Application of Wearable Sensors in Parkinson’s Disease: Future Prospective

1. Introduction

Parkinson's Disease (PD) is a complex neurodegenerative disorder that presents with a wide range of motor (tremors, bradykinesia, postural instability) and non-motor symptoms, including autonomic disorders, behavioral and cognitive deficits [1]. In late stage, levodopa-related motor complications include: a) off effects, where the duration of the drug's beneficial effect decreases with each dose; b) delayed ON/noON fluctuations, where the time between taking levodopa and the clinical effect is either delayed or nonexistent; and c) random ON-OFF fluctuations [2]. Traditional methods of assessing PD symptoms rely on subjective clinical evaluations (e.g. diaries, questionnaires), which can be prone to bias and variability [3,4].

However, the emergence of wearable sensor technology has revolutionized the monitoring and management of PD by enabling objective, continuous, and remote assessment of key motor symptoms. Wearable sensors are small, lightweight devices that can be worn on the body to capture various physiological parameters, such as movement, tremor, gait, balance, cognitive and sleep patterns [5,6]. These sensors utilize a range of supporting networking technologies, including accelerometers, gyroscopes, magnetometers, and electromyography, to collect data on motor and non-motor symptoms in PD patients [7]. While some methods have suggested using many sensors on various body parts, more contemporary methods often utilize sensors on the foot, wrist, fingers, cellphones with wrist adapters, or devices that resemble watches [7,8,9,10,11]. Additionally, these methods have been combined with threshold approaches, automatic classification algorithms, and deep learning techniques [7-11} to offer inexpensive and non-invasive remote surveillance systems. The data collected by wearable sensors can be analyzed in real-time or stored for later review, providing valuable insights into disease progression, treatment efficacy, and quality of life through developments in machine learning [7,8] (Figure 1.).

Previous reviews focused on wearable technologies [9] such as smartphone apps [10], inertial sensors using maschine learning models to predict motor / non- motor symptoms of PD [7,11]. This narrative review aims to provide a more comprehensive overview of the current applications, benefits, challenges, and future directions of wearable sensors in the management of PD.

2. Materials and Methods

I reviewed articles published during the 2000–2024 period, focusing on wearable sensors in PD patients. I searched PUBMED, SCOPUS and EMBASE from 2000–2024 using the following terms: “Parkinson’s disease”, “Wearable sensors”, ‘’PDMonitor®’’, ‘’PKG®system’’, ‘’STAT-ONTM’’, ‘’ KinesiaONETM ‘’, ‘’FeetMe® Monitor’’, ‘’DynaPort7TM’’, ‘’Mobility Lab™’’, ‘’accelerometers’’, ‘’gyroscopes’’, ‘’magnetometers”, “Smartphones”, “Smart homes’’ using Boolean operators. Including criteria were original research papers, reviews, and meta-analyses that have been published in peer-reviewed publications in English language. The snowballing procedure was carried out to screen the references of each selected article for potential extra papers to cover the current key evidence following by a full-text evaluation of pertinent research. For the narrative synthesis, my initial categorization was based on the benefits of wearable sensors in the management of PD, followed by a discussion of limitations and potential future directions.

3. Results

3.1. Wearable, Inertial Sensors and FDA-Approved Devices

Small electrical devices called inertial sensors, sometimes referred to as motion sensors, monitor acceleration, angular velocity, and sometimes the intensity of magnetic fields. An object's orientation and motion in three dimensions are determined using these measurements. Accelerometers, gyroscopes, and magnetometers are examples of inertial sensors that are frequently used to record a variety of motion-related data.Wearable motion sensors have shown promise in the everyday life monitoring of PD patients due to their compact size, low weight, and low energy requirements [12,13,14]. Table 1 summarizes research on wearable PD sensor.

By detecting unexpected fluctuations in acceleration, accelerometer sensors enable the monitoring of posture and movement. Gyroscopes, on the other hand, detect rotational movements and angular velocity. Magnetometers evaluate heading and orientation tracking by measuring orientation and recording alterations in the magnetic field. The wrist accounts for 18.2% of all accelerometer, gyroscope, and magnetometer attachments, with the lower leg (13.6%), waist (12.1%), upper leg (9.1%), and foot (6.1%) following in order of frequency [15]. When used together, these tools provide unbiased information on changes in posture and motor function. The most widely used sensors are a magnetometer, gyroscope, and tri-axial accelerometer. An inertial measurement unit (IMU), or three-dimensional construct with linear and angular velocity, is produced when they are integrated into a single device [16,17,18]. Small devices enable extended free walking outside labs for assessing PD motor symptoms, particularly due to the Hawthorne observation effect, which suggests that free-living activities with varying complexities and distractions may reduce attention.

Inertial sensors can differentiate PD from healthy controls based on movement patterns, enabling machine learning algorithms to identify specific features. They can also monitor PD progression, capture changes in movement patterns, and provide insights into the disease course [16,17,18,19,20,21,22,23,24]. Longitudinal data can inform treatment decisions and assess efficacy, such as tremor severity and gait disturbances.

Inertial sensors have been used in observational studies to advance digital progression biomarkers for PD such as the Personalized Parkinson Project [25], the Cincinnati Cohort Biomarker Program (CCBP) [26], the Wearable Assessment in The Clinic and at Home in PD (WATCH-PD) [27,28], the Oxford Parkinson Disease Centre (OPDC) discovery cohort [29] and the ALAMEDA study [30]. These devices objectively measure the efficacy of new interventions by capturing changes in motor symptoms over trials. By documenting variations in motor symptoms during the trial, these tools are used to objectively assess the effectiveness of novel therapies, such as medicine, physical therapy, and deep brain stimulation. Inertial sensors can better assess the efficacy of novel treatments and provide insightful information about reactions to therapy by delivering continual and quantitative data. Inertial sensors are frequently employed in clinical trials to evaluate main objectives pertaining to the degree of motor symptom severity and functional improvement. For instance, gyroscopes record improvements in gait metrics, whereas accelerometers detect changes in the intensity of tremor [25,26,27,28]. Experts can ascertain whether the therapy significantly affects motor symptoms and functional capacities by examining this data. Inertial sensors may offer secondary outcome measurements in addition to primary endpoints. These gadgets can provide a thorough understanding of a person's motor function and general mobility by recording information on movement pattern, activity level, and postural change. The effect of the treatment on routine tasks and quality of life—two crucial secondary goals in PD research—can be evaluated using this data. With everything taken together, inertial sensors are essential to the planning and implementation of PD clinical trials because they offer precise and quantitative assessments of functional skills and motor symptoms [25,26,27,28,29,30]. The accuracy and dependability of trial results could be increased by this technique, which would eventually result in the creation of more potent therapies.

NICE recommends wearable devices like PDMonitor, PKG, STAT-ON, Kinesia360, and KinesiaU for remote monitoring of PD to inform treatment, as they belong to Class I or IIa medical devices [31] (Table 2).

The chronology of the above FDA-cleared wearable devices for PD is shown in Figure 2.

The PDMonitor® is a Class IIa medical device designed to monitor patients with PD. It consists of five IMU sensing devices placed on wrists, shins, and waist (1) a SmartBox used to collect, process and upload data to the Cloud; (2) Five sensing/monitoring devices each with internal data storage (up to 7 days) that can transfer data when docked into the SmartBox; and (3) accessories for attaching the devices to the patient. The system was used to assess 65 PD patients in-hospital and at home for at least 7 hours/day [32]. The device calculated the probability of the patient being OFF based on gait, tremor, FoG, and bradykinesia.

The PKG®system, first FDA-cleared in 2014, is a medical device for PD patients. It captures motor symptoms, monitors medication administration, detects device removal, and monitors movement during sleep. The watch, attached to the wrist, collects data 6-10 days and has a rechargeable battery and accelerometer. The data is translated into bradykinesia and dyskinesia scores, which generate a graphical report. The device also measures motor complications, including fluctuations and dyskinesia scores. The device has shown high sensitivity and specificity in identifying tremors and distinguishing between fluctuators and non-fluctuators vs the gold-standards, ie, the AIMS score (dyskinesia) and UPDRS-III (bradykinesia) [33,34].

The STAT-ONTM is a wearable sensor that uses real-time machine learning algorithms to measure inertial signals. It weighs 86g and consists of two ultralow triaxial nano-accelerometers and two microcontrollers using Bluetooth Low Energy wireless technology. The sensor records gait, step length, FoG, falls, and trunk or neck dyskinesia. Despite its high sensitivity and specificity, it has been compared to patient self-reporting. The device has been successful in identifying trunk dyskinesia, FoG, and falls, with high sensitivity>93% and specificity75–98%[35]. The sensor is user-friendly and allows for input of events like medication use.

The KinesiaONETM is a Class I medical device that evaluates motor signs of PD using an iPad mini loaded with KinesiaONE software, an electronic diary, a wearable sensor, and a charge pad [33]. The mobile application uses TETRAS and UPDRS to help users through activities, and the sensor captures angular velocity and linear acceleration. Severity scores are determined by algorithms when the data is uploaded to a secure cloud repository. Two wireless motion sensors are used by the Kinesia360TM gadget to measure body position, steps taken throughout routine tasks, bradykinesia, dyskinesia, and tremor. Numerous investigations have demonstrated the device's dependability, with good sensitivity and specificity [36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53].

The most recent iteration of the DynaPort Hybrid technology, the DynaPort7TM, evaluates movement by integrating gyroscopes and acceleration sensors. Among other things, it gauges gait rhythmicity, steps, and walking patterns. The DynaPort7 has been proven to accurately identify misstep occurrences in a huge cohort of senior adults who do not have dementia or Parkinson's disease [54]. According to in-hospital evaluations, MDS-UPDRS part III was substantially correlated with seven mobility parameters [55]. The DynaPort gadget is acceptable and pleasant.

The FeetMe® Monitor is a real-time monitoring device that measures gait parameters and plantar pressure in the insole. The insole is equipped with a single Inertial Measurement Unit (IMU) that comprises a three-dimensional accelerometer and a three-dimensional gyroscope, enabling the detection of movement across six axes. It measures key gait cycle events and other factors stride length, stride duration, stance duration, swing duration, step duration, single support duration. The device is attached to an insole and has been proven effective in PD patients [56,57,58,59,60].

The Mobility Lab™ device uses wearable inertial sensors to monitor gait and balance in patients with PD. It includes six wireless sensors, software for user and patient guidance, and automated analysis. The system can detect subtle parkinsonian gait in early PD, evaluate parkinsonian gait/posture response to levodopa, and evaluate global gait and balance impairment. It also detects FoG episodes with high agreement with clinical raters, making it a valuable tool for gait and postural analysis in hospitalized patients[61].

3.2. Acoustic, Optical and Electrical and Electrochemical Sensors

Acoustic sensors can analyze speech patterns to assess vocal function and speech intelligibility in PD. They can also quantify gait parameters, such as step length and velocity. Wan et al. [62] used data from early disease patients' persistent vowel phonations to determine the severity of the disease. Deng et al. [63] showed in another investigation that PD may be effectively identified using voice, gait/rest, and tapping data. It's interesting to note that sustained vowels increased these data's discriminative power [64]. Clinicians can assess the effectiveness of speech treatment, monitor the course of the condition, and improve communication techniques in Parkinson's disease by tracking voice characteristics over time. Optical sensors can detect gait parameters and finger tapping in PD, enabling healthcare professionals to monitor treatment strategies and adjust accordingly. EMG sensors can assess muscle activity and movement patterns in PD, identifying muscle groups affected by the disease and guiding personalized rehabilitation strategies. Advances in electroencephalographic (EEG) technology have enabled the use of these sensors to monitor motor symptoms in PD, such as impulse control disorders and heart electrical activity. These technologies can help healthcare providers assess cardiac function and intervene early to mitigate adverse events.

An electrochemical biosensor is defined as "a self-contained integrated device able to provide specific quantitative or semi-quantitative analytical information, by using a biological recognition element (i.e., biochemical receptor), retained in direct spatial contact with an electrochemical transduction element" by the International Union of Pure and Applied Chemistry (IUPAC) [65]. One study, for instance, showed how to use a wearable electrochemical aptasensor for circulating blood α-synuclein oligomers, which has good sensitivity and specificity [66]. Rapid quantification and real-time monitoring are still challenges, despite the fact that the severity of motor symptoms is correlated with the plasma levodopa concentration [67]. These encouraging investigations unquestionably call for more investigation and advancement in along with possible use in different biofluids, including as urine, tears, and saliva.

3.3. Smartphones

Smartphones' extensive use, versatility, and multitasking make them potentially essential tools for identifying common symptoms of PD [68]. To investigate fine-motor deficits in early PD, for instance, Iakovakis et al. [69] employed natural smartphone touchscreen typing. A remote study with 253 individuals who contributed 252,000 typing sessions through the iPrognosis mobile app (Apple Computer Inc., Cupertino, CA, USA) and a dataset of 36,000 typing sessions from 39 subjects were used to validate this methodology. The findings were noteworthy and showed a close relationship with MDS-UPDRS part III. likewise, digital assessments and traditional clinical scales, such as the MDS-UPDRS, Symbol Digit Modalities Test, and Montreal Cognitive Assessment, demonstrated a good connection in the 12-month WATCH-PD multicenter longitudinal trial conducted with cellphones [26].

3.4. Smart Domotics

The word "domotics" is derived from the Latin "domus," implying "house," and "tics," which refers to computer science, robotics, and telematics. In addition to providing overall convenience, safety, and autonomy, domotics and smart homes can be a rich source of real-time data for ongoing monitoring and management in PD[70]. Combining wearable technology with stationary equipment, such as video and temperature, pressure, and movement sensors, may have benefits in thoroughly describing patterns of motion in the home [71]. Furthermore, smart mattresses, automated electrical equipment, and voice-activated lights provide real advantages to those with incapacitating illnesses [71]. Devices for virtual reality and leisure time could be employed in home physiotherapy to help people with PD with their balance and walk. For instance, a recent study conducted at home with a Nintendo® Wii console and wireless controller found that constant visual input made it easier to focus and execute movements [72].

4. Discussion

There are several key benefits of using wearable sensors in the management of PD. Firstly, wearable sensors enable continuous and objective monitoring of motor/non-motor symptoms, offering a more accurate assessment of disease progression compared to traditional episodic evaluations [7,8,9,10,11,12,13,15,18,30,32,33,52,53,55]. Ultimately, wearables are suitable for measuring autonomic status in real time, such as with ingestible respiratory function devices and cuffless blood pressure monitors. Inter-rater variability, symptom swings in and out of clinic, and sensitivity, infrequent rating scales and questionnaires are some of the shortcomings these solutions address with relation to discrepancies [9]. This real-time data can help clinicians personalize treatment plans, adjust medication dosages, and optimize therapeutic interventions to improve patient outcomes in and out of the clinical setting [25,28,73].One example of wearable technology is a non-invasive vibrotactile glove that uses implanted mechanical triggers to stimulate PD patients' fingertips and reduce motor symptoms including tremor and gait [73]. Additionally, wearable sensors can detect subtle changes in motor symptoms that may not be apparent during clinic visits, providing early indicators of disease progression or treatment response [39,40,42,43,45,46]. Physiology-driven inputs from wearable technology can be utilized to guide customized PD treatments like deep brain stimulation (DBS). More precisely, PD patients' shank velocity from IMUs can be utilized as a kinematic control variable for directing dynamic STN-DBS intervention in real-time to ameliorate freezingand gait dysfunction [74]. Furthermore, wearable sensors enable remote monitoring and telemedicine, allowing patients to receive care from the comfort of their homes and reducing the burden of frequent hospital visits [9].

Despite the numerous benefits of wearable sensors, there are also several challenges and limitations to consider. One of the main challenges is the integration of sensor data into clinical practice, as clinicians may not have the resources or expertise to interpret and utilize the complex data generated by wearable sensors. In addition, issues related to data privacy, security, and regulatory compliance must be carefully addressed to ensure the ethical use of sensor data in healthcare settings. Furthermore, wearable sensors may not be suitable for all PD patients, particularly those with cognitive impairments or mobility issues that prevent them from wearing or using the devices effectively. Notably, decreased score on these smartphone cognitive tests was associated with a lower score on the traditional cognitive assessment [9]. Finally, there is a need for standardization of sensor technologies, data analytics, and outcome measures to facilitate the adoption of wearable sensors in routine clinical practice.

The future of wearable sensors in the management of PD holds great promise for advancing precision medicine and personalized care. With the continued development of artificial intelligence and machine learning algorithms, wearable sensors have the potential to predict disease progression, optimize treatment regimens, and identify novel biomarkers for early diagnosis and intervention [75,76,77,78,79].Data enhancement and domain adjustment might enhance the generalizability of models in the healthy-to-PD transition scenario, and deep neural network models based on healthy data may be able to effectively identify actions carried out by PD patients [76].However, the recent PD deep learning models are inadequate against optimal dataset management problems, effective time series data analysis problems, real-time sensor handling problems, data regularization issues, and home-based PD monitoring solutions [76,77]. To resolve these issues, pant body sensor networks are the driving force behind the novel deep clustering model's autonomous health monitoring and analysis paradigm, which enables quick and efficient decision-making [78]. Additionally, the integration of wearable sensors with smart devices and mobile applications can enhance patient engagement, self-management, and adherence to treatment plans. A current study using IMUs combined with convolutional neural network modeling to measure FOG in PD individuals evaluated sensor sites that optimized patient adherence for individualized at-home motor symptom tracking [79]. Collaborative efforts among researchers, clinicians, industry partners, and regulatory agencies will be essential to drive innovation and adoption of wearable sensors in PD care.

5. Conclusions

Wearable sensors represent a powerful and innovative tool for monitoring and managing motor symptoms in PD. By providing objective, continuous, and remote assessment of key motor symptoms, wearable sensors offer numerous benefits for clinicians, patients, and caregivers. However, the widespread adoption of wearable sensors in clinical practice will require addressing challenges related to data interpretation, privacy, compliance, and standardization. Moving forward, collaborative efforts and interdisciplinary approaches will be essential to leverage the full potential of wearable sensors in improving outcomes and quality of life for PD patients. Future prospects for wearable sensors in PD include the development of more advanced and user-friendly devices, integration with other technologies such as artificial intelligence and telemedicine, and validation in large-scale clinical trials. These advancements could revolutionize the management of PD by enabling early diagnosis, personalized treatment plans, and improved quality of life.