Integrating Wearable Sensor Signal Processing with Unsupervised Learning Methods for Tremor Classification in Parkinson’s Disease

Bioengineering, Год журнала: 2025, Номер 12(1), С. 37 - 37

Опубликована: Янв. 6, 2025

Tremor is one of the most common symptoms Parkinson's disease (PD), assessed using clinician-assigned clinical scales, which can be subjective and prone to variability. This study evaluates potential unsupervised learning for classification assessment tremor severity from wearable sensor data. We analyzed 25 resting signals 24 participants (13 PD patients 11 controls), focusing on motion intensities derived accelerometer recordings. The k-means clustering algorithm was employed, achieving a accuracy 76% versus non-tremor states. However, performance decreased multiclass (57.1%) binary severe mild (71.4%), highlighting challenges in detecting subtle intensity variations. findings underscore utility enabling scalable, objective analysis. Integration such models into systems could improve continuous monitoring, enhance rehabilitation strategies, support standardized assessments. Future work should explore advanced algorithms, enriched feature sets, larger datasets robustness generalizability.

Язык: Английский

---

Computer Vision in Clinical Neurology

JAMA Neurology, Год журнала: 2025, Номер unknown

Опубликована: Фев. 17, 2025

Importance Neurological examinations traditionally rely on visual analysis of physical clinical signs, such as tremor, ataxia, or nystagmus. Contemporary score-based assessments aim to standardize and quantify these observations, but tools suffer from clinimetric limitations often fail capture subtle yet important aspects human movement. This poses a significant roadblock more precise personalized neurological care, which increasingly focuses early stages disease. Computer vision, branch artificial intelligence, has the potential address challenges by providing objective measures signs based solely video footage. Observations Recent studies highlight computer vision measure disease severity, discover novel biomarkers, characterize therapeutic outcomes in neurology with high accuracy granularity. may enable sensitive detection movement patterns that escape eye, aligning an emerging research focus stages. However, accessibility, ethics, validation need be addressed for widespread adoption. In particular, improvements usability algorithmic robustness are key priorities future developments. Conclusions Relevance technologies have revolutionize practice objective, quantitative signs. These could enhance diagnostic accuracy, improve treatment monitoring, democratize specialized care. Clinicians should aware their complement traditional assessment methods. further focusing validation, ethical considerations, practical implementation is necessary fully realize neurology.

Язык: Английский

---

Validity of tremor analysis using smartphone compatible computer vision frameworks

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Апрель 18, 2025

Abstract Computer vision (CV)-based approaches hold promising potential for the classification and quantitative assessment of movement disorders. To take full advantage this potential, pipelines need to be validated against established clinical electrophysiological gold standards. This study examines validity Mediapipe (by Google) Vision Apple) smartphone-enabled hand detection frameworks tremor analysis. Both were tested in virtual experiments with simulated tremulous hands determine optimal camera position minimum detectable amplitude frequency. then compared optical motion capture (OMC), accelerometry, ratings 20 patients. CV accurately measured peak Significant correlations found between CV-assessed amplitudes Essential Tremor Rating Assessment Scale (TETRAS) scores. However, accuracy estimation OMC as ground truth was insufficient application. In conclusion, CV-based analysis is an accurate simple tool Further improvements are needed.

Язык: Английский

---

Identification of smile events using automated facial expression recognition during the Autism Diagnostic Observation Schedule (ADOS-2): a proof-of-principle study

Frontiers in Psychiatry, Год журнала: 2025, Номер 16

Опубликована: Май 1, 2025

Introduction The diagnosis of autism spectrum disorder (ASD) is resource-intensive and associated with long waiting times. Digital screenings using facial expression recognition (FER) are a promising approach to accelerate the diagnostic process while increasing its sensitivity specificity. aim this study examine whether identification smile events FER in an utilisation population reliable. Methods From video recordings children undergoing Autism Diagnostic Observation Schedule (ADOS-2) due suspected ASD, sequences showing non-smile were identified. It being investigated reliably recognizes corresponds human rating. Results based on action unit mouthSmile accurately identifies 96.43% specificity 96.08%. A very high agreement raters (κ = 0.918) was achieved. Discussion This demonstrates that can principle be identified clinical autism. Further studies required generalise results.

Язык: Английский

---

Validation of computer vision technology for analyzing bradykinesia in outpatient clinic videos of people with Parkinson's disease

Journal of the Neurological Sciences, Год журнала: 2024, Номер 466, С. 123271 - 123271

Опубликована: Окт. 15, 2024

Язык: Английский

---

Hand Tremor Characterization from a Spatiotemporal Convolutional Representation

Ingeniería, Год журнала: 2024, Номер 29(3), С. e21091 - e21091

Опубликована: Ноя. 27, 2024

Context: Parkinson’s Disease (PD) is a neurodegenerative disorder related to dopamine deficiency that mainly entails motor conditions such as slowness of movement, postural instability, limb tremor, rigidity, and decreased range motion. Tremor, defined rhythmic uncontrolled movement limbs, the most prevalent symptom in PD. In clinical routine, tremors are assessed quantified by observing hands following resting patterns. These configurations include voluntary muscular contractions tremor perception reduction, which leads noisy signals. The assessments also subjective depend on expertise professionals determine whether associated with Method: This work introduces deep volumetric representation characterizes PD patterns recording conditions. strategy includes convolutional architecture extracts spatiotemporal correlated propagated through different layers until discrimination between control subjects achieved. Moreover, set explainability maps computed backpropagating output gradients into convolutionally learned spatio-temporal maps. Results: method was evaluated 80 videos (five patients five subjects), reporting an average accuracy 92.5% perfect sensitivity score configuration. As for scheme, proposed obtained 90% 80%. Conclusions: approach showed efficacy regarding localization patterns, recovering information while preserving spatial temporal representation. allows visualizing from patients.

Язык: Английский

---

Enhanced Computer Vision Technique for Differentiating Tremor Types

Опубликована: Окт. 23, 2024

Язык: Английский

---