| Field |
Value |
| NCT Number |
NCT06663826 |
| Status |
Recruiting |
| Phase |
Observational (Device Development) |
| Sponsor |
machineMD AG |
| Collaborators |
University of Zurich, University Hospital Zurich, University of Exeter, gaitQ Limited |
| Study Type |
Observational / Patient Registry |
| Target Enrollment |
100 patients |
| Study Duration |
15 days per patient |
| Locations |
University Hospital of Zurich, Switzerland |
Parkinson's disease (PD) is one of the most common neurodegenerative diseases worldwide, affecting approximately 1% of the population over 65 years of age 1. Current PD diagnosis relies heavily on:
- Patient history and subjective symptom descriptions
- Clinical assessments using the Movement Disorder Society (MDS) criteria
- Neurological examination including the MDS-UPDRS (Unified Parkinson's Disease Rating Scale)
- Short walking tests such as the 3-meter Timed Up and Go (TUG)
However, these approaches suffer from significant limitations:
- Subjectivity: Diagnosis depends on the patient's recall and the physician's interpretation
- Temporal variability: Symptoms fluctuate through and between days, making "snapshot" clinic assessments potentially unrepresentative
- Inability to track progression: Manual examinations lack precise instrumentation to accurately monitor disease progression
- Early detection gap: Current criteria cannot reliably detect subclinical, early-stage signs
Several oculo-visual abnormalities have been described in PD. Research indicates that abnormal ocular motor function occurs in 75-87.5% of people with PD 2. These dysfunctions may:
- Precede motor symptoms
- Follow motor symptoms
- Provide valuable information for early disease detection
- Serve as biomarkers for disease progression
The most commonly reported ocular motor dysfunctions in PD include:
- Saccadic impairments: Abnormalities in rapid eye movements
- Smooth pursuit deficits: Difficulty tracking moving objects
- Vergence dysfunction: Problems with focusing on near vs. far objects
Gait impairments are among the most common and disabling symptoms of PD 3, including:
- Freezing of gait (FOG): An inability to initiate or maintain normal walking patterns
- Festinating gait (FSG): Shortening of stride length with elevated step frequency
Both FOG and FSG contribute to an increased risk of falls and fall-related injuries. Objective, continuous remote gait monitoring would enable:
- Real-time tracking of gait impairments
- Objective disease progression monitoring
- Personalized care delivery
Collect ocular motor, pupil, and gait data from people with PD to develop and compare machine learning models for diagnosing and monitoring PD.
-
Clinical correlation: Correlate ocular motor, pupil, and gait parameters with clinical parameters including:
- Disease stage
- Disease duration
- Age of onset
- Medication status
- MDS-UPDRS score
-
Real-world evidence: Collect RWE data regarding health economics parameters to assess the properties, effects, and impacts of deployed health technologies.
-
Scientific advancement: Contribute to the scientific understanding of PD, potentially uncovering new insights into disease patterns, progression, and treatment response.
This is an exploratory open-label single-centre research project. Each patient will undergo:
| Visit |
Duration |
Assessments |
| Visit 1 |
3 hours |
MDS-UPDRS, neos examination, standard ocular motor/pupil exam, gait assessment |
| Visit 2 (after 2 weeks) |
2 hours |
Repeat assessments |
| Home monitoring (2 weeks) |
Daily |
TUG test (15m walk, 5 sit-to-stand, 5-minute walk) |
The neos device is a medical device approved for objective ocular motor and pupil measurement. It provides:
- High-precision eye tracking
- Pupil response measurements
- Standardized oculomotor assessments
- Quantitative data for machine learning
A consumer device enabling objective and continuous remote gait monitoring, consisting of:
- IMU (Inertial Measurement Unit) sensor placed on the patient's back
- Wearable leg sensor for gait analysis
The machine learning algorithms will be trained on a clinical dataset comprising:
- 50 PD patients
- Healthy individuals (data from another study)
- 12 patients with other parkinsonian disorders (atypical parkinsonism)
Data types collected:
- Ocular motor data from neos
- Ocular motor and pupil data from standard clinical examination
- Gait data from GaitQ senti (leg sensor)
- Gait data from IMU sensor (back)
- Demographic information: age, sex, ethnicity, eye color
- Clinical information: disease stage, disease duration, age of onset, medication, MDS-UPDRS score
- Diagnosis of Parkinson's disease or another parkinsonian syndrome (atypical Parkinson's)
- Refractive error between -6 and +4 diopters (both eyes)
- Informed consent documented per signature
- Able to self-report history of daily gait freezing and/or festination
- Able to walk unsupported or using an aid for at least 5 minutes
- Other known neurological diseases
- Current medication/drugs that could influence ocular motor tasks (e.g., benzodiazepines, alcohol, stimulants, recreational drugs) — except Parkinson's medications
- Incapacity to understand and comply with the examination (e.g., advanced cognitive decline)
- Any injury or disorder that may affect eye movement measurements or balance (other than PD)
- Any skin conditions or broken skin in the calf and behind the knee area
- Lack of access to WiFi for home monitoring
| Measure |
Description |
Timeframe |
| Machine learning model development |
Train algorithms using clinical dataset for PD diagnosis and monitoring |
1 year |
| Measure |
Description |
Timeframe |
| Correlation with clinical parameters |
Correlation between ocular motor parameters and disease stage, duration, onset age, medication, MDS-UPDRS |
1 year |
This trial represents a significant step toward objective, quantitative PD diagnosis by:
- Replacing subjectivity with precision: Using high-quality sensor data instead of subjective clinical interpretation
- Enabling earlier detection: Quantitative ocular motor and gait parameters may identify PD before overt motor symptoms
- Improving disease monitoring: Continuous remote monitoring provides longitudinal data beyond clinic visits
- Personalizing care: Objective measurements enable tailored treatment strategies
By integrating machine learning with high-quality sensor data, this approach aims to:
- Improve diagnostic accuracy beyond current clinical criteria
- Enable earlier PD detection
- Provide objective disease progression tracking
- Support personalized treatment decisions
| Investigator |
Role |
Affiliation |
| Konrad Weber, Prof. Dr. med. |
Principal Investigator |
University of Zurich |
| Ana Coito, Ph.D. |
Contact |
machineMD AG |
| Pia Massatsch, Ph.D. |
Contact |
machineMD AG |
| Erika Han, MD |
Sub-Investigator |
University Hospital Zurich |