Automated nystagmus detection: Accuracy of slow-phase and quick-phase algorithms to determine the presence of nystagmus |
Ariel A. Winnick,
Chih-Chung Chen, Tzu-Pu Chang , Yu-Hung Kuo , Ching-Fu Wang , Chin-Hsun Huang , Chun-Chen Yang |
This study investigates the accuracy of automated nystagmus detection using video-oculography (VOG) and machine learning algorithms. It evaluated individual parameters, such as slow-phase velocity, SP duration ratio, saccadic difference, and saccadic ratio, which showed good diagnostic performance with AUCs between 0.791 and 0.896, and sensitivities and specificities around 70–88%. Combining these features through logistic regression models further improved accuracy, with the best models achieving AUCs over 0.91. Despite promising results, limitations include data collected from a single device, specific testing conditions, and a low prevalence of nystagmus, which affected positive predictive values. The study highlights the potential of integrating multiple parameters and machine learning to enhance automated nystagmus detection, suggesting promising directions for future research and clinical application in vestibular disorder diagnosis.
|
The fixation suppression test can uncover vertical nystagmus of central origin in some patients with dizziness |
Anand K Bery,
Ching-Fu Wang, Daniel R Gold, Tzu-Pu Chang |
The study shows that vertical nystagmus due to central vestibular issues is often subtle, weak, and suppressed by fixation, making it hard to detect clinically. Removing fixation during testing reveals these signs, with low SPV (around 2°/s), enhancing diagnostic accuracy. Detecting such nystagmus early can improve diagnosis of central causes like stroke or cerebellar disease. This highlights the need for advanced VNG devices capable of automated fixation suppression and precise low-velocity nystagmus detection. Software using AI for real-time analysis, training tools, and portable solutions can help clinicians identify subtle signs effectively. Telemedicine-compatible devices can enable remote assessments, broadening access and improving early diagnosis.
|
Prelimanry data on a novel smart glasses system for measuring the angle of deviation in strabismus |
Lung-Chi Lee,
Kathy Ming Feng, Pei-Chi Chuang, Yi-Hao Chen & Ke-Hung Chien |
This study introduces the NeuroSpeed (NSS) smart glasses system for measuring strabismus, comparing its automated eye movement recordings with traditional PACT in 70 patients. Quantitative results showed a strong correlation (r=0.969) and minimal bias, indicating reliable measurements for deviations under 40 prism diopters. Qualitatively, the device offers advantages like a comfortable, head-mounted design, quick analysis, and the ability to review recorded videos—streamlining and objectifying evaluations. Industry-wise, NSS advances ophthalmic diagnostics by providing a portable, automated tool that enhances accuracy, reduces examiner workload, and supports broader clinical and remote applications as it continues to improve.
|
Clinical Spectrum of Positional Vertigo in an Outpatient Setting |
Chih-Chung Chen,
Chen-Yu Wang, Po-Yueh Chen, Mei-Chien Chen, Ting-Yi Lee, Hsun-Hua Lee |
This study analyzed over 2,400 outpatient cases to explore the clinical spectrum of positional vertigo (PV). It found that PV is the most reported dizziness pattern, primarily caused by benign paroxysmal positional vertigo (BPPV) and vestibular migraine (VM). Notably, VM frequently leads to PV without typical positional nystagmus (PN) or any detectable PN, presenting significant diagnostic challenges. The findings emphasize that managing PV without PN is complex due to diagnostic uncertainty, highlighting the value of structured patient questionnaires to improve accuracy. Industry contributions include the development and validation of a patient-reported dizziness questionnaire (PREDIQT), which streamlines history-taking and enhances early diagnosis, supporting clinicians in differentiating vestibular disorders more efficiently.
|
Trends and Challenges of Wearable Multimodal Technologies for Stroke Risk Prediction |
Yun-Hsuan Chen,
Mohamad Sawan |
The Neurobit system, as discussed in the paper, exemplifies the potential of wearable multimodal devices that combine EEG and fNIRS for stroke risk prediction. Qualitatively, it offers a lightweight, user-friendly solution capable of continuous monitoring, which could be integrated into IoT networks to facilitate real-time assessments. Quantitatively, the system's effectiveness is supported by its ability to capture diverse physiological signals that enhance prediction accuracy beyond traditional measures. Industry contributions include advancing wearable neurovascular monitoring technologies, which could lead to more accessible, early detection systems that significantly improve stroke management and prevention strategies.
|