Fitbit (Google) is a leading line of wearable fitness trackers and smartwatches now owned by Google (NASDAQ: GOOGL), originally founded in 2007 and headquartered in San Francisco, California. The company's devices have been extensively researched for Parkinson's disease monitoring applications, making them relevant to the neurodegenerative disease research community.
Fitbit devices leverage sophisticated sensor arrays to track movement, heart rate, sleep, and other physiological parameters—data that researchers have harnessed to quantify motor symptoms including tremor, bradykinesia, and gait disturbances in Parkinson's disease patients. With Google's acquisition of Fitbit completed in 2021, the platform now sits at the intersection of consumer wearables and clinical-grade health monitoring[@google].
| Attribute |
Details |
| Parent Company |
Google (Alphabet) |
| Headquarters |
San Francisco, California |
| Founded |
2007 |
| Acquired by Google |
2021 |
| Ticker |
NASDAQ: GOOGL |
¶ Fitbit Devices and Parkinson's Disease
Fitbit wearables offer several features directly relevant to PD symptom tracking and research. The evolution of Fitbit sensor technology has progressively improved the capability to capture movement disorders, though important limitations exist for clinical-grade applications.
¶ Sensors and Capabilities by Device Generation
| Sensor |
Capability |
PD Relevance |
| 3-axis accelerometer |
Movement and activity detection |
Tremor and bradykinesia detection |
| Heart rate sensor (optical) |
Continuous heart rate monitoring |
Autonomic dysfunction assessment |
| Sensor |
Capability |
PD Relevance |
| Enhanced accelerometer |
Higher sampling rates |
Better tremor frequency analysis |
| Altimeter |
Floor counting |
Activity level quantification |
| SpO2 sensor |
Blood oxygen levels |
Sleep apnea detection common in PD |
| EDA sensor (Sense series) |
Electrodermal activity |
Stress and autonomic response |
| Sensor |
Capability |
PD Relevance |
| Multi-sensor array |
Integrated motion processing |
Continuous symptom monitoring |
| Continuous heart rate |
HRV analysis |
Cardiac autonomic dysfunction |
| Skin temperature |
Temperature tracking |
Potential fever detection |
| cEDA sensor |
Continuous EDA |
Emotional stress monitoring |
| GPS |
Location tracking |
Mobility assessment |
Fitbit devices track multiple movement parameters relevant to Parkinson's disease:
- Step count and daily activity: Objective measure of overall mobility
- Exercise modes and intensity: Quantification of physical activity levels
- Sleep stages and quality: Assessment of sleep disturbances common in PD
- Heart rate variability (HRV): Indicator of autonomic nervous system function
- Active zone minutes: Combined metric of activity intensity
- Tremor detection: Limited capability in some devices
Multiple research studies have evaluated Fitbit devices for Parkinson's disease monitoring. The consumer-grade accelerometers in Fitbit devices offer a cost-effective alternative to research-grade inertial measurement units (IMUs), though with important tradeoffs in precision and validation.
Motor Symptom Monitoring
The seminal study by Fereshtehnejad et al. (2019) demonstrated that consumer-grade wearable sensors, including Fitbit devices, could objectively measure Parkinson disease severity[@wearable_pd_2019]. The study found that:
- Accelerometer data correlated with clinical motor scales (MDS-UPDRS)
- Device-based metrics could differentiate PD patients from controls
- Home-based monitoring captured symptom fluctuations
Matsumoto et al. (2020) specifically evaluated the feasibility of Fitbit devices for monitoring motor symptoms in Parkinson's disease[@fitbit_feasibility]. Their findings indicated:
- Acceptable accuracy for detecting tremor and bradykinesia
- Limitations in quantifying very low-frequency movements
- Potential for longitudinal monitoring
Sleep Analysis
Parkinson's disease is associated with numerous sleep disturbances:
Fitbit sleep tracking, while not clinical-grade polysomnography, provides longitudinal sleep data that can identify patterns and changes over time. Studies have used Fitbit data to characterize sleep behavior in PD cohorts[@sleep_pd].
Activity Levels and Gait
Research has validated Fitbit step counts and activity tracking in Parkinson's disease:
- Correlation with standard gait assessments
- Detection of bradykinesia through reduced activity
- Monitoring of daily fluctuations in mobility[@activity_monitoring]
Medication Response Monitoring
Wearable sensors can capture the temporal pattern of medication response ("on-off" fluctuations) in Parkinson's disease. Fitbit devices have been used to:
- Track symptom fluctuations throughout the day
- Measure response to levodopa dosing
- Identify dyskinesias through characteristic movement patterns[@wearable_levodopa]
While Fitbit devices offer valuable research capabilities, important limitations exist:
| Limitation |
Impact |
| Sampling rate |
Limited to ~32 Hz (lower than research IMUs) |
| Sensor precision |
Consumer-grade, not clinically validated |
| Frequency range |
May miss very low-frequency tremor (<2 Hz) |
| Validation |
Limited clinical validation in PD |
| Algorithm opacity |
Proprietary algorithms not transparent |
| Limited accessibility |
Raw data access requires developer API |
| Device |
Key Sensors |
PD Research Utility |
| Fitbit Sense 2 |
cEDA, SpO2, skin temp, GPS |
Autonomic monitoring, sleep |
| Fitbit Charge 6 |
Heart rate, SpO2, GPS |
Activity tracking, HRV |
| Fitbit Inspire 4 |
Basic accelerometer |
Step counts, activity |
| Fitbit Versa 4 |
SpO2, GPS, NFC |
Comprehensive monitoring |
| Device |
Research Applications |
| Fitbit Charge HR |
Early HRV studies |
| Fitbit Alta HR |
Sleep pattern analysis |
| Fitbit Ionic |
First smartwatch with SpO2 |
| Fitbit Inspire HR |
Low-cost activity monitoring |
The following peer-reviewed studies have evaluated Fitbit or similar consumer wearables for Parkinson's disease:
-
Fereshtehnejad et al., 2019 - Objective measurement of Parkinson disease severity using wearable sensors
- Demonstrated correlation between wearable sensor metrics and clinical scales
- Validated use of accelerometer-based measures for PD severity
-
Matsumoto et al., 2020 - Feasibility of Fitbit for monitoring motor symptoms in Parkinson's disease
- Evaluated consumer-grade devices specifically
- Found acceptable accuracy for major motor symptoms
-
Heldman et al., 2016 - Clinician versus machine: reliability of wearable sensors
- Compared consumer vs. research-grade sensors
- Established reliability benchmarks
-
Heijmans et al., 2019 - Digital biomarkers in Parkinson's disease
- Comprehensive review of digital biomarkers
- Role of consumer wearables in PD monitoring
Fitbit devices are being integrated into clinical workflows:
- Remote patient monitoring: Tracking symptoms at home
- Clinical trial endpoints: Objective measure of motor function
- Telehealth integration: Data sharing with healthcare providers
- Deep brain stimulation programming: Movement data for optimal settings
- Rune Labs StrivePD: Primarily Apple Watch integration, but exploring other platforms
- Kinesia: Research-grade wearable system (not Fitbit-specific)
- PDMonitor: European PD monitoring platform
- Hinge Health: Digital musculoskeletal health (includes PD)
Fitbit provides web APIs for data access:
- Fitbit Web API: Cloud-based data access
- Fitbit SDK: Device app development
- Fitbit Studio: Custom clock face development
¶ Data Export and Analysis
Researchers can access Fitbit data through:
- Direct API access: For registered applications
- Manual export: CSV download from Fitbit dashboard
- Third-party platforms: Integration with research platforms
Google completed its acquisition of Fitbit in 2021 for approximately $2.1 billion. The acquisition has several implications for Parkinson's disease research and digital health:
- AI/ML capabilities: Google's expertise in machine learning could improve symptom analysis
- Integration with health records: Potential for EHR integration
- Scale and resources: Greater development resources
- Android ecosystem: Broader device compatibility
¶ Concerns and Considerations
- Privacy: Collection of health data by Google
- Data practices: How health data may be used or shared
- Competitive landscape: Consolidation in wearable market
Beyond motor symptoms, Fitbit devices can contribute to monitoring autonomic dysfunction, which affects up to 70% of Parkinson's disease patients:
- Orthostatic hypotension: Common in PD
- Heart rate variability: Reduced HRV is a biomarker[@hrv_pd]
- Resting heart rate: Elevated in some PD patients[@autonomic_pd]
- Continuous heart rate monitoring: Detects HRV changes
- HRV tracking: Available on premium devices
- Sleep heart rate: Overnight heart rate patterns
| Parameter |
Fitbit Devices |
Research-Grade IMUs |
| Cost |
$50-300 |
$500-5000+ |
| Sampling rate |
~32 Hz |
100-1000 Hz |
| Accuracy |
Consumer |
Clinical/research |
| Validation |
Limited |
Extensive |
| Form factor |
Wrist-worn |
Multiple sites |
| Battery life |
5-14 days |
Hours to days |
- Improved algorithms: Machine learning for PD-specific analysis
- Clinical validation: More rigorous studies in PD populations
- Regulatory clearance: Potential FDA clearance for PD monitoring
- Integration with therapies: Connected to medication delivery systems
Google's health initiatives may enhance Fitbit's capabilities:
- Fitbit Premium: Enhanced health insights
- Google Fit: Integration with broader health ecosystem
- AI research: Applied machine learning for symptom detection
The broader wearable technology landscape continues to evolve with implications for Parkinson's disease monitoring:
Recent advances in sensor technology are improving wearable capabilities for neurological applications:
- Micro-electromechanical systems (MEMS): Improved accelerometer precision
- Inertial measurement units (IMUs): Multi-axis motion tracking
- Biochemical sensors: Sweat-based biomarker detection
- Flexible electronics: Improved comfort and contact
The integration of machine learning with wearable data offers new possibilities:
- Deep learning models: Automated symptom classification
- Pattern recognition: Detection of subclinical manifestations
- Predictive algorithms: Forecasting disease progression
- Personalized baselines: Individualized symptom tracking
¶ FDA Regulatory Landscape
The regulatory environment for digital health devices continues to develop:
- FDA Digital Health Center of Excellence: Established to support innovation
- Software as Medical Device (SaMD): Regulatory framework development
- Real-world evidence: Acceptance for regulatory decisions
- Digital biomarker qualification: Ongoing efforts for standardization
Fitbit data can support movement disorder specialists in several ways:
- Continuous data collection: Objective symptom tracking between visits
- Medication response curves: Visual representation of "on-off" fluctuations
- Activity trends: Quantified mobility changes over time
- Sleep quality metrics: Documentation of nocturnal symptoms
- Data-driven adjustments: Objective basis for medication changes
- Symptom correlation: Identifying triggers and patterns
- Progression tracking: Documenting disease course
Fitbit devices support rehabilitation professionals:
- Step count accuracy: Objective measure of mobility
- Cadence analysis: Monitoring walking pattern changes
- Balance assessment: Activity-based risk evaluation
- Activity tracking: Monitoring prescribed exercises
- Progress documentation: Long-term functional assessment
- Motivation tools: Goal-setting features
Fitbit enables large-scale natural history data collection:
- Cohort monitoring: Tracking large PD populations
- Subtype characterization: Identifying disease patterns
- Environmental factors: Correlating activity with outcomes
Wearable-derived endpoints are increasingly used in trials:
- Exploratory endpoints: Digital biomarker measures
- Progression markers: Objective disease progression indicators
- Patient-reported outcomes: Supplementing traditional measures
Several factors affect Fitbit data quality for PD applications:
- Wrist position: Consistency important for accuracy
- Dominant vs. non-dominant: May affect measurements
- Looseness: Affects sensor contact
- Temperature: Battery and sensor performance
- Humidity: Skin contact sensor accuracy
- Altitude: Pressure sensor calibration
- Skin tone: Optical sensor accuracy variations
- Tattoos: May affect sensor readings
- Edema: Common in PD, affects fit
Researchers should consider several processing factors:
- Frequency analysis: FFT for tremor frequency detection
- Windowing: Appropriate epoch selection
- Filtering: Noise reduction techniques
- Time-domain features: Mean, variance, skewness
- Frequency-domain features: Peak frequency, power spectral density
- Non-linear features: Entropy measures
¶ Competitive Landscape
The consumer wearable market for PD applications includes several platforms:
| Platform |
Strengths |
Limitations |
| Fitbit |
Large user base, established research |
Limited raw data access |
| Apple Watch |
High sampling rate, research programs |
iOS-only |
| Garmin |
Sports-focused, battery life |
Less health focus |
| Samsung |
Global presence, Tizen OS |
Limited PD research |
For clinical research, specialized devices offer advantages:
- Gait monitors: Research-grade accelerometers
- Motion capture systems: Optical tracking
- Inertial measurement units: Multiple sensor fusion
- Comprehensive systems: Combined physiological monitoring
¶ Privacy and Data Security
Fitbit data collection raises important privacy considerations:
- Movement data: Accelerometer readings
- Heart data: Rate, HRV, rhythm
- Sleep data: Stages, quality metrics
- Location data: GPS tracking
- Research partnerships: Data sharing with researchers
- De-identified datasets: Available for research
- User consent: Required for data use
Fitbit must comply with various health data regulations:
- HIPAA: Health data protection requirements
- GDPR: European data protection
- State laws: Various state privacy regulations
The broader wearable health market continues to grow:
- Global wearables: $60B+ market by 2025
- Health-focused segment: Growing rapidly
- PD-specific applications: Niche but expanding
- Platform lock-in: Ecosystem advantages
- Research partnerships: Academic collaborations
- Clinical integration: Healthcare system adoption
Digital health investments continue in the PD space:
- Digital therapeutics: Software-based treatments
- Remote monitoring: Telehealth integration
- AI/ML: Intelligent analytics
Fitbit devices have been used in Parkinson's disease clinical trials:
- Natural history studies: Characterizing PD progression
- Therapeutic trials: As exploratory endpoints
- Device studies: Evaluating wearable interventions
- Remote monitoring: Reduces clinic visits
- Continuous data: Longitudinal symptom tracking
- Patient compliance: Well-accepted devices
- Cost-effective: Lower than research-grade alternatives
- Regulatory: Meeting FDA biomarker qualification
- Data quality: Ensuring consistency
- Validation: Proposing clinically meaningful endpoints
Parkinson's disease tremor is characterized by:
- Frequency: 4-6 Hz resting tremor
- Pattern: "Pill-rolling" movement of the fingers
- Variability: Influenced by medication state and stress
Fitbit accelerometers can detect tremor through frequency analysis[@gait_analysis_wearables]:
- Power spectral density analysis reveals characteristic tremor frequencies
- Machine learning algorithms can distinguish tremor from voluntary movement
- Longitudinal tracking captures tremor severity changes over time
Research Findings:
Studies have demonstrated that wrist-worn accelerometers like Fitbit can distinguish between:
- Tremor-dominant PD vs. postural instability/gait difficulty subtypes
- On-state vs. off-state based on tremor characteristics
- Tremor severity correlated with MDS-UPDRS tremor subscore[@tremor_quantification]
Bradykinesia (slowness of movement) is a cardinal symptom of Parkinson's disease[@bradykinesia_detection]:
Detection Methods:
- Reduced arm swing amplitude
- Slower walking speed
- Decreased activity counts during awake hours
- Prolonged movement initiation time
Fitbit Metrics for Bradykinesia:
- Step count reduction: Correlates with overall mobility decline
- Activity intensity: Active zone minutes decrease
- Movement variability: Higher day-to-day variance in PD patients
Gait disturbances in PD include[@gait_pd]:
- Reduced stride length
- Shuffling gait
- Freezing of gait (FOG)
- Postural instability
Fitbit Contributions to Gait Analysis:
| Gait Parameter |
Fitbit Capability |
Research Utility |
| Step count |
Accurate |
Daily mobility tracking |
| Stride estimation |
Limited |
Research-grade IMUs needed |
| Gait cadence |
Available |
Turning analysis |
| Postural sway |
Limited |
Specialized devices |
Freezing of gait detection using wearables has been extensively studied[@freezers_detection]:
- Accelerometer patterns during FOG events show characteristic "shuffling"
- Algorithms can detect FOG with reasonable sensitivity/specificity
- Fitbit data could supplement clinical FOG assessments
Levodopa-induced dyskinesias (LIDs) are a common complication of long-term PD therapy[@dyskinesia_monitoring]:
Dyskinesia Characteristics:
- Involuntary, irregular movements
- Often correlate with peak plasma levodopa
- Can be choreiform or dystonic
Fitbit Detection Potential:
- Characteristic movement patterns different from tremor
- Time-locked to medication dosing
- Could alert clinicians to dyskinesia onset
Sleep disturbances are among the most common non-motor symptoms in PD[@rem_sleep_disorder]:
Prevalence:
- REM sleep behavior disorder (RBD): 30-50% of PD patients
- Insomnia: Up to 60% of patients
- Sleep apnea: 20-40% prevalence
Fitbit Sleep Tracking Capabilities:
| Feature |
Capability |
Clinical Relevance |
| Sleep stages |
REM, light, deep detection |
RBD screening |
| Total sleep time |
Objective measurement |
Insomnia monitoring |
| Sleep efficiency |
Percentage of time in bed asleep |
Sleep quality |
| Wake episodes |
Nighttime awakenings count |
Fragmented sleep |
| Restless sleep |
Movement during sleep |
Periodic limb movement |
Limitations:
- Not diagnostic for RBD (requires polysomnography)
- Cannot distinguish sleep apnea from other causes
- May overestimate sleep in bed
Autonomic dysfunction affects most PD patients[@autonomic_dysfunction_wearables]:
Common Manifestations:
Cardiac Assessment with Fitbit:
- Heart rate variability: Reduced HRV in PD[@hrv_pd]
- Resting heart rate: Elevated in some patients
- Heart rate trends: Overnight patterns may indicate autonomic issues
Research Applications:
- Correlation with standard autonomic testing
- Monitoring of autonomic symptoms over time
- Early detection of dysautonomia
¶ Data Infrastructure and Analytics
The Fitbit platform provides several data access mechanisms[@fitbita]:
Data Types Available:
- Activity (steps, calories, distance)
- Heart rate (intraday, resting, variability)
- Sleep (stages, efficiency, duration)
- Body (weight, BMI, fat)
- Devices (device-specific data)
API Limitations for Research:
- Rate limits on data access
- Intraday data requires premium for some endpoints
- Data granularity may not meet research standards
Several research platforms have integrated Fitbit data:
Rune Labs StrivePD:
- Primarily Apple Watch focus
- Exploring other wearable platforms
- Integrates with clinical systems
Kinesia (Great Lakes Neurotechnologies):
- Research-grade wearable system
- Validated algorithms for PD
- FDA-cleared for some applications
PDMonitor:
- European CE-marked device
- Designed specifically for PD
- Multi-sensor approach
Consumer wearable data quality for research requires attention[@motion_sensors_accuracy]:
Factors Affecting Quality:
- Device placement (wrist position)
- Skin contact quality
- Battery state
- Algorithm updates
Best Practices:
- Document device model and firmware version
- Validate data completeness
- Cross-check with clinical assessments
- Account for missing data periods
Fitbit and Google have developed machine learning capabilities[@人工智能_parkinson]:
Algorithm Categories:
- Activity classification
- Sleep stage estimation
- Heart rate anomaly detection
- Fall detection
PD-Specific Applications:
- Symptom severity scoring
- Medication response prediction
- Disease progression modeling
- Digital phenotype generation
Academic researchers have developed custom algorithms:
Deep Learning Models:
- CNNs for movement pattern analysis
- RNNs for time-series prediction
- Transformers for longitudinal modeling
Validation Requirements:
- Comparison to clinical scales
- Multi-site validation
- Longitudinal stability testing
Fitbit devices have various regulatory statuses:
| Device Type |
FDA Status |
Implications |
| Fitness tracker |
General wellness |
Not medical device |
| Heart rate monitor |
Class II (some) |
Cleared for certain uses |
| Pulse oximeter |
Class II |
SpO2 measurement |
| ECG |
Class II |
FDA-cleared on some devices |
The FDA is actively working on digital biomarker qualification[@pd_remote_monitoring]:
Qualified Biomarkers:
- Digital fixed gait pattern (under qualification)
- Longitudinal activity measures
Challenges:
- Demonstrating clinical relevance
- Cross-device validation
- Standardization of endpoints
Potential regulatory developments include:
- Software as Medical Device (SaMD): Regulatory framework for algorithms
- Real-world evidence: Use of real-world data for approvals
- Digital Therapeutics: FDA-approved digital treatments
Wearable monitoring offers potential cost savings:
| Cost Category |
Traditional Care |
Wearable-Enhanced |
| Clinic visits |
Regular |
Reduced frequency |
| Hospitalization |
Common |
Early detection |
| Clinical trials |
Expensive |
Remote monitoring |
| Overall care |
Variable |
Potentially lower |
Fitbit data integration into healthcare systems:
- EHR integration: Some health systems incorporating wearables
- Remote patient monitoring: CMS reimbursement codes
- Telehealth platforms: Data sharing capabilities
¶ Patient and Clinician Perspectives
Wearable devices generally have high acceptance in PD populations:
- Non-invasive monitoring
- Minimal burden on daily life
- Engages patients in their care
- Provides objective feedback
Barriers:
- Technical literacy
- Device discomfort
- Privacy concerns
- Cost for premium features
Clinicians have mixed views on consumer wearables:
- Appreciation for objective data
- Concerns about data quality
- Liability questions
- Workflow integration challenges
Value Propositions:
- Continuous vs. episodic data
- Home-based monitoring
- Patient engagement
- Resource optimization
Next-generation sensors under development:
- Improved accelerometers: Higher precision and sampling
- Non-invasive glucose monitoring: Metabolic tracking
- Continuous blood pressure: Cardiovascular monitoring
- Advanced sleep sensing: EEG-derived metrics
Algorithmic improvements anticipated:
- Personalized algorithms: Individual baseline tracking
- Disease-specific models: PD-tuned algorithms
- Predictive analytics: Early warning systems
- Multi-modal fusion: Combining multiple sensors
Wearable-connected therapeutic systems:
- Closed-loop delivery: Sensor-triggered medication
- DBS programming: Movement data for optimization
- Rehabilitation feedback: Physical therapy guidance
Fitbit devices represent a promising platform for Parkinson's disease monitoring and research. While consumer-grade sensors have inherent limitations compared to research-grade equipment, they offer significant advantages in terms of cost, accessibility, and patient compliance.
The current evidence supports the use of Fitbit devices for:
- Long-term activity monitoring
- Sleep pattern assessment
- Heart rate variability analysis
- Research data collection in large cohorts
Future developments including improved algorithms, regulatory clearance for PD-specific applications, and integration with clinical workflows could significantly enhance the utility of consumer wearables in Parkinson's disease management.
As the field of digital health continues to evolve, Fitbit and similar platforms will likely play an increasingly important role in the intersection of consumer technology and clinical neuroscience. The key challenge remains bridging the gap between convenience and clinical validity—a gap that ongoing research and technological development continue to narrow.
- Fitbit Official Website
- Google Fitbit Acquisition
- Fitbit Health API
- Parkinson's Foundation - Technology and PD
- Fereshtehnejad SM, Khoshavi M, Lokhandwala S, et al, Objective measurement of Parkinson disease severity using wearable sensors, Neurology (2019)
- Matsumoto JY, Griffin L, Stufflebam A, et al, Feasibility of consumer-based accelerometers for monitoring motor symptoms in Parkinson's disease, J Parkinsons Dis (2020)
- Heldman DA, Espay AJ, LeWitt PA, et al, Clinician versus machine: reliability of wearable sensors for assessing Parkinson's disease, J Parkinsons Dis (2016)
- Heijmans M, Haber A, Graf J, et al, Digital biomarkers in Parkinson's disease, Nat Rev Neurol (2019)
- Ossig C, Reichmann H, Storch A, et al, Treatment strategies and motor fluctuation endpoints in Parkinson's disease, J Neural Transm (2016)
- Mayor RJ, Ahmad F, Mukherjee S, et al, Heart rate variability analysis in Parkinson's disease, Clin Auton Res (2018)
- Amara AW, Chahan S, Parkinson's disease sleep disturbances, Curr Neurol Neurosci Rep (2019)
- Tsanas A, Little MA, McSharry PE, et al, Novel smartphone and wearable sensor-based assessment of levodopa-induced dyskinesia, J Parkinsons Dis (2012)
- Lord S, Galna B, Coleman S, et al, Accuracy of wearable sensors for measuring gait and postural sway in Parkinson's disease, Mov Disord (2014)
- Hasan H, Burrows M, Lee K, et al, Wearable sensor-based quantification of bradykinesia in Parkinson's disease, IEEE J Biomed Health Inform (2020)
- Kubota KJ, Chen JA, Grossman M, et al, A systematic review of wearable sensors for Parkinson's disease monitoring, Sensors (2021)
- Mera A, Heldman DA, Espay AJ, et al, Feasibility of home-based wearable sensor monitoring in Parkinson's disease clinical trials, J Parkinsons Dis (2017)
- Jain S, Goldstein DS, Cardiovascular autonomic dysfunction in Parkinson's disease, Parkinsonism Relat Disord (2012)
- van Uem JM, Maetzler W, White DK, et al, Accuracy of activity monitors for measuring daily activity in Parkinson's disease, Neurology (2016)
- Deuschl G, Beck CA, Bain PG, Tremor quantification in clinical practice, Parkinsonism Relat Disord (2012)
- Parkinson's Foundation, Technology and Parkinson's Disease
- Del Din S, Godfrey A, Rochester L, Validation of an accelerometer-based method for detecting gait events, Sensors (2016)
- Rovini E, Maremmani C, Cavallo F, How wearable sensors can support Parkinson's disease management, Sensors (2019)
- Lin M, Levine W, Huan Y, et al, Artificial intelligence in Parkinson's disease, Nat Rev Neurol (2020)
- Nieuwboer A, Rochester L, Hummelen L, et al, Reliability of gait event detection in freezing of gait, J Neurol (2013)
- Godinho C, Domingos J, de Cunha M, et al, A systematic review of Parkinson's disease rehabilitation, J Parkinsons Dis (2016)
- Iranzo A, Santamaria J, Tolosa E, The role of sleep disorders in Parkinson's disease, Nat Rev Neurol (2013)
- Postuma RB, Gagnon JF, Montplaisir J, Autonomic dysfunction in Parkinson's disease, Nat Rev Neurol (2013)
- Dobkin BH, Dorsch A, The promise of wearable motion sensors, JAMA Neurol (2016)
- Lakshminarayana R, Wang D, Burn D, et al, Using smartphone and wearable devices for Parkinson's monitoring (2017)
- Fisher JM, Nguyen D, Yang J, et al, Levodopa-induced dyskinesia measurement in Parkinson's disease (2016)