Measuring functional hand use in children with unilateral cerebral palsy using accelerometry and machine learning.

AIM: To investigate wearable sensors for measuring functional hand use in children with unilateral cerebral palsy (CP). METHOD: Dual wrist-worn accelerometry data were collected from three females and seven males with unilateral CP (mean age = 10 years 2 months [SD 3 years]) while performing hand tasks during video-recorded play sessions. Video observers labelled instances of functional and non-functional hand use. Machine learning was compared to the conventional activity count approach for identifying unilateral hand movements as functional or non-functional. Correlation and agreement analyses compared the functional usage metrics derived from each method. RESULTS: The best-performing machine learning approach had high precision and recall when trained on an individual basis (F1 = 0.896 [SD 0.043]). On an individual basis, the best-performing classifier showed a significant correlation (r = 0.990, p < 0.001) and strong agreement (bias = 0.57%, 95% confidence interval = -4.98 to 6.13) with video observations. When validated in a leave-one-subject-out scenario, performance decreased significantly (F1 = 0.584 [SD 0.076]). The activity count approach failed to detect significant differences in non-functional or functional hand activity and showed no significant correlation or agreement with the video observations. INTERPRETATION: With further development, wearable accelerometry combined with machine learning may enable quantitative monitoring of everyday functional hand use in children with unilateral CP.

Personalized adapted locomotor training for an individual with sequelae of West Nile virus infection: a mixed-method case report.

BACKGROUND: West Nile virus (WNV) can have severe consequences, including encephalitis and paralysis. Purpose: To describe the benefits of intensive locomotor training (LT) for an individual with a previous WNV infection resulting in chronic paraplegia. Case Description: The patient, who became a wheelchair user following standard rehabilitation, began LT 3 years post infection. Her goals included standing and walking with an assistive device and transferring independently. The intervention consisted of bodyweight-supported treadmill training and overground training, which involved walking, balancing, strengthening, and transferring activities. Outcomes: Following 5 months of LT, the patient ambulated independently with a walker at a speed = 0.34m/s. She walked 110.1 metres in 6 minutes and increased her Berg Balance Scale score by 17 points. These improvements were either maintained or further increased 3 months post LT. The patient's perspectives on LT were collected through a semi-structured interview. A conventional content analysis, which uses data to drive themes, revealed three themes: (1) recalibrating goals, (2) outcomes (i.e. physical and psychological benefits, such as a sense of accomplishment), and (3) challenges of LT and effective coping strategies. Conclusions: The patient demonstrated improved balance and walking abilities. Intensive LT was feasible and effective for this individual with chronic paraplegia due to WNV infection.

Falls after spinal cord injury: a systematic review and meta-analysis of incidence proportion and contributing factors.

INTRODUCTION: Falls are detrimental to the well-being of individuals with spinal cord injury (SCI). To establish effective fall prevention initiatives, a comprehensive understanding of falls after SCI is needed. OBJECTIVES: To report the incidence proportion of falls and summarize the factors contributing to falls in individuals with SCI. STUDY DESIGN: Systematic review and meta-analysis. METHODS: Eight databases were searched. Abstracts/full articles were screened by two researchers independently. Data concerning study design, participant characteristics, and the incidence proportion, factors, and consequences of falls were extracted. Risk of bias was assessed using a domain-based approach that considered sampling and measurement bias. The incidence proportions of falls were pooled for ambulators and wheelchair users separately using random-effects meta-analyses, and compared descriptively for inpatients and community-dwelling individuals. Fall-related factors were organized according to the Biological, Behavioral, Social & Economic and Environmental Model of fall risk. RESULTS: The search resulted in 1706 articles; 24 unique studies were included. The risk of sampling bias was high. All but one study focused on community-dwelling individuals; 78% (95% confidence interval 73-83%, I2 = 0%) of ambulators and 69% (95% confidence interval 60-76%, I2 = 59%) of wheelchair users fell ≥1 over 12 months. In contrast, only 13% of inpatients fell. Most fall-related factors were categorized as biological (e.g., muscle weakness), behavioral (e.g., inattentiveness) or environmental (e.g., uneven surfaces). CONCLUSIONS: Falls are frequent among community-dwelling individuals with SCI. A variety of biological, behavioral, and environmental factors contribute to falls, some of which are modifiable and may be addressed through SCI-specific fall prevention initiatives.

Validating Accelerometry as a Measure of Arm Movement for Children With Hemiplegic Cerebral Palsy.

BACKGROUND: For children with hemiplegic cerebral palsy (HCP), rehabilitation aims to increase movement of the affected arm. However, no validated measure objectively examines this construct in pediatric practice or daily life. OBJECTIVE: The objective of this study was to evaluate the criterion and known-groups validity of accelerometry as a measure of arm movement in children and adolescents with HCP. DESIGN: This was a prospective cross-sectional study. METHODS: Twenty-seven children and adolescents with typical development (3.4-13.9 years old) and 11 children and adolescents with HCP (4.7-14.7 years old; Manual Ability Classification System rating I or II) wore accelerometers on their wrists while engaged in 20 minutes of play, which included intermittent intervals of stillness and vigorous movement of the arms. Vector magnitude (VM) values identified the presence (VM > 2.0 counts per epoch) and absence (VM ≤ 2.0 counts per epoch) of arm movement for every 2-second epoch. Video was simultaneously recorded; each 2-second interval of footage was scored as "movement" or "no movement" for each arm. RESULTS: Agreement between accelerometry and video observation was greater than or equal to 81%, and the prevalence-adjusted and bias-adjusted κ value was greater than or equal to 0.69 for both groups of participants; these results supported the criterion validity of accelerometry. The ratio of nondominant arm movement to dominant arm movement measured by accelerometry was significantly greater in participants with typical development (mean [SD] = 0.87 [0.09]) than in participants with HCP (mean = 0.78 [0.07]) on the basis of 10 age- and sex-matched pairs; these results supported known-groups validity. LIMITATIONS: The small sample size of the group with HCP prevented the stratification of data by age. Participants with HCP had high or moderately high function of the affected arm; hence, the findings do not apply to children and adolescents with more significant hemiparesis. CONCLUSIONS: Accelerometry is a valid measure of arm movement in children with HCP and children without HCP. These findings contribute to the development of innovative upper limb assessments for children with hemiparesis.

Influence of upper limb movement patterns on accelerometer measurements: a pediatric case series.

OBJECTIVE: Previous studies showed success using wrist-worn accelerometers to monitor upper-limb activity in adults and children with hemiparesis. However, a knowledge gap exists regarding which specific joint movements are reflected in accelerometry readings. We conducted a case series intended to enrich data interpretation by characterizing the influence of different pediatric upper-limb movements on accelerometry data. APPROACH: The study recruited six typically developing children and five children with hemiparetic cerebral palsy. The participants performed unilateral and bilateral activities, and their upper limb movements were measured with wrist-worn accelerometers and the Microsoft Kinect, a markerless motion-capture system that tracks skeletal data. The Kinect data were used to quantify specific upper limb movements through joint angle calculations (trunk, shoulder, elbow and wrist). Correlation coefficients (r) were calculated to quantify the influence of individual joint movements on accelerometry data. Regression analyses were performed to examine multi-joint patterns and explain variability across different activities and participants. MAIN RESULTS: Single-joint correlation results suggest that pediatric wrist-worn accelerometry data are not biased to particular individual joint movements. Rather, the accelerometry data could best be explained by the movements of the joints with the most functional relevance to the performed activity. SIGNIFICANCE: This case series provides deeper insight into the interpretation of wrist-worn accelerometry data, and supports the use of this tool in quantifying functional upper-limb movements in pediatric populations.

The Feasibility of Functional Electrical Stimulation to Improve Upper Extremity Function in a Two-year-old Child with Perinatal Stroke: A Case Report.

AIMS: To evaluate the effectiveness and feasibility (i.e. tolerability, adherence) of functional electrical stimulation (FES) for the upper extremity (UE) in a two-year-old child with perinatal stroke. METHODS: Forty hours of FES over eight weeks was prescribed. FES to the hemiplegic triceps, extensor carpi radialis longus and brevis, extensor carpi ulnaris and extensor digitorum was timed with reaching during play. Assessments were performed before, during, and two months post-intervention. UE function (Melbourne Assessment 2 (MA2), Assisting Hand Assessment (AHA)) and spasticity (Modified Tardieu with electrogoniometry and electromyography) were measured. The mother completed a semi-structured interview post-intervention. Descriptive statistics were used for adherence and UE measures. A repeated-measures ANOVA compared Modified Tardieu parameters (e.g. catch angle) over time. Conventional content analysis was used for the interview data. RESULTS: The child completed 39.2/40 hours. Immediately post-intervention, improvements were observed on MA2's Range of Motion subscale and catch angle (Modified Tardieu, p < 0.001). Two months post-intervention, improvements were observed on MA2's Accuracy and Fluency subscales. No change in AHA score occurred. Three themes emerged from the interview: (1) Ingredients for program success; (2) Information about the FES device; and (3) The child's response. CONCLUSIONS: UE FES was feasible in a two-year-old child with hemiplegia.