Motor Assessment With the STEGA iPad App to Measure Handwriting in Children.

IMPORTANCE: Handwriting and the fine motor control (hand and fingers) underlying it are key indicators of numerous motor disorders, especially among children. However, current assessment methods are expensive, slow, and subjective, leading to a lack of knowledge about the relationship between handwriting and motor control. OBJECTIVE: To develop and validate the iPad precision drawing app Standardized Tracing Evaluation and Grapheme Assessment (STEGA) to enable rapid quantitative assessment of fine motor control and handwriting. DESIGN: Cross-sectional, single-arm observational study. SETTING: Academic research institution. PARTICIPANTS: Fifty-seven typically developing right-handed children ages 9 to 12 yr with knowledge of cursive. OUTCOMES AND MEASURES: Predicted quality, measured as the correlation between handwriting letter legibility (Evaluation Tool of Children's Handwriting-Cursive [ETCH-C]) and predicted legibility (calculated from STEGA's 120 Hz, nine-variable data). RESULTS: STEGA successfully predicted handwriting (r2 = .437, p < .001) using a support vector regression method. Angular error was the most important aspect of STEGA performance. STEGA was much faster to administer than the ETCH-C (M = 6.7 min, SD = 1.3, versus M = 19.7 min, SD = 5.2). CONCLUSIONS AND RELEVANCE: Assessment of motor control (and especially pen direction control) may provide a meaningful, objective way to assess handwriting. Future studies are needed to validate STEGA with a wider age range, but the initial results indicate that STEGA can provide the first rapid, quantitative, high-resolution, telehealth-capable assessment of the motor control that underpins handwriting. What This Article Adds: The ability to control pen direction may be the most important motor skill for successful handwriting. STEGA may provide the first criterion standard for the fine motor control skills that underpin handwriting, suitable for rehabilitation research and practice.

High-resolution HLA typing by long reads from the R10.3 Oxford nanopore flow cells.

Nanopore sequencing has been investigated as a rapid and cost-efficient option for HLA typing in recent years. Despite the lower raw read accuracy, encouraging typing accuracy has been reported, and long reads from the platform offer additional benefits of the improved phasing of distant variants. The newly released R10.3 flow cells are expected to provide higher read-level accuracy than previous chemistries. We examined the performance of R10.3 flow cells on the MinION device in HLA typing after enrichment of target genes by multiplexed PCR. We also aimed to mimic a 1-day workflow with 8-24 samples per sequencing run. A diverse collection of 102 unique samples were typed for HLA-A, -B, -C, -DPA1, -DPB1, -DQA1, -DQB1, -DRB1, -DRB3/4/5 loci. The concordance rates at 2-field and 3-field resolutions were 99.5% (1836 alleles) and 99.3% (1710 alleles). We also report important quality metrics from these sequencing runs. Continued research and independent validations are warranted to increase the robustness of nanopore-based HLA typing for broad clinical application.

Measuring Activities of Daily Living in Stroke Patients with Motion Machine Learning Algorithms: A Pilot Study.

Measuring activities of daily living (ADLs) using wearable technologies may offer higher precision and granularity than the current clinical assessments for patients after stroke. This study aimed to develop and determine the accuracy of detecting different ADLs using machine-learning (ML) algorithms and wearable sensors. Eleven post-stroke patients participated in this pilot study at an ADL Simulation Lab across two study visits. We collected blocks of repeated activity ("atomic" activity) performance data to train our ML algorithms during one visit. We evaluated our ML algorithms using independent semi-naturalistic activity data collected at a separate session. We tested Decision Tree, Random Forest, Support Vector Machine (SVM), and eXtreme Gradient Boosting (XGBoost) for model development. XGBoost was the best classification model. We achieved 82% accuracy based on ten ADL tasks. With a model including seven tasks, accuracy improved to 90%. ADL tasks included chopping food, vacuuming, sweeping, spreading jam or butter, folding laundry, eating, brushing teeth, taking off/putting on a shirt, wiping a cupboard, and buttoning a shirt. Results provide preliminary evidence that ADL functioning can be predicted with adequate accuracy using wearable sensors and ML. The use of external validation (independent training and testing data sets) and semi-naturalistic testing data is a major strength of the study and a step closer to the long-term goal of ADL monitoring in real-world settings. Further investigation is needed to improve the ADL prediction accuracy, increase the number of tasks monitored, and test the model outside of a laboratory setting.