate metric called Task Score and quantified any significant change in performance including the potential influence of learning. We recruited 48 individuals to the TIA cohort and 28 individuals to the migraine cohort. Individuals in both groups displayed impairments on robotic tasks within 2 weeks of symptom cessation and also at approximately 1 year after symptom cessation, most commonly in tests of cognitive-motor integration. Up to 51.3% of people in the TIA cohort demonstrated an impairment on a given task within 2-weeks of symptom resolution, and up to 27.3% had an impairment after 1 year. In the migraine group, these numbers were 37.5% and 31.6%, respectively. We identified that up to 18% of participants in the TIA group, and up to 10% in the migraine group, displayed impairments that persisted for up to 1 year after symptom resolution. Finally, we determined that a subset of both cohorts (25-30%) experienced statistically significant deteriorations in performance after 1 year. People who have experienced transient neurological symptoms, such as those that arise from TIA or migraine, may continue to experience lasting neurological impairments. Most individuals had relatively stable task performance over time, with some impairments persisting for up to 1 year. However, some individuals demonstrated substantial changes in performance, which highlights the heterogeneity of these neurological disorders. These findings demonstrate the need to consider factors that contribute to lasting neurological impairment, approaches that could be developed to alleviate the lasting effects of TIA or migraine, and the need to consider individual neurological status, even following transient neurological symptoms.

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.

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

BACKGROUND: Multiple sclerosis (MS) causes pervasive motor, sensory and cognitive dysfunction. The Expanded Disability Status Scale (EDSS) is the gold standard for assessing MS disability. The EDSS is biased towards mobility and may not accurately measure MS-related disabilities in the upper limb or in cognitive functions (e.g. executive function). OBJECTIVE: Our objectives were to determine the feasibility of using the Kinarm robotic system to quantify neurological deficits related to arm function and cognition in MS patients, and examine relationships between traditional clinical assessments and Kinarm variables. METHODS: Individuals with MS performed 8 robotic tasks assessing motor, cognitive, and sensory ability. We additionally collected traditional clinical assessments and compared these to the results of the robotic assessment. RESULTS: Forty-three people with MS were assessed. Most participants could complete the robotic assessment. Twenty-six (60%) were impaired on at least one cognitive task and twenty-six (60%) were impaired on at least one upper-limb motor task. Cognitive domain task performance correlated most strongly with the EDSS. CONCLUSIONS: Kinarm robotic assessment of people with MS is feasible, can identify a broad range of upper-limb motor and sensory, as well as cognitive, impairments, and complements current clinical rating scales in the assessment of MS-related disability.

Statistical measures of motor, sensory and cognitive performance across repeated robot-based testing.

BACKGROUND: Traditional clinical assessments are used extensively in neurology; however, they can be coarse, which can also make them insensitive to change. Kinarm is a robotic assessment system that has been used for precise assessment of individuals with neurological impairments. However, this precision also leads to the challenge of identifying whether a given change in performance reflects a significant change in an individual's ability or is simply natural variation. Our objective here is to derive confidence intervals and thresholds of significant change for Kinarm Standard Tests™ (KST). METHODS: We assessed participants twice within 15 days on all tasks presently available in KST. We determined the 5-95% confidence intervals for each task parameter, and derived thresholds for significant change. We tested for learning effects and corrected for the false discovery rate (FDR) to identify task parameters with significant learning effects. Finally, we calculated intraclass correlation of type ICC [1, 2] (ICC-C) to quantify consistency across assessments. RESULTS: We recruited an average of 56 participants per task. Confidence intervals for Z-Task Scores ranged between 0.61 and 1.55, and the threshold for significant change ranged between 0.87 and 2.19. We determined that 4/11 tasks displayed learning effects that were significant after FDR correction; these 4 tasks primarily tested cognition or cognitive-motor integration. ICC-C values for Z-Task Scores ranged from 0.26 to 0.76. CONCLUSIONS: The present results provide statistical bounds on individual performance for KST as well as significant changes across repeated testing. Most measures of performance had good inter-rater reliability. Tasks with a higher cognitive burden seemed to be more susceptible to learning effects, which should be taken into account when interpreting longitudinal assessments of these tasks.

Assessing various sensorimotor and cognitive functions in people with epilepsy is feasible with robotics.

BACKGROUND: Epilepsy is a common neurological disorder characterized by recurrent seizures, along with comorbid cognitive and psychosocial impairment. Current gold standards of assessment can quantify cognitive and motor performance, but may not capture all subtleties of behavior. Here, we study the feasibility of assessing various upper limb sensorimotor and cognition functions in people with epilepsy using the Kinarm robotic assessment system. We quantify performance across multiple behavioral domains and additionally consider the possible effects of epilepsy subtype and medication. METHODS: We recruited individuals with a variety of epilepsy subtypes. Participants performed 8 behavioral tasks that tested motor, cognitive, and sensory domains. We collected data on the same tasks from a group of control participants that had no known neurological impairments. We quantified performance using Task Scores, which provide a composite measure of overall performance on a given task and are adjusted for age, sex, and handedness. RESULTS: We collected data from 46 individuals with epilepsy and 92 control participants. The assessment was well-tolerated, with no adverse events recorded. Cognitive tasks testing spatial working memory, executive function, and motor response inhibition were the most frequently impaired in the epilepsy cohort, with 33/46 (72%) being outside the normal range on at least one of these tasks. Additionally, 29/46 (63%) were impaired on at least one task testing primarily motor skill, and 14/46 (30%) were impaired on a proprioceptive sensory task. People with either focal epilepsy or generalized epilepsy performed significantly worse on both motor and cognitive tasks than control participants after correcting for multiple comparisons. There were no statistical differences between generalized and focal epilepsy groups on Task Scores. Finally, individuals taking topiramate trended toward having worse performance on a spatial working memory task than other individuals with epilepsy who were not taking topiramate. CONCLUSIONS: Kinarm robotic assessment is feasible in individuals with epilepsy and is well-tolerated. Our robotic paradigm can detect impairments in various sensorimotor and cognitive functions across the population with epilepsy. Future studies will explore the role of epilepsy subtype and medications.

The Impact of Transient Ischemic Attack (TIA) on Brain and Behavior.

Transient ischemic attack (TIA) was originally defined as self-resolving focal cerebral ischemia with symptoms lasting <24 h. The newer definition also added the limitation that there should be no evidence of acute brain tissue infarction, to recognize that acute injury to the brain can result from ischemia of <24-h duration. However, several recent findings suggest that having a TIA correlates with deficits that can persist far beyond the resolution of clinical symptoms, even in the absence of imaging evidence of ischemic tissue injury. These deficits may be the result of subtle perturbations to brain structure and/or function that are not easily appreciated using the standard clinical and imaging tools that are currently employed in practice. Here, we will discuss evidence that suggests that TIA may lead to lasting changes to the structure and function of the brain.

The feasibility of using robotic technology to quantify sensory, motor, and cognitive impairments associated with ALS.

OBJECTIVE: We used the KINARM robot to quantify impairments in cognitive and upper-limb sensorimotor performance in a cohort of people with amyotrophic lateral sclerosis (ALS). We sought to study the feasibility of using this technology for ALS research, to quantify patterns of impairments in individuals living with ALS, and elucidate correlations between robotic and traditional clinical behavioral measures. METHODS: Participants completed robot-based behavioral tasks testing sensorimotor, cognitive, and proprioceptive performance. Performance on robotic tasks was normalized to a large healthy control cohort (no neurological impairments), adjusted for age. Task impairment was defined as performance outside the 95% range of controls. Traditional clinical tests included: Frontal Assessment Battery (FAB), ALS Functional Rating Scale-Revised (ALSFRS-R), and Montreal Cognitive Assessment (MoCA). RESULTS: Seventeen people with ALS were assessed. Two participants reported pain or discomfort from the robot's seat and 2 others reported discomfort from arm position during the assessment (both rectified and did not affect exam completion). Participants were able to perform the majority of the robotic tasks, although 9 participants were unable to complete 1 or more tasks. Between 20 and 69% of participants displayed sensorimotor impairments; 19 and 69% displayed cognitive task impairments; 25% displayed proprioceptive impairments. MoCA was impaired in 9/17 participants; 10/17 had impaired performance on FAB. MoCA and FAB correlated well with robot-based measures of cognition. CONCLUSION: Use of robotic assessment is generally feasible for people with ALS. Individuals with ALS have sensorimotor impairments as expected, and some demonstrate substantial cognitive impairments.

Using principal component analysis to reduce complex datasets produced by robotic technology in healthy participants.

BACKGROUND: The KINARM robot produces a granular dataset of participant performance metrics associated with proprioceptive, motor, visuospatial, and executive function. This comprehensive battery includes several behavioral tasks that each generate 9 to 20 metrics of performance. Therefore, the entire battery of tasks generates well over 100 metrics per participant, which can make clinical interpretation challenging. Therefore, we sought to reduce these multivariate data by applying principal component analysis (PCA) to increase interpretability while minimizing information loss. METHODS: Healthy right-hand dominant participants were assessed using a bilateral KINARM end-point robot. Subjects (Ns = 101-208) were assessed using 6 behavioral tasks and automated software generated 9 to 20 metrics related to the spatial and temporal aspects of subject performance. Data from these metrics were converted to Z-scores prior to PCA. The number of components was determined from scree plots and parallel analysis, with interpretability considered as a qualitative criterion. Rotation type (orthogonal vs oblique) was decided on a per task basis. RESULTS: The KINARM performance data, per task, was substantially reduced (range 67-79%), while still accounting for a large amount of variance (range 70-82%). The number of KINARM parameters reduced to 3 components for 5 out of 6 tasks and to 5 components for the sixth task. Many components were comprised of KINARM parameters with high loadings and only some cross loadings were observed, which demonstrates a strong separation of components. CONCLUSIONS: Complex participant data produced by the KINARM robot can be reduced into a small number of interpretable components by using PCA. Future applications of PCA may offer potential insight into specific patterns of sensorimotor impairment among patient populations.