To evaluate the association between finger dexterity and handwriting proficiency in schoolchildren, and to describe the distribution of pencil grasp patterns during a standardized copy task. This school-based cross-sectional analytical study (June-August 2025) used complete enumeration of students in Grades 3-5. After parental consent, teachers completed the Handwriting Proficiency Screening Questionnaire (HPSQ); students performed a 2-3-line copy task in English for pencil-grasp classification and underwent O'Connor finger dexterity testing under standardized classroom conditions. Among 500 students, 170 (34.0%) had poor handwriting (HPSQ ≥ 14). Age, grade, and sex were comparable between the poor and good handwriting groups. Dexterity was markedly worse in the poor handwriting group [O'Connor finger time (s): 388 (71.5) vs. 309.7 (59.9); P < 0.001]. Finger time moderately correlated with HPSQ scores (r = 0.466). Grasp patterns differed between groups (P < 0.001); dynamic tripod was less common in the poor handwriting group (33.5 vs. 50.0%), and non-mature grasps had higher odds of poor handwriting (OR 2.01, 95% CI 1.38-2.93). Poor legibility and performance-time subscores were higher with non-mature grasps and increased with slower finger times. In multivariable analysis, slower finger dexterity and higher grade independently predicted greater difficulty. One-third of students had poor handwriting. Slower O'Connor finger dexterity independently predicted teacher-rated difficulty, while non-mature grasp patterns, though more prevalent in poor handwriting group, were not independently predictive. Findings support school screening that pairs HPSQ with a brief dexterity test, with interventions targeting fine motor control rather than grasp retraining.
The assessment of handwriting is fundamental for identifying difficulties, which may have long-term negative consequences. However, standard evaluation typically focuses only on the final handwritten product. For this reason, Italian guidelines recommended supporting traditional evaluation with digital tools to also analyze the handwriting process. A sensorized ink pen used on paper was employed by over 700 students, ranging from first grade in Italian primary school to third grade in lower secondary school, to execute two tasks of the BVSCO-3, the gold standard for handwriting assessment. From sensorized ink pen data, handwriting indicators in the domains of Time, Force, Smoothness, Tilt, and Frequency were extracted. These indicators were then analyzed to examine their correlation with clinical scores, to model cross-sectional trends across grades, and to identify handwriting difficulties. The correlation analysis revealed significant relationships between the indicators and the clinical score, particularly for the Time domain. A cross-sectional statistical analysis showed that the indicators follow developmental trends compatible with handwriting learning curves reported in the literature: for many indicators, a performance plateau was reached in grade 3, from both motor and processing perspectives. Lastly, binary classification models successfully distinguished subjects with handwriting difficulties (based on BVSCO-3 results) from proficient writers. The sensorized ink pen allowed uncovering relevant characteristics of children's handwriting process, while guaranteeing ecological data acquisition conditions. Its use could pave the way for a prompt identification of handwriting difficulties in school settings, thus facilitating an efficient referral to clinical services.
Handwriting is a hierarchical cognitive-motor activity requiring the integration of motor execution, visuospatial processing, working memory, and executive control. Digital handwriting technology enables simultaneous assessment of process (kinematics) and product (performance outcomes), offering a theoretically grounded approach to detecting cognitive vulnerability in aging. This study examined whether kinematic handwriting features differentiate institutionalized older adults with and without cognitive impairment and whether these features predict handwriting product performance under varying cognitive-motor demands. Fifty-eight participants (20 cognitively healthy; 38 cognitively impaired), classified using education-adjusted MMSE cutoffs, completed pen-control tasks (DOTS, LINES) and four handwriting-speed tasks (two copy, two dictation) on a digitizing tablet. Nine standardized kinematic variables were analyzed using logistic and multiple linear regression models with correction for multiple comparisons. Pen-control tasks (DOTS, LINES) did not significantly discriminate between the two groups, the handwriting-speed tasks, particularly dictation, revealed significant group differences. Temporal efficiency and stroke organization variables (e.g., Duration, Number of Strokes) significantly contributed to classification in high-demand tasks. Among cognitively healthy participants, associations between kinematic and product measures were limited, suggesting preserved compensatory mechanisms. Conversely, cognitively impaired individuals exhibited stronger process-product coupling, with Start Time, Vertical Size, and Duration significantly predicting handwriting performance in dictation tasks. Handwriting kinematics, especially temporal and stroke-related features, are sensitive indicators of cognitive impairment when assessed under high cognitive-motor load. These findings support the use of digitally mediated handwriting tasks-particularly dictation paradigms-as ecologically valid, low-cost tools for screening and monitoring cognitive decline in older adults. ClinicalTrials.gov, NCT06483438.
Neurodegenerative diseases such as Parkinson's disease (PD) and Alzheimer's disease (AD) present a significant and growing challenge to the healthcare systems worldwide. Both conditions are progressive and often undetected early, making timely diagnosis crucial. Recently, breakthroughs in computer vision and artificial intelligence have enabled the development of non-invasive and cost-effective screening and decision-making tools, allowing for earlier detection of the disease. This review serves as a comprehensive guide, providing structured insights into computational research methods for automated detection of PD and AD, with focus on handwriting analysis as a subtle behavioral biomarker of neurological impairment. A range of methodologies is examined, including static and dynamic handwriting assessment, feature engineering procedures, deep learning and classical ML-based approaches. The analysis emphasizes the most effective methods, the handwriting features found to be most revealing, the datasets most used in the literature, and the performance levels reported for each disease. Several studies report that transfer learning based on convolutional neural networks and transformer-based architectures for Parkinson's disease diagnosis is regularly able to achieve high accuracy, frequently above 95% on benchmark datasets. In contrast, Alzheimer's disease research is progressively benefitting from multimodal approaches combining kinematic and spatial handwriting features to capture cognitive and motor changes. Structured summaries of publicly available handwriting datasets are provided, and critical advancements, ongoing challenges, and future research priorities are discussed. The integration of insights across the studies, through this work, aims to assist researchers and clinicians in the development and translation of handwriting-based, AI-guided diagnostic tools for neurodegenerative diseases.
Extrapyramidal symptoms (EPS) are motor side effects commonly induced by antipsychotic medications and can lead to measurable changes in handwriting patterns. These symptoms affect both the spatial and temporal characteristics of writing, including stroke thickness, direction and the rate of directional change. To model these complex variations, we propose a novel Liquid Generative Adversarial Network (LiquidGAN), which combines the adaptive dynamics of liquid neural networks with the data generation capability of GANs. Handwriting data were collected from 94 patients with confirmed EPS and 30 healthy controls using Archimedean spiral patterns drawn with both hands. A total of 211 images were processed for both binary and multiclass classification using a pretrained ResNet50 model. The pretrained ResNet50 achieved 92% accuracy and 97% precision in the binary classification task; however, its performance dropped significantly to 57% accuracy in multiclass classification, indicating limited capability in capturing fine-grained EPS severity variations. In contrast, the proposed LiquidGAN demonstrated excellent performance in the binary classification task, achieving 97% accuracy and 98% precision. More importantly, LiquidGAN substantially outperformed the baseline in the more challenging multiclass setting, achieving 70% accuracy and precision across four classes (mild, moderate, severe, and control). This shows that the diverse dataset from the liquidGAN significantly improves the HOG-ANN classification and effectively captures complex and subtle handwriting variations associated with different EPS severity levels that conventional models such as ResNet50 fail to distinguish. In addition, LiquidGAN generated diverse and realistic synthetic handwriting samples, yielding improved Fréchet Inception Distance (FID), precision, and recall compared with style GAN. These findings demonstrate that handwriting biomarkers, when analyzed through dynamic generative learning, offer an effective and non-invasive approach for monitoring extrapyramidal side effects in clinical settings.
Chinese language practice, especially handwriting and typewriting practice, has always been a key method for mastering Chinese in Chinese L2 learning. However, current research on blended practice modalities that combine handwriting and typewriting remains insufficient. This study used a pilot study with 30 international students to compare the associative patterns of handwriting practice and blended practice (a sequential multimodal practice wherein handwriting instruction was followed by typewriting) on students' Chinese skills performance, Chinese learning motivation, and attitude. Results indicated that students using the blended practice were associated with significantly better Chinese skills performance, as well as higher levels of motivation and more positive attitudes compared to those using only handwriting. Exploratory path analysis identified a notable direct association between practice modality and Chinese skills performance; however, the pathways through motivation and attitude were not statistically detectable. These findings suggest that Chinese language teachers may consider designing lessons that incorporate this sequential blended practice, which may support improvements in L2 Chinese performance, motivation, and attitudes. Furthermore, second language learners should actively apply this practice modality to improve their Chinese L2 learning performance.
Parkinson's disease (PD) is a significant mental health condition, and patients greatly benefit from prompt diagnosis and treatment if the disease is identified early. One powerful method for diagnosing PD at an early stage is the analysis of hand-drawing and handwriting samples from individuals with PD. The novelty of this research lies in developing a handwriting and hand-drawing Parkinson's Disease (HwdPD) framework for detecting PD. This framework utilizes hand-drawing and Arabic handwriting samples, which have been observed to be effective in detecting PD. Based on VGG19 and transformer, the proposed framework was tested using a real standard dataset containing 63 hand-drawn images, namely spiral, wave, and ellipse samples, with 30 samples from PD and 33 from healthy patients. The dataset also contains Arabic handwriting samples, namely "eight" and hello ("لو"). These images were processed by using augmentation to enhance the performance of the HwdPD framework. This enhancement of the image area was fed to the classification algorithm (transformer ViT-B16 and VGG19). ViT-B16 scored high accuracy, with 100% in spiral, wave, and ellipse images. In handwriting samples ("eight"), the system successfully achieved a high percentage of 100%. This system emphasizes the potential to improve diagnostic accuracy and assist clinical decision-making by evaluating its performance on these datasets. The HwdPD framework demonstrates potential for identifying PD biomarkers, which may lead to improved diagnostics.
The decoding of latent neural states from observable signals is a key focus of modern brain-AI research. Although most neural decoding models are based on electrophysiological recordings, peripheral motor outputs also convey information about circuit-level dynamics. Handwriting is a channeled behavioral signal that indexes the health of the cortico-basal ganglia-thalamo-cortical loop. In Parkinson's disease (PD), dopaminergic loss disrupts this circuit, leading to tremor oscillations, micrographia, and movement irregularities. The challenge of decoding behavior-encoded neural signals from handwriting images constitutes a principled neural signal interpretation problem. We propose a physics-informed and interpretable AI framework for decoding basal ganglia motor dysfunction through harmonic oscillator perturbation analysis. Six energy-inspired measures are extracted to quantify different aspects of motor system dynamics: intensity variation, spatial gradients, multi-scale stability, deviation variability, directional anisotropy, and cross-scale interactions. The transparent mapping between computational models and neurophysiological processes is enabled by these steps, which have their basis in mechanistic theories of amplitude modulation and oscillatory instability. High levels of discrimination were achieved when 594 handwriting trials (279 with PD and 315 controls) were assessed using repeated 10-fold cross-validation for spiral, circle, and meander tasks. Support Vector Machines achieved 84.06% accuracy and 93.56% sensitivity, with highly significant group differences across all features (p < 10-33; Cohen's |d| = 0.87-1.51). Through the integration of physics-based modeling and interpretable machine learning, the proposed framework extends neural signal interpretation beyond direct neural recordings, establishing handwriting as a low-cost, behaviorally encoded biomarker of circuit state and advancing AI-driven decoding of brain dysfunction.
The digital transition in school settings is reshaping children's learning, writing practices, and mental health trajectories. This narrative review examines whether handwriting still matters in contemporary hybrid educational environments from a multidisciplinary perspective. Evidence from neuroscience, developmental psychology, educational sciences, pediatrics, and child psychiatry was narratively synthesized, with attention to handwriting, digital exposure, learning, emotional regulation, and vulnerable populations. Handwriting uniquely integrates fine motor control, visuomotor coordination, orthographic processing, attention, and embodied cognition, supporting early literacy, memory consolidation, conceptual learning, and reflective writing. Conversely, excessive or poorly mediated digital exposure may interact with attentional fragmentation, sleep disruption, online stressors, problematic use, and internalizing symptoms, particularly in vulnerable children and adolescents. Digital tools remain essential for personalization, accessibility, and compensatory support, especially for students with neurodevelopmental or learning difficulties. Handwriting and digital technologies should not be framed as competing educational paradigms. A developmentally sensitive hybrid model is needed, preserving handwriting during key stages of literacy and self-regulation while integrating digital tools as purposeful, individualized resources for learning, inclusion, and school mental health promotion.
Parkinson's disease (PD) affects fine motor control and produces measurable abnormalities in handwriting and drawing. This study proposes a rigorously evaluated multimodal framework for PD detection that combines a Vision Transformer (ViT) for spiral and meander image analysis with an XGBoost classifier operating on 54 carefully engineered kinematic features extracted from multichannel handwriting signals. To assess how multimodal integration should be performed, both intermediate feature-level fusion and late decision-level fusion were evaluated under a strict 5-fold subject-wise cross-validation protocol, with supplementary sample-level analysis, on the NewHandPD dataset. The visual stream consistently outperformed the acquisition stream as a single modality, while both fusion strategies improved performance by exploiting complementary spatial (visual) and motor information. Intermediate fusion achieved the highest apparent discriminative performance, reaching 97.7% accuracy on spiral drawings and 98.5% accuracy on meander drawings, whereas late fusion provided more interpretable and modular behavior, with best subject-level results of 93.94% accuracy and AUC = 0.9687 for spiral, and 92.42% accuracy with AUC = 0.9770 for meander. These findings suggest that multimodal handwriting analysis can be an effective approach for Parkinson's disease detection on the NewHandPD dataset, although further validation on larger and independent cohorts is required.
Background: Carpal tunnel release (CTR) surgery may have an impact on the speed and accuracy of handwriting and digital typing using a computer or mobile devices. Methods: In this prospective cohort study, patients undergoing CTR surgery of the dominant hand, between 18 and 70 years of age, with frequent use of a QWERTY keyboard and smartphone were included. A baseline Visual Analog Scale (VAS) for subjective alteration of writing and Quick Disabilities of the Arm, Shoulder and Hand (QuickDASH) were collected at a maximum 3-month follow-up. Digital typing speed (word per minute, wpm) and accuracy (number of mistakes) were tested. Handwriting was assessed by means of direct supervision of one investigator. Results: Of the 30 enrolled patients, 23 (76.7%) completed the 3-month follow-up and were included in the final analysis (mean age 55 ± 12.3 years). Pre- and post-surgery improvements in keyboard typing speed (14.8 ± 6.8 wpm to 17.6 ± 5 wpm) and mobile texting speed (16.7 ± 5.9 wpm to 21.7 ± 6.5 wpm) were significantly improved over time. Accuracy improved significantly only in keyboard typing, where the mean number of errors was reduced (13.1 ± 8.2 to 9.9 ± 5.6). QuickDASH scores decreased significantly (39.1 ± 9.1 to 17 ± 6). Conclusions: CTR surgery was associated with improved typing speed and reduced the number of errors (19% and -24%, respectively) as well as texting speed (30%). This improvement may be relevant in daily and occupational activities, as reported in the previous literature.
BackgroundAI chatbots are increasingly used in language education, but their adoption for pronunciation and handwriting instruction remains underexplored. This study examined factors influencing Chinese K-12 foreign language teachers' adoption of AI chatbots for these perceptual-motor teaching tasks. MethodsSurvey data were collected from 615 teachers, with 526 valid responses analyzed using confirmatory factor analysis and structural equation modeling.ResultsTeachers mainly used AI chatbots for lesson planning and assignment design, while direct use for pronunciation and handwriting instruction was limited. Perceived ease of use positively predicted perceived usefulness, trust, self-efficacy, and behavioral intention. Perceived value strongly predicted perceived usefulness. Trust and self-efficacy predicted behavioral intention, which predicted actual use, whereas perceived usefulness had no significant direct effect on behavioral intention.DiscussionAdoption was influenced more by usability, trust, and implementation confidence than by perceived usefulness alone.ConclusionTeacher training should emphasize evaluating AI-generated feedback and translating chatbot outputs into reliable corrective guidance and repeated practice.
Handwritten bedside medication lists remain common in healthcare, especially in low-resource countries, presenting challenges for digitization and automated safety checks. The introduction of large-language models may present an opportunity to facilitate digitization of handwritten medication lists. This study evaluated the accuracy of three GPT-based language models in recognizing handwritten medication lists, presented in Dutch. Thirty-three participants transcribed a list of 10 medications with dosage and administration instructions. These lists were processed by each model and scored for correct medication name, dosage, frequency, and route. The effect of writer characteristics on LLM performance was assessed using the Mann-Whitney U test (sex, handedness, ink colour) and Kruskal-Wallis H-test (handwriting style: cursive, print, mixed). GPT 4.1 achieved the highest accuracy, followed by GPT4o, both outperforming GPT4o-mini (p < 0.001). Recognition strongly correlated with human legibility (ρ = 0.655; p < 0.001). Print handwriting and blue ink resulted in higher recognition than cursive or mixed styles and black ink. Complex dosing instructions were most error-prone. GPT-based optical character recognition showed potential for scalable digitization of handwritten prescriptions, however, human oversight remains essential to ensure medication safety. Future research should validate performance in real-world, multilingual settings.
With the rapid development of flexible electronics and human-computer interaction technologies, the demand for high-performance flexible strain sensors for human motion and physiological signal monitoring has grown sharply. This study developed a sandwich-structured flexible strain sensor based on multi-walled carbon nanotube/halloysite nanotube (MWCNT/HNT) nanocomposites. A polydopamine (PDA) modification layer was first formed on the polydimethylsiloxane (PDMS) film via in situ polymerization of dopamine (DA), which significantly improved the surface hydrophilicity and enhanced interfacial adhesion between PDMS and the subsequent conductive layer. The MWCNT/HNT mixed conductive layer was then uniformly sprayed onto the PDA@PDMS film, followed by PDMS encapsulation to form the final structure. The sensor integrates the excellent conductivity of MWCNTs and the mechanical reinforcement capability of HNTs, achieving a wide strain-sensing range, high sensitivity with GF values of 14.46 at 0-40% strain and 35.55 at 40-85% strain, fast response/recovery times of 250/300 ms, and excellent cyclic stability over 7500 cycles. It can sensitively detect physiological signals such as swallowing, pulse and breathing, as well as motion signals including joint bending and mouse clicking. Moreover, it can distinguish subtle finger movements during English letter writing, showing broad application potential in wearable healthcare devices and human-computer interaction systems.
Cellulose-based materials simultaneously exhibit degradability, biocompatibility, and cost-effectiveness, making them promising eco-friendly components for fabricating and advancing wearable electronics. However, irreversible slippage between cellulose fibers under cyclic loading poses a significant challenge to achieving stable conversion of mechanical stimuli into electrical signals. Herein, we developed a piezoresistive sensor based on cellulose handsheets using a screen-printing-inspired strategy. With the addition of 10 wt% multi-walled carbon nanotubes (MWCNTs), continuous percolating networks formed as robust mechanical scaffolds and reconfigurable conductive pathways, simultaneously enhancing the structural stability of cellulose handsheet and boosting its conductivity to 27.9 S/m. The resulting conductive cellulose handsheets were imprinted into microisland arrays as stress-intensifying architectures and orthogonally integrated to construct piezoresistive sensors. The synergistic effect of percolating networks and microisland arrays endowed the sensors with high sensitivity (58.1 kPa-1 within 0-5.6 kPa), excellent durability (>10,000 cycles), a wide workable pressure range (0-60 kPa), and fast response and recovery times (140/80 ms). Practical applications of the proposed sensors were demonstrated for real-time activity detection and health monitoring. Furthermore, with the assistance of a deep learning algorithm, the sensor achieved 78.8% recognition accuracy for handwritten multi-letter words, highlighting its potential to advance sustainable and intelligent human-machine interactions in wearable electronics.
暂无摘要（点击查看详情）
Handwritten text recognition (HTR) in examination scenarios has gained increasing attention for its role in intelligent grading systems. However, existing studies have not systematically modeled the complex handwriting phenomena inherent in exam settings, hindering a comprehensive understanding of the recognition challenges and limitations of current methods. Specifically, handwriting artifacts pose significant challenges to recognition models in two complementary aspects: sequentially, they disrupt the reading order and lead to non-monotonic sequences, while visually, they distort character structures and induce attention drift. To enable systematic benchmarking of exam handwriting, we first construct BNU-Exam-HTR, a large-scale dataset of handwritten exam text, and establish BNU-Exam-Benchmark, a fine-grained evaluation framework defining 12 representative challenges observed in real exam handwriting. To overcome these challenges, we further propose EduOCR, a recognition model with a collaborative dual-branch decoder. The Sequential Symbol Module (SSM) uses autoregressive decoding to handle non-monotonic sequences, while the Permutation-Aware Prediction Head (PPH) simulates artifact perturbations to guide the shared encoder in distinguishing characters from noise, thus stabilizing attention and mitigating alignment errors. Extensive experiments show that EduOCR consistently outperforms state-of-the-art HTR models, OCR tools, and multimodal large language models across all 12 challenges, demonstrating superior robustness and adaptability.
Handwriting impairment is a distinctive symptom of movement disorders (MDs), such as Parkinson's disease (PD) and Parkinsonian syndrome (PDS). Custom-designed digital handwriting (DHW) tools were developed to strengthen the diagnosis of MDs. Participants with MDs from the Neurology and PET departments at the First Affiliated Hospital of Guangzhou Medical University were recruited to perform digital handwriting (DHW) tasks, which included drawing a line and a cube, and writing a sentence in Chinese (Ch), English (E), and Korean (K). AV133 PET-CT/MRI was used to detect the expression of vesicular monoamine transporter 2 (VMAT2) in the brain. An attention-based one-dimensional convolutional neural network (1D-CNN) model was developed to evaluate DHW indicators for PD diagnosis. A total of 197 participants with MDs and 160 matched controls were included, among whom 55 MD patients underwent AV133 PET-CT/MR for VMAT2 quantification. The machine learning model demonstrated that the DHW scores showed excellent discrimination between MDs and controls, with an area under the receiver operating characteristic curve (AUC) of 0.982. In subgroup analyses, the L + E + K and L + E + K + Cu task combinations effectively distinguished PD from PDS, with AUC values of 0.727 and 0.717, respectively. The diagnostic agreement between the DHW and PET-VMAT2 results reached 85.45%. PET-AV133 imaging revealed a significant 33.7% downregulation of VMAT2 expression throughout the putamen in PD patients compared with PDS patients, with dramatic reductions in the posterior putamen. DHW scores from the L + E + K + Cu task combination were significantly correlated with VMAT2 expression in the caudate nucleus (q = 0.0005), specifically within the bilateral caudate heads and bodies. Custom-designed DHW tasks can serve as effective digital markers for PD diagnosis.
Micrographia is a symptom characterized by abnormally small handwriting and is a common motor symptom in Parkinson's disease. Two subtypes-consistent micrographia (CM) and progressive micrographia (PM)-are thought to reflect distinct underlying mechanisms. CM has been linked to bradykinesia and often improves with dopaminergic treatment, whereas PM typically persists and may involve altered sensorimotor processing. However, the relative contributions of motor execution, feedback integration, and motor awareness to these subtypes remain unclear. This study investigated whether awareness of motor control differs between people with Parkinson's disease (PwP) and healthy controls (HC), and whether such differences are associated with micrographia subtypes. Forty-five PwP and twenty age-matched HC completed a handwriting task and a control detection task (CDT), which assesses the ability to distinguish self-generated from externally generated movements. Based on handwriting metrics, PwP were classified as having no micrographia, CM, or PM. CDT accuracy was significantly lower in PwP than in HC, with particularly reduced performance in individuals with PM. These findings suggest that PM is associated with reduced motor awareness, potentially involving altered prediction-feedback integration, and may inform the development of therapeutic approaches that complement dopaminergic treatment.