USP E-Portfolio
Research Projects
Click on project descriptions to be led to current projects.
Publications
See recent publications and links to related websites.
July 2025
Video analysis reveals early signs of Bradykinesia in REM sleep behavior disorder and Parkinson’s disease
Research Credits
February 2025
VisionMD: an open-source tool for video-based analysis of motor function in movement disorders
Analysis of movement velocity feature in fingertapping in populations of healthy, RBD, and PD using video-based assessment.
PI & Department
- Diego Guarín, Ph.D.
- College of Health and Human Performance, Applied Physiology & Kinesiology Department
Example of Data Output
This is a Paragraph. Click on "Edit Text" or double click on the text box to start editing the content.
Start Date
June 2024
Paper published February 2025
Project Description & Progression
The Movement Estimation & Assessment Laboratory has been working towards creating an Artificial Intelligence tool that can pull objective information from frontal plane videos of subjects performing tasks such as opening and closing the palm, finger tapping, speaking, etc. The purpose of this research is to have a baseline for Parkinsonian symptoms that may be too small for physicians rating with the UPDRS-3 scale to notice.
My specific project is focusing on analyzing a class of sequential tasks that expose bradykinesia. These tasks include finger tapping, hand movement, leg agility, and toe tapping. Processing the data and interpreting the data is a collective lab effort. Recognizing significant differences, through statistical formulas, between the RBD and PD population shows that one population can decline into another. As the undergrad on this project I was tasked with data entry, data interpretation, and paper/poster writing. "VisionMD: An Open-Source Tool for Video-Based Analysis of Motor Function in Movement Disorders" has been accepted for publication in npj Parkinson's Disease. There is also a public website with the paper and tutorials on the tool (visionmd.ai). Poster presentations on the tool and using it to compare velocity movement features of healthy control, idiopathic rem sleep behavior disorder (iRBD), and Parkinson's Disease (PD) were completed at 2024 HHP AI days, 2024 UF AI Days, and 2025 Spring Undergraduate Research Symposium.
Tracking progression of motor task performance in atypical parkinsonian diseases and comparing to PD.
PI & Department
- Diego Guarín, Ph.D.
- College of Health and Human Performance, Applied Physiology & Kinesiology Department
Example of Data Output
s
Start Date
August 2025
Project Description & Progression
Using frontal view videos of UPDRS scored patients and their possible diagnosis's, ranging from possible Parkinson's Disease (PD) Multiple System Atrophy (MSA), and Progressive Supranuclear Palsy (PSP), run through VisionMD to identify different motor attributes between the diseased groups. There are more than 100 subjects which makes this study promising. Our lab presented this work at a poster presentation for UF AI Days 2025 named "Differentiation of Atypical Parkinsonism and Parkinson’s Disease utilizing motor task analysis with VisionMD" and found that there were significant differences between PSP and MSA in movement frequency.
Currently, I have processed 20 subjects with probable MSA and PSP performing the hand movement task found in the UPDRS 3. At the same time, 20 subjects of PD patients performing the same task were processed. These findings were compared and are being cumulated in a poster presentation for D.K. Stanley Day 2026 in the Health and Human Performance Department at UF. The poster is titled "Differentiating Atypical Parkinsonism and Parkinson's Disease Utilizing Frontal Video Based Analysis of Motor Assessments" and its abstract is as follows:
"Parkinsonism comprises idiopathic Parkinson’s disease (PD) and atypical syndromes such as progressive supranuclear palsy (PSP) and multiple system atrophy (MSA). Early differentiation is challenging due to overlapping motor feature decline and highly variable rates of progression. Patients with atypical parkinsonism often exhibit more rapid functional decline, highlighting the need for objective markers that support earlier diagnosis. VisionMD is an open-source platform designed to generate quantitative metrics of motor performance in multiple tasks such as gait, hand movement, and finger tapping. This tool has been used to identify significant differences between finger tapping of diseased groups, but the disease duration was not taken into consideration.
This poster presents the application of frontal video–based motion analysis to evaluate hand movement and finger tapping across three disease groups (MSA, PD, and PSP) with comparable disease duration. The population of subjects have disease durations that range from 2-3 years for all three groups. By identifying objective motor differences, this work aims to contribute to earlier and more accurate recognition of atypical parkinsonian syndromes. Statistical analysis results demonstrate a significant difference in finger tapping amongst the three groups in amplitude decay, mean opening velocity, and mean speed. Further finger tapping analysis, demonstrates significant differences between MSA and PSP in opening speed and task cycle duration."
The goal of this project has evolved to validate the usage of VisionMD to differentiate between atypical parkinsonism and Parkinson's Disease.