Michelle J. Johnson

Background: Reliable EEG biomarkers of brain-network engagement could personalize action-observation (AO) therapy after stroke. Objective: Identify functional-connectivity (FC) metrics that most robustly decode AO stimuli. Methods: Five right-handed adults (21-29 y) viewed 120 video trials (robot/human limb actions + controls) while 32-channel EEG was recorded. Ten central-region channels were filtered (alpha and beta bands) and FC matrices (10×10) computed using coherence (COH), imaginary coherence (iCOH), phase-locking value, partial directed coherence (PDC) and spectral Granger causality (SpcG). A graph neural network (GNN) was trained with stratified 15-fold cross validation for two tasks: (1) AO-condition decoding : six classes -- human-left, human-right, robot-left, robot-right, baseline, and landscape; (2) Action-type decoding : five upper-limb actions -- air punch, back-and-forth arm swing, lateral arm swing, overhead arm raise, and wave. Results: iCOH achieved the highest performance across both tasks (macro-AUC 0.997&1.000; balanced accuracy 0.96 - 1.00). Directed metrics PDC and SpcG also performed strongly (macro-AUC ≥ 0.99). Findings persisted despite class imbalance and small sample size. Conclusions: Volume-conduction-invariant (iCOH) and directed FC measures provide robust signatures of motor-and-cognitive-network engagement during AO. These EEG markers may inform adaptive AO therapy or BCI-guided rehabilitation post-stroke. Larger cohorts will validate clinical utility.

After stroke, motor control deficits are a common cause of disability. An optimal control model could provide insight into how motor and cognitive impairments influence post-stroke sensorimotor control. We present an initial attempt at modeling how post-stroke individuals perform a robot-based preview tracking task. Specifically, we fit a feedback controller with added delay and position bias to the impaired and less-impaired limbs of 11 stroke patients. We use an inverse linear quadratic regulator (LQR) to find the cost function for each participant. We also use linear regression to determine relationships between model parameters and clinical assessments of motor and cognitive impairments. Our main findings are that post-stroke sensorimotor control can be modeled as an optimal feedback controller and adding a delay and bias constant may improve the performance of this model. We also comment on how model parameters may be related to clinical assessments via linear regression. Future work should model more individuals and evaluate kinematic performance.

Rehabilitation therapy can be more effective and engaging when interactive technologies are involved. To enhance this experience, we integrated a haptic/biometric-based (HBB) Dynamic Difficulty Adjustment (DDA) system into the enAblegames ${}^{\text {TM}}$ platform, which already uses body tracking for therapeutic gaming. This system adapts game and haptic difficulty in real time based on each patient's biometric data and performance, making therapy more personalized. We tested this system with 11 participants, comparing their experiences with and without DDA. The results were promising-36% preferred DDA-enhanced games, compared to just $\mathbf{7 \%}$ for non-DDA, and in a single-game scenario, preference for DDA increased by 50 %. These early findings suggest that HBB-DDA can make rehabilitation more engaging and tailored to individual needs. While more research is needed to understand its full impact, this system has the potential to improve patient experience and therapy outcomes.

Action Observation (AO) therapy leverages the mirror neuron system (MNS) and may support motor recovery in neurorehabilitation. In this study, we integrated Flo v2, a humanoid robot equipped with grippers and object detection system, into an AO therapy paradigm with electroencephalography (EEG) monitoring. Flo v2's enhanced design enables the execution of upper-limb actions, either transitive (involving object interaction such as grasping a cup) or intransitive (gesture-based without object manipulation such as waving). The robot control is synchronized with EEG recording to facilitate the investigation of cortical responses during AO tasks. We also conducted a case study to assess of the upgraded robot system's feasibility. Three healthy participants observed and imitated robot-performed actions, where the robot actor was in person or on videos. Exploratory analyses of EEG signals examined sensorimotor mu event-related desynchronization (ERD) during video-based and in-person AO tasks. Results indicated stronger responses during bimanual and transitive AO in the in-person settings. However, individual variability in cortical responses was evident, with one subject showing less pronounced ERD patterns, and that comparisons of mu ERDs across different types of action in video-based AO settings were inconsistent among subjects. Flo v2's enhancements demonstrated its feasibility as a tool for robot-mediated AOE therapy and highlighted potential for further neurorehabilitation research.

Stroke, a condition in which blood flow has been restricted to parts of the brain, has been known to cause spasticity, a condition in which the muscles abnormally tense up through an overactivation of the stretch reflex. One low-cost solution to this problem is the use of self-operated passive stretching devices that can be operated with minimal assistance. The goal of this project was to develop and evaluate a new self-stretching wrist and finger device using an iterative design cycle. The device was designed with the following two stretching mechanisms: a rotating dial and an inflatable grip. Three stages of the device prototypes were developed: early, late, and final. After each stage, a usability survey was conducted with a group of healthy controls to rate 0-10 in criteria of overall comfort, unexpected pain, upper arm muscle tension, and device security of the arm. After, design changes were made to address common feedback. In the end, we saw that prioritizing the reduction of pain, decrease of unnecessary upper arm muscle tension, and increase of device security of the arm proved to be valuable measures of increasing overall device comfort. In addition, electromyography (EMG) data was collected with a healthy control subject to verify that the device was able to properly initiate the wrist flexors and extensors through the flexion and extension movement, and that each muscle group was isolated properly to their respective movements through the device.

Action observation (AO) therapy is a promising rehabilitative treatment for motor and language function in individuals recovering from neurological conditions, such as stroke. This pilot study aimed to investigate the potential of humanoid robots to support AO therapy in rehabilitation settings. The brain activity of three healthy right-handed participants was monitored with electroencephalography (EEG) while they observed eight different actions performed by two agents, a human actor and a robot, using their left and right arms. Their event-related spectral perturbations (ERSPs, changes in the spectral power of neural oscillations in response to an event or stimulus, compared to baseline) in sensorimotor regions were analyzed. The single-subject analysis showed variability in ERSP patterns among all participants, including power suppression in sensorimotor mu and beta rhythms. One participant showed stronger responses to "robot" AO conditions than to "human" conditions. Strong and positive correlations in ERSP across all conditions were observed for almost all participants and channels, implying common cognitive processes or neural networks at play in the mirror neuron system during AO. The results support the feasibility of using EEG to explore differences in neural responses to observation of robot- and human-induced actions.

BACKGROUND: Major traumatic injury in the pediatric population requires further evaluation to improve patient outcomes. Relatively few Canadian studies have investigated pediatric trauma using population-based data. Our objectives were to describe the epidemiology of pediatric major trauma in Nova Scotia and identify factors associated with in-hospital mortality. METHODS: Retrospective cohort study of pediatric major trauma patients (age <18 years) injured in Nova Scotia over a 17-year period (April 2001-March 2018). Data were collected from the Nova Scotia Trauma Registry. Characteristics were compared between patient subgroups using t-tests, chi-square analyses and Fisher's exact test. Temporal trends were evaluated using the Mann-Kendall test. Incidence and mortality rates were mapped using ArcGIS Pro. A multivariate logistic regression model was created to assess for factors associated with in-hospital mortality. RESULTS: A total of 1258 injuries were observed over the 17-year study period. The incidence of pediatric major trauma was 41.7 per 100,000 person-years. Most patients were male (819/1258; 65.1 %) and resided in urban areas (764/1258; 60.7 %). Blunt trauma accounted for 86.2 % (1084/1258) of injuries, and motor vehicle collisions were the most common cause (448/1258; 35.6 %). Incidence and mortality rates were highest in the 15-17 year age group, with a trend towards increasing incidence among females (p = 0.011). Mortality was 17.2 % (217/1258) of patients; 10.9 % (137/1258) died pre-hospital. No trends were detected in mortality rates. The regression model showed increased odds of in-hospital mortality for every point increase in the ISS (OR 1.05; 95 % CI 1.02 to 1.09) and for every unit decrease in scene GCS (OR 0.63; 95 % CI 0.56-0.71). Rural patients were 2 times more likely to die in-hospital versus urban patients (OR 2.40; 95 % CI 1.01-5.69), and patients injured at home were 6 times more likely to die compared to those injured in other locations (OR 6.19; 95 % CI 1.01-38.11). CONCLUSION: Pediatric trauma remains a major public health issue in Canada and beyond. Greater efforts are required to expand our understanding of trauma epidemiology and develop targeted injury prevention strategies, especially for rural inhabitants.