Abstract
Digital biomarkers that remotely monitor symptoms have the potential to revolutionize outcome assessments in future disease-modifying trials in Parkinson’s disease (PD), by allowing objective and recurrent measurement of symptoms and signs collected in the participant’s own living environment. This biomarker field is developing rapidly for assessing the motor features of PD, but the non-motor domain lags behind. Here, we systematically review and assess digital biomarkers under development for measuring non-motor symptoms of PD. We also consider relevant developments outside the PD field. We focus on technological readiness level and evaluate whether the identified digital non-motor biomarkers have potential for measuring disease progression, covering the spectrum from prodromal to advanced disease stages. Furthermore, we provide perspectives for future deployment of these biomarkers in trials. We found that various wearables show high promise for measuring autonomic function, constipation and sleep characteristics, including REM sleep behavior disorder. Biomarkers for neuropsychiatric symptoms are less well-developed, but show increasing accuracy in non-PD populations. Most biomarkers have not been validated for specific use in PD, and their sensitivity to capture disease progression remains untested for prodromal PD where the need for digital progression biomarkers is greatest. External validation in real-world environments and large longitudinal cohorts remains necessary for integrating non-motor biomarkers into research, and ultimately also into daily clinical practice.
Similar content being viewed by others
Introduction
Since the last decade, the quest for disease-modifying therapies in Parkinson’s disease (PD) has intensified1. Measuring clinical disease progression accurately and objectively during these trials is of great importance to assess treatment efficacy. Currently, gold-standard outcome measures for such trials are the Movement Disorders Society scales for motor symptoms and daily functioning2. Use of these scales comes with several challenges: they require substantial assessment time, are conducted episodically and are based on subjective interpretation, which leads to considerable measurement errors over time3.
In the past decade, we have seen tangible advancements in the remote digital assessment of the motor symptoms of PD, both in clinical practice and in research settings4,5,6,7,8. Digital sensors can monitor symptoms non-invasively and often passively (without need for active human intervention). The quantification of symptoms and signs at a higher frequency and for longer time frames (even continuously for several digital outcomes) makes such digital biomarkers a more objective, more convenient and potentially more sensitive alternative to the episodic in-clinic assessment of disease progression. As such, several digital biomarkers of motor signs have been included in research settings9,10,11.
Much less attention has been paid to non-motor symptoms, in both research and clinical practice, apart from olfactory dysfunction as a supportive criterion in the MDS diagnostic criteria12. Yet, non-motor symptoms have a substantial negative impact on quality of life in affected individuals13, precede motor symptoms during the prodromal phase up to two decades14 and are subject to daily fluctuations just as motor symptoms, favoring longitudinal measurement over episodic in-clinic assessment15. The earlier occurrence of non-motor symptoms during the prodromal phase makes them potentially more sensitive as progression biomarkers in disease-modifying trials in prodromal populations and more suitable as diagnostic biomarker for the inclusion of at-risk or prodromal individuals in trials, and also as diagnostic marker in clinical practice. Having access to digital biomarkers that can objectively and remotely monitor non-motor symptoms across the spectrum of disease severity paves the way for more patient-relevant and unobtrusive outcomes in research. The importance of such personalization is further emphasized by the recent Food and Drug Administration (FDA) guidelines on patient-focused digital health and drug testing16,17.
In this review, we provide an overview of the current developmental status, validity and reliability of digital, passive non-motor biomarkers for the prodromal and clinically manifest phase of PD. In contrast to previous reviews that were limited to PD studies18,19,20, we did not restrict our search strategy to the PD field, as we recognize that various non-motor symptoms may also be present in other diseases. Therefore, we aimed to investigate whether digital biomarkers that have been developed in these other areas might also have relevance for the PD field. In addition, we will focus on both diagnostic and progression non-motor biomarkers and critically appraise the feasibility of implementing specific non-motor biomarkers in trials. Lastly, we make recommendations for future research based on our review.
Results
Selection of sources of evidence
From PubMed and EMBASE, a total of 18,725 articles and 4678 conference abstracts were screened. 391 articles were eligible, and 119 were eventually included after full-text screening. The flowchart in Fig. 1 depicts the article selection process. The most important reasons for exclusions after title-abstract screening were that a biomarker did not measure the intended symptom or that the measurement of a biomarker was not (fully) passive. A reference manager (Endnote®) library file outlining the complete search strategy and final selection is available from the authors.
Flowchart of the article selection process.
Characteristics of sources of evidence
Of the 20 PD non-motor symptoms, seventeen had one or more proposed passive digital biomarkers. Eight of the 65 studies included in the main table (Table 1) included individuals with PD, and two included individuals with RBD. No other prodromal PD populations were included. Of all eligible studies, most focused on measures of sleep (n = 97), heart rate variability (n = 79), mood (n = 71) and cognition (n = 53). For hallucinations, one study was included. No studies reported eligible digital biomarkers for olfactory dysfunction, erectile dysfunction or color vision disturbances.
An overview of the TRL for both prodromal and manifest PD per non-motor symptom is presented in Table 2. As different non-motor symptoms have different times of onset and progression across the disease course21, we additionally indicated the predictive value of each non-motor symptom in early disease in Table 2. Most proposed digital biomarkers were either a proof of concept or early validation study in healthy controls. Biomarkers for the four sleep-related domains (sleep disturbance, rapid eye movement (REM) sleep behavior disorder, excessive daytime sleepiness and fatigue) were furthest developed with TRLs ranging between 4 and 6 for clinically manifest PD and a TRL of 4 for prodromal PD, with fatigue as exception (TRL 2). The TRL of autonomic function-related biomarkers varied between 2 (bladder dysfunction, orthostatic hypotension) and 6 (heart rate variability) for manifest PD, and were consistently lower for prodromal PD, with the highest TRL of 4 for heart rate variability. Biomarkers for GI-related symptoms had a TRL of 3, and markers for cognitive and neuropsychiatric symptoms had a TRL of 3 (for anxiety and hallucinations) or 4 (for both cognitive deficits and depressive symptoms), and this was identical for manifest and prodromal PD. Sleep-related biomarkers, heart rate variability, constipation, cognitive deficits and depressive symptoms had the highest TRL.
Synthesis of results
A top 3 of the digital biomarkers per non-motor symptom is presented in Table 1. In the Supplementary Table 2, the remaining selected articles per non-motor symptom are listed.
Critical appraisal of digital biomarkers for non-motor PD symptoms
Sleep disturbance
Out of 97 eligible studies, seven studies that assessed sleep using passive digital biomarkers were included22,23,24,25,26,27,28. Five of these included people with PD, four of which had a sample size >200, and two (proof-of-concept) studies included healthy individuals. Most studies used wrist accelerometry signals as part of a wrist-worn multimodal device. Two experimental studies tested a smart pillow that integrates a skin temperature and sweat sensor with sleep features28, and a smart mattress that integrates heart rate variability and respiration sensors with in-bed movement sensors27. The following sleep-related features were investigated: immobile bouts, sleep fragmentation (estimated by movement sensors or heart rate variability), and rest-activity circadian rhythm.
With regard to accelerometry, immobility during sleep could predict abnormal polysomnography (the gold standard) with sensitivity and specificity ≥80%22. No data was available on the sensitivity to longitudinal disease progression. However, one study reports an association between a composite (digital) sleep score and disease duration in 44 individuals with PD (~1-point decrease per disease year relative to healthy controls score of 13)22. One different accelerometry study (n = 48) reports that early PD could be distinguished from healthy controls by measuring turning duration and velocity during sleep (3.2 vs. 1.9 s and 16.2 vs 23.4 degrees/second, p ≤ 0.001)24. With regard to the experimental studies, the mattress-integrated sensor system reports promising heartrate tracking by measuring RR intervals (r2 = 0.99, versus ECG)27, but reports no performance metrics for estimating sleep parameters. Similarly, conceptual testing of a pillow-integrated sensor system suggests sensitivity to changes in sweat rate and temperature overnight in healthy individuals (no statistics)28. TRL ranges between 2 and 6 due to the lack of (longitudinal) data in free-living conditions and uptake as exploratory outcomes in trials. Still, two of the included digital biomarkers are available and approved by regulatory agencies as commercial products, but not for sleep monitoring specifically, limiting TRL to 5 for manifest PD22,23.
REM sleep behavior disorder
Eight studies measured REM sleep behavior disorder. Three studies included people with RBD (n = 35–97), and five studies included people with manifest PD and RBD (n = 22–70). All studies had a control group of either healthy controls or individuals with non-RBD sleep disorders. RBD was measured by quantifying nightly movement distribution, wake bouts, immobile bouts and rest-activity cycles, all using wrist actigraphy29,30,31,32,33,34,35,36.
An algorithm that measures movement distribution has >90% accuracy in diagnosing RBD in people with PD in free-living conditions33. Two other sensors that measure rest-activity cycles and immobile bouts are hampered by low sensitivity (~60%)29,30 despite high (~90%) specificity. One study demonstrates increased sensitivity by adding an RBD questionnaire31. Studies are short-term and have not explored differences between disease severity levels. Furthermore, sensitivity to disease progression has not been investigated. Three sensors are available for use in research. As studies have not yet deployed such sensors as secondary outcomes in trials, TRL is 529,30,33.
Excessive daytime sleepiness
From thirteen eligible studies, seven studies were included25,26,37,38,39,40,41. Five studies included people with PD (n = 28–106, one observational study with n = 239 PD26), one large observational study included elderly (n = 2920) and one experimental in-lab study included healthy controls (n = 30). Excessive daytime sleepiness was defined as immobility during daytime as measured by actigraphy, apart from one experimental study that measured low-frequency/high-frequency (LF/HF) ratio using wrist-worn photoplethysmography.
Concordance for immobile bout detection suggestive of sleep between actigraphy and polysomnography was 85%, mostly due to immobility detection by actigraphy not detected by polysomnography38. A large observational study demonstrated that daytime sleeping (napping) >1 h was predictive of PD at least two years prior to diagnosis (odds ratio 1.96, 95% CI 1.25–3.08), whereas the gold-standard Epworth Sleepiness Scale (ESS) was not. If daytime napping was combined with a positive ESS, the odds for later-onset PD further increased (2.52, 95% CI 1.21–5.27)40. However, sensitivity to disease progression was not investigated, although one early study suggests no progression in number of daytime naps25. LF/HF ratio as a measure of daytime drowsiness levels in an in-lab study showed promising accuracy of 92% for estimating LF/HF ratio from photoplethysmography, but was not tested in free-living conditions. Due to the lack of a large validation trial, TRL remains 4.
Fatigue
Four out of nineteen eligible studies measuring fatigue were included42,43,44,45. All studies included healthy individuals. Three small studies (n = 3–35) proxied fatigue by assessing frequency- and time-domain parameters from RR intervals, including low-frequency amplitude and total spectral power). HRV parameters were occasionally combined with other autonomic features, such as respiration rate or skin impedance.
One proof-of-concept used short recordings from smartphone-measured gaze to estimate fatigue levels. Studies only consisted of in-lab proof-of-concept tests on a small number of controls, demonstrating moderate to high accuracy in lab conditions (75–85%). Algorithms have not been assessed in real-world uncontrolled settings and not in the PD population, limiting TRL to 2.
Pain
Three of seven eligible studies were included that measured pain passively46,47,48. Pain has only been investigated in controlled conditions among postoperative patients, people with dementia and individuals with human immunodeficiency virus (HIV). Studies were small to moderate in size (n = 20–68). Proxies for pain include skin impedance, a combination of psychomotor activity patterns and sleep parameters, and facial micro-expressions. The ground truth used to quantify pain differed from ordinal five-point scales to validated pain scales.
In controlled conditions (at rest or low-intensity activities including walking, sitting, coughing), accuracy for passive pain detection using skin impedance46 and circadian motor patterns48 ranged between 60–86%, depending on pain severity. One mobile application that measures facial micro-expressions using a smartphone cameras had a high correlation (r = 0.88) with a gold-standard pain scale, if combined with questionnaires in the mobile application. Due to early proof-of-concept phase of studies, TRL was limited to 3.
Reduced heart rate variability
79 studies were eligible, of which fifteen were included49,50,51,52,53,54,55,56,57,58,59,60,61,62,63. One study included people with PD (n = 47), and the other studies (n = 5–74) included healthy individuals, and sometimes elderly specifically. Heart rate variability was measured using photoplethysmography in a wide range of different measurement locations, including on chest straps, as part of wrist-worn multimodal smartwatches (i.e. containing additional sensors, mostly accelerometry), unimodal wrist-worn sensors or finger-worn sensors, and one sensor integrated into a smart T-shirt63. Most sensors were tested in controlled conditions over short follow-up durations up to a few weeks, in resting states and during exercise. No studies longitudinally investigated sensitivity to disease progression.
Accuracy and error rates of various commercially available sensors against gold-standard (electrocardiogram) show wide differences between devices of similar modality (e.g. wrist-worn, finger-worn, chest-worn). For example, smartwatches demonstrated a large difference between controlled and free-living conditions, with error rates for time domain parameters such as root mean square of the successive differences between RR intervals (RMSSD) between 4 and 13% against the gold standard52,53,54,55. During the night, sensors appear to have lowest error rates for RMSSDD, likely due to the lower prevalence of motion artifacts62. One observational pilot study in PD demonstrated differences in several frequency analysis parameters from RR intervals, including the high-frequency (HF) power, total frequency (TF) power and the low-frequency/high-frequency power (LF/HF) ratio between early PD, advanced PD and healthy controls56. As longitudinal studies and a first clinical trial in prodromal PD include heart rate variability as outcomes, TRL is 664,65.
Bladder dysfunction
From 20 eligible studies, four small studies (n < 50) were included that described passive and non-invasive bladder dysfunction measures by a smart toilet, lower abdomen near-infrared sensor and smartphone microphone66,67,68,69. Most studies were proof-of-concept tests in healthy controls, and one study included individuals with overactive bladder or outlet obstructions.
In a toilet-integrated sensor system, computer-vision uroflowmeter correlated well with standard uroflowmetry for urination duration and volume (r = 0.92–0.96)66. Smartphone microphones correlated well with flow rate and volume ( > 0.90), although this was only tested in men68,69. Due to the limited level of validation, TRL was 2.
Erectile dysfunction
No eligible digital biomarkers have been reported for erectile dysfunction.
Low skin impedance
Of 22 eligible studies, four studies measuring skin impedance were included46,54,58,60. All studies included healthy controls (n = 72 in total). Skin impedance was measured by a wrist-worn or finger-worn sensor, apart from one study investigating a smart vest that integrates multiple physiological parameters60.
Overall, the studies lacked extensive performance testing, except for one study which showed high agreement between a wrist-worn smartwatch and a gold standard (without reporting performance metrics). Furthermore, the same device could accurately predict pain severity, strengthening its construct validity as a proxy for pain (accuracy 86%). Measuring skin impedance passively with a wearable device is difficult because movements hinder stable contact with the electrodes and cause noisy data. Yet, wristbands are a feasible option with reasonable measurement accuracy for everyday use. Most studies focused on in-lab testing, but we lack findings regarding the sensitivity of skin impedance measures to longitudinal disease progression. The TRL for these sensors is therefore 3.
Hyperhidrosis
From 25 eligible studies, four studies were included70,71,72,73. One study was an early-validation test in athletes (n = 312), and three studies were proof of concepts in healthy individuals (n < 5). Sweat rate was measured using humidity sensors on skin patches, an ear-worn impedance sensor or a wrist-worn sensor. One patch saves data locally, the other patch necessitates taking a photo of the patch for colorimetric analysis.
Sweat rate as measured by wearable patches correlated well with gold-standard sensors (hygrometry) in-lab and during controlled exercise or resting conditions in healthy controls (r > 0.80). However, most patches are validated in the sports context, skin placement is occasionally difficult and most sensors are not suitable for long-term measurement, limiting TRL to 3.
Aberrant skin temperature rhythm
From 22 eligible studies, four studies were included that measured skin temperature passively in healthy controls (n = 57 in total) and people with PD (n = 12)74,75,76,77. Temperature sensors were either integrated into a commercialized smartwatch or finger-worn sensor, or a separate patch.
The patches slightly underestimate the skin temperature (~0.6 degrees Celsius), but show a high mean accuracy across various hot/cold settings (accuracy −0.09 degrees Celsius; precision 0.05 degrees Celsius). The finger-worn sensor is still in an early developmental stage, but the smartwatch is already being applied as outcome measure. This smartwatch study suggests an altered circadian skin temperature pattern in individuals with manifest PD75. However, the accuracy and longitudinal sensitivity to measure skin temperature rhythm in free-living conditions for manifest PD are unclear, limiting the TRL for these sensors to 4.
Orthostatic hypotension
Seven out of fourteen eligible studies that passively measured blood pressure were included60,78,79,80,81,82,83. These studies included healthy controls. Sample sizes ranged considerably, including two smaller studies (n = 10 and 25), two medium-sized studies (n = 91 and 97) and one large study (n = 1057). Blood pressure was measured using a cuffless wrist or finger sensor or a smart T-shirt that integrates a cuff-based sensor with other physiological parameters.
The cuffless sensors showed considerable correspondence to gold standard measurements, with mean differences of −0.1 and 0.0 (SD 3.5–3.6) mmHg for diastolic blood pressure (correlation with gold standard >0.86 across studies). This mean difference is larger in lying and standing position (−0.62 mmHg). The finger worn sensor achieved similar performance metrics with a mean error of 0.9 mmHg systolic and −3.2 mmHg diastolic. The smart vest is still in an early developmental phase but the descriptive results show promising face validity. Cuffless sensors are less intrusive, but are hampered by accuracy of placement and movement artifacts. Whether this is problematic for studies over longer follow-up durations has not been studied. Furthermore, none of the studies have measured sensitivity to drop in blood pressure after standing up, which is necessary for quantifying orthostatic hypotension. Therefore, the digital biomarkers in this category are in TRL 2.
Constipation
Five out of twelve eligible studies were included that measured constipation66,84,85,86,87. These studies included 174 people with PD, 57 healthy controls and 7 people with chronic diarrhea. Three studies measured constipation by an ingestible smart pill, one by measuring abdominal electrodermal activity and one by analyzing excreta through a smart toilet system.
Measurements of the smart pill show construct validity, as transit times correlated with constipation severity (r = 0.32) and gold standard scintigraphy (r = 0.95). The study investigating the abdominal electrodermal patch reported no test performance metrics, but showed face validity using descriptive results. The smart toilet system is still in an early developmental phase, but stool classification agreed considerably with expert opinion (area under the curve >0.89). Taken together, the sensors show good test performance even in real-life settings. However, the association between sensor based constipation measures and PD disease severity or progression has yet to be investigated. Therefore, the TRL is at 4.
Dysphagia
Of 23 eligible studies, five were included88,89,90,91,92. In total, they tested 28 people with PD, 97 healthy controls and 80 people with dysphagia or who were being tested for dysphagia.
One device measured dysphagia directly through surface-electromyography of the swallowing muscles, showing a better signal-to-noise ratio compared to conventional electromyography patches. Another device recorded swallowing sounds through a neck-worn microphone, showing a sensitivity of 93.9% to detect a swallow. A submental patch in combination with a mechano-acoustic chest sensor showed a sensitivity of 80% and specificity of 60% to classify dysphagia. Other approaches include a laryngeal displacement sensor, which could distinguish people with PD from healthy controls by their longer swallow duration (p < .05) and an intra-oral pressure sensor molded in a mouthpiece (93% accuracy to detect a swallow). Although the proposed sensors perform adequately in controlled environments, submental sensors are quite intrusive to wear for longer durations and. Furthermore, the current sensors have not been tested for their sensitivity to track dysphagia progression. Therefore, TRL was 3 for these biomarkers.
Cognitive deficits
From 51 eligible studies, we included eighteen of which we discuss the most advanced ones here93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109. Studies were large cohorts of elderly (total n > 4000) with one cohort including a subgroup of people with manifest PD. Cognitive deficits have been proxied by various modalities. Most studies deployed wrist accelerometry, for example to track rest-activity rhythm, four studies investigated smart homes containing activity and movement sensors in multiple rooms, and one study deployed smartphone keystroke as a proxy for cognitive dysfunction. Seven larger studies (most n > 1000), of which five deployed accelerometry and two a smart home, had longitudinal follow-up up to five years, and demonstrated sensitivity to progression towards MCI or dementia93,94,95,96,97,98.
One smart home study demonstrated an area under the curve of 0.77 for cyclomatic daily activity complexity99, while accelerometry studies report hazard ratios for developing dementia of 1.39 for lower rest-activity rhythm in healthy controls and 1.97 for high intra-day activity variability in mild cognitive impairment. Although most sensors are in an early to late validation stage and have not been tested in PD, especially circadian rhythm and walking features show promising performance in cross-sectionally and longitudinally predicting cognitive decline over a years-long prodromal period. The only study including people with PD demonstrated an association between nocturnal sleep disturbance and cognitive function (working memory r = 0.28; verbal memory r = 0.23), although it is unclear whether cognitive deficits disturb sleep or that this association reflects co-occurrence of symptoms100. Taken together, biomarkers for cognitive function are being extensively validated yet focus more on the general than the PD-specific population. The TRL for measuring cognitive deficits using sensors is therefore 4.
Anxiety
Three out of thirteen eligible studies were included, encompassing 120 healthy controls and 265 people with anxiety or panic disorders110,111,112. Anxiety was measured using smartphone-recorded speech, a wrist-sensor tracking accelerometry and skin and cardiac autonomic features, or wrist-worn actigraphy monitoring the wake-sleep rhythm. Studies were highly heterogeneous in follow-up duration and measured modality of anxiety, ranging from acute anxiety to progression of generalized anxiety over 18 years.
In algorithmically classified smartphone-recorded speech, higher levels of generalized anxiety were associated with the use of fewer reward-related words (r = −0.29). Another study showed that the skin and cardiac autonomic features could accurately detect (acute) psychological stress in 87.2% of the instances, including sitting, running and biking. Lastly, actigraphy was leveraged to track the wake-sleep rhythm, demonstrating predictive value for deterioration of anxiety over an up-to-18 year period (area under the curve = 0.70, sensitivity 84.6%, specificity 52.7%, accuracy 68.7%). The deployed sensors are wearable and feasible to use in everyday life, except for continuous microphone recordings of speech which can be intrusive for privacy reasons. Two studies are in an early developmental stage, whereas one study included longitudinal data. Due to the lack of PD-specific data, TRL for these sensors is 3.
Depressive symptoms
64 articles studied the measurement of depressive symptoms using digital biomarkers, of which eighteen were included113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130. We discuss the most relevant ones here. Studies mostly included healthy controls (total n > 1200). One study included individuals with depressive symptoms (n = 20) but no study included people with PD. Depression was usually defined by a standardized depression scale or ecological momentary assessment. Studies typically proxied depressive symptoms with smartphone use data (such as screen time and frequency of use) and smartwatches or wrist-worn sensors that contain photoplethysmography to track heart rate, accelerometers to track activity and rest-activity rhythm, or a combination of these sensors.
Overall, accuracy and sensitivity are good in proof-of-concept observational studies (≥80%), but study follow-up did usually not exceed months, apart from one study that had a 10-week follow-up. Predictive accuracy of depressive symptoms was highest for integrated physiological and both physical and smartphone activity (>80%)114,115, although one study in a relatively young heterogeneous population of controls (most <64 years) also reached an accuracy of 96% with smartphone data only against the Patient Health Questionnaire-8113. Despite extensive and promising early validation results, there is a lack of longitudinal and PD-specific studies, limiting TRL to 4.
Hallucinations and delusions
One observational pilot study in 79 people (27 hallucinating and 23 non-hallucinating people with PD, 29 controls) investigated rest-activity rhythm using a wrist-worn accelerometer.
The study demonstrates lower interdaily stability, greater night-time activity and reduced relative amplitude of activity in hallucinators (p < .05), also after correction for clinical status. This observational study lacks large-scale validation and longitudinal testing, but forms an important first step to distill proxies for hallucinations in PD (TRL 3)131.
Other non-motor symptoms
For olfactory dysfunction and color vision disturbances, no passive digital biomarkers are available or in development.
General feasibility of digital biomarkers in clinical trials
Multiple studies investigated the feasibility of deploying digital biomarkers with short durations of follow-up in PD, or in populations with other neurodegenerative diseases132,133,134. Populations under investigation included clinically manifest PD (including early stages of PD), other neurodegenerative disorders (e.g. Alzheimer’s disease) and healthy elderly. There were no longitudinal studies that investigated the feasibility in individuals with prodromal disease. Usually, single sensors were investigated, of which wrist accelerometry and smartphone apps were the most common ones. Occasionally, the feasibility of multiple digital biomarkers was assessed. For example, individuals at risk of Alzheimer’s disease wore an Integrated multimodal wrist-worn heart rate variability and accelerometry sensor and had a positive attitude towards current and continued use135. One study investigated the short-term compliance and wearability of up to five sensors concomitantly in people with a variety of manifest neurodegenerative diseases, including PD. 98% wore at least three of five sensors throughout a one-week period. Compliance was lower during the night and already reduced at the end of this week132.
Various studies investigated longer-term compliance in manifest PD. For example, 13-week compliance (median percentage of data collection) of a combination of a smartphone (app) and smartwatch (accelerometer) that aimed to collect digital signals 24/7 was 68% (mean data collection 16.3 h per day). Supported by a helpdesk and participant feedback when data collection declined, data collection had decreased by 23% after 13 weeks136. One large study conducted in nearly 400 people with early PD demonstrated a median smartwatch wear time of 21.1 h per day, with a 5.4% drop-out7. Of note, the high compliance was achieved using a multifaceted approach to promote participant engagement, including a helpdesk, personal assessor, newsletters and participant events64,137. After a six-month trial of passive smartphone data collection, nearly all people with early PD had a positive attitude towards longer use138. With regard to innovative approaches of digital biomarkers, a pilot study of passive in-home sensor monitoring over multiple months was also generally feasible in aging military veterans139.
Positive predictors for better compliance were a seamless-fit design, comfortable wear and either a completely unobtrusive and passive design or a clear added benefit to the participant’s daily life in elderly. For a variety of commercially available devices included in the current study, multiple requirements were not met133. For example, although wrist-worn sensors were generally considered easiest-to-use133,134, limited battery capacity and battery life decline is accompanied by burdensome daily charging and the inability to wear the sensor throughout the day or night. Regarding added benefit for participants, the attitude towards feedback of personal health or activity data is variable and depends on the specific population, disease status and the modality that is measured. For example, compliance with passive digital monitoring of mental health was low over three months in middle-aged veterans140.
Discussion
This study provides an overview of the development of digital biomarkers to measure the presence, severity and rate of progression of non-motor symptoms of PD. We demonstrate that various sleep-related digital biomarkers have been validated in clinically manifest PD, and that biomarkers for REM sleep behavior disorder and excessive daytime sleepiness have been validated in prodromal and at-risk populations. Heart rate variability in particular can be measured passively with high accuracy and likely correlates with disease progression, and promising developments for mood and cognitive features are ongoing in non-PD research fields. For example, By contrast, for several other common non-motor symptoms of PD, the development of passive digital biomarkers is still in an early stage. Specifically, several biomarkers for neuropsychiatric symptoms, autonomic dysfunction and gastro-intestinal features have been developed and tested in healthy participants and patient populations with a neurodegenerative disorder, but these have not yet been validated externally and have uncertain sensitivity to disease progression in PD. In the next sections, we discuss the leading, most promising and least feasible digital biomarkers for PD non-motor symptoms.
Digital biomarkers with high potential
The most promising biomarkers for prodromal PD are currently in the domain of sleep-related symptoms. Digital biomarkers for REM sleep behavior disorder and excessive daytime sleepiness have already been validated in prodromal cohorts. REM sleep behavior disorder and excessive daytime sleepiness are both predictors for later conversion to manifest PD, making these suitable as selection or progression biomarkers for disease modifying trials141. Furthermore, accelerometry-based sleep characteristics such as decreased sleep efficiency and turn velocity have been validated in multiple PD cohorts and correlate well with gold-standard polysomnography. Such features are predictive of PD onset multiple years before diagnosis142 and likely remain predictive of progression throughout the disease course143,144. Therefore, a continued development, longitudinal validation and feasibility assessment of digital markers for sleep, REM sleep behavior disorder and excessive daytime sleepiness in real-world situations could be helpful to monitor disease progression in prodromal disease-modifying trials. Besides sleep-related symptoms, biomarker development for heart rate variability is in the highest developmental stage of all symptoms. At present, studies indicate that heart rate variability is already disturbed in people with REM sleep behavior disorder145,146, reliably distinguishes PD from non-PD147 and is cross-sectionally associated with disease severity148. However, the sensitivity to disease progression is uncertain. Still, we recommend that future trials that recruit a prodromal or manifest PD population could consider including heart rate variability as exploratory digital progression outcome, given the low measurement burden and the option to capture progression of autonomic dysfunction throughout the course of PD.
Opportunities for digital biomarkers
Digital biomarkers with high potential but with a limited current developmental stage include neuropsychiatric features, cognitive function, constipation and orthostatic hypotension. These non-motor symptoms have relatively high predictive value in prodromal PD141,149 and throughout the disease course150,151, although the course of progression of orthostatic hypotension is not well-understood152. We therefore consider it likely that in addition to sleep-related features and heart rate variability, these non-motor symptoms will have high value in composite disease progression scores. A recent review outlines perspectives for deploying wearable technology in monitoring depression153. Measurement techniques for constipation and orthostatic hypotension are objective and direct, and primarily require validation in PD cohorts. As an alternative to passive proxies of neuropsychiatric features and cognitive function that are currently under development, sensitive active alternatives are already available. For example, widespread use of mobile phones enables frequent text analysis of brief digital questionnaires154 and recording and analysis of voice samples to screen for depressive symptoms155. Active alternatives to measure cognitive impairment include remote speech assessments156 and gait tests using accelerometry sensors157. Active alternatives are currently commonly deployed as digital biomarker solution in recent disease modification trials158 and have shown good longitudinal feasibility in PD138. Several initiatives for further development of passive digital biomarkers for neuropsychiatric features are currently underway, such as digital deep phenotyping initiatives159,160.
For markers of autonomic dysfunction, many proposed sensors were poorly integrated for use in PD trials. For example, for hyperhidrosis monitoring, a combination of an accurate patch70, remote synchronization71 and longer measurement duration possibilities72 is necessary for passive monitoring opportunities. From a technological point of view, the development of unobtrusive and specific passive markers for dysphagia (although primarily a motor symptom), hallucinations, pain and bladder dysfunction is most challenging, despite some of these having high value in predicting disease progression161,162,163.
Passive digital biomarkers were not available at all for color vision disturbances, erectile dysfunction and olfactory dysfunction. However, well-validated active or self-administered tests are available for erectile dysfunction and olfactory dysfunction164,165, and under development for color vision disturbances166,167. Recent evidence suggests that olfactory dysfunction might be predictive of symptom decline or a more malignant PD phenotype168 and that non-invasively measuring olfactory bulb function directly shows promise as a diagnostic marker and possibly as a progression marker169. Further development of remote active digital tests for these non-motor symptoms is likely the most rational way forward.
Ethical considerations
There are several ethical considerations regarding widespread digital biomarker development for and deployment in clinical trials. First, to what extent novel digital biomarkers should be patient-centric and patient-relevant compared to more sensitive but potentially clinically irrelevant markers is currently debated170,171. The American Food and Drug Administration (FDA) has recently released guidance documents on patient-focused drug development and argues for a combination of sensitive markers that directly reflect patient-relevant outcomes. This guideline followed on the public debate about the designation of expensive orphan drug status based on outcome measures without direct clinical relevance16. Including patients in the development and evaluation of new biomarkers is often overlooked, which hampers long-term implementation for various reasons172. In this study, people with PD were part of the assessment to give better insight in longer-term feasibility, including relevance of measurement, participant burden and limitations of use, which is essential in development of novel digital applications173,174. Lastly, data ownership and privacy issues require a solid debate, especially with digital biomarkers of non-motor symptoms, intrusive passive markers monitoring daily activity, mobile phone typing and speech as well as geographical and sometimes even semantic location. This subject is outside the scope of this review and has been carefully reviewed elsewhere175.
Limitations
This study has several limitations. Despite the existence of various roadmaps172 and frameworks171,176,177,178 for the development, analysis and critical appraisal of digital biomarkers, TRL classification of sensors in digital biomarkers is hampered by the high heterogeneity in validation methods and study reporting. Due to this heterogeneity across symptom categories, the complexity of the included research could not be reduced into one validated scale for assessment of validity, reliability or feasibility. In addition, we restricted ourselves to peer-reviewed literature, which does not necessarily reflect either the latest stage of development or commercial availability which can be found through commercial channels or in grey literature. Furthermore, the selection of non-motor symptoms was based on the MDS-NMSS, which does not include PD non-motor symptoms previously unappreciated such as (central) respiratory dysfunction, for which several early-stage digital sensors are under development179,180,181,182,183. Lastly, inherent to many non-motor symptoms, objective measurement is often limited to measuring proxies of those symptoms. Notably, similar proxies are deployed for different non-motor symptoms. For example, skin impedance is included both as sign of autonomic dysfunction and as proxy for pain, and psychomotor activity or sleep-wake cycle are leveraged as a proxy for cognitive function, depression, anxiety and hallucinations. This suggests a (current) lack of specificity of such sensors as proxies of non-motor outcomes, and might hamper the interpretation of disease progression. Measuring proxies complicates the translation from both non-PD to PD and from lab studies to real-world situations, in which various known and unknown factors might interfere with those proxies with negative effect on test performance178. Combing multiple measurement modalities might alleviate such challenges184. Such a composite multi-symptom score has additional merit for improved sensitivity and more consistent predictive power across individuals. A major strength of this study is the comprehensive search strategy expanded beyond the PD-field. This strategy yielded several additional markers for various non-motor symptoms undetected with a previous scoping review that was limited to PD populations18. We envision that this broader search gives PD researchers insight in high-potential developments to be translated to their own portfolio or promote interdisciplinary collaboration. Thereby, this study provides a basis for further discussions on the role of digital outcomes in disease modification trials, all of which were still deemed exploratory in a recent consensus paper185.
Conclusion
Here, we have critically reviewed the current state of the art of non-motor digital biomarkers for PD, from inside and outside the PD research field. Furthermore, we have provided insight into the opportunities and challenges for the further development of these important biomarkers. Passive digital sleep-based biomarkers can currently be deployed in clinical trials as secondary outcome, and autonomic biomarkers and neuropsychiatric biomarkers show the highest short-term potential. However, these first require further internal validation before they can be engaged as exploratory outcomes. Future studies may use our critical reflection as a starting point for further digital biomarker development, for subsequent deployment in clinical trials, and ultimately for their application in clinical practice.
Methods
Protocol and registration
This study adheres to the PRISMA-ScR guidelines for scoping reviews and uses the PRISMA-S guidelines for reporting the search strategy. The protocol was pre-registered through the Open Science Framework (https://osf.io/ubyc2/).
Eligibility criteria
All digital biomarkers for non-motor symptoms were considered if the measured modality was objective (not directly influenced by the participant, such as with electronic diaries) and collected data passively. This means that the biomarker should take measurements during natural uninterrupted behavior of individuals in daily life, contrary to active monitoring which requires specific tasks to be performed. Passive measurement was an inclusion criterion due to the level of objectiveness and long-term compliance, important for the feasibility of longitudinal cohorts and clinical trials.
If a potential marker was an indirect proxy of a non-motor symptom, the marker had to be a clear reflection of and closely related to the respective non-motor symptom to be eligible. For example, free-living activity levels were considered as a potential marker of fatigue, but association studies merely reporting co-occurrence of two symptoms were not eligible. Biomarkers could be measured through any modality, irrespective of the shape and location of the measurement application, as long as it was non-invasive. Examples include wearables (e.g. smartwatch, bracelet), furniture-integrated sensors (smart bed, smart toilet), software (mobile phone apps, computer programs), automated video recording and recording of use (computer).
All peer-reviewed articles in English published after 2006 were considered. In addition, we included all conference abstracts published in the last 5 years, which allowed for inclusion of the latest developments. References of included articles were scanned for additional eligible studies. All non-motor symptoms as recognized in the Non-Motor Symptom Scale (NMSS) of the Movement Disorders Society (MDS-NMSS) were considered, including dysphagia (although primarily a motor symptom). Skin impedance and temperature were included as additional putative early signs of autonomic dysfunction186.
Information sources and search
PubMed and EMBASE were initially screened systematically for eligible studies on November 25th, 2021. The search was updated in December 2023.
The structured search string consisted of two elements. The first element included synonyms of all non-motor symptoms or signs, and the second contained a string of synonyms of digital biomarker device modality-related terms. The search strings are made available in the Supplementary Table 1. Because of the large number of articles, and to avoid decision-making bias, the selection process was divided into two stages.
Stage 1: Selection of sources of evidence based on article scope
Four authors (JJD, RB, EP, MM) independently evaluated the eligible entries. Each entry was assigned to two authors, such that all articles were screened and considered for inclusion twice. First, all articles were screened for title and abstract, after which full-text screening was conducted. The main selection criterion was whether the article met the eligibility criteria (see above). Disagreements were resolved in consensus meetings with these four authors.
From all selected full-texts, we extracted a predefined set of variables in a standardized extraction form. Included variables were first author, title, year of publication, journal, study type and objective (validation study, proof of concept, etc.), the measured non-motor symptom, tested population(s), and main field of development (PD or other).
Stage 2: Selection of sources of evidence based on relevance
Based on all selected full-text articles, the same four authors independently selected a top 5 for each non-motor symptom presenting the most advanced state of the field, to reduce the large body of developments into a clear overview of the state of the art, which is the main goal of this study. The selection was based on 1) methodological rigor of the study design (sample size, repeated measures, comparisons made), 2) test performance (validity and reliability) and 3) strength of the evidence for the association between the measured modality and the non-motor symptom (when the digital marker was an indirect proxy). Due to the high heterogeneity in symptoms, study design and reported test parameters within each symptom category, we argued there was no universal scale suitable for assessing such heterogeneous parameters across symptom categories. Therefore, instead, studies were evaluated within each symptom category based on the pre-defined set of three parameters. Any discrepancies in the selection were discussed in multiple meetings by the four authors until consensus was reached.
For the top 5 articles, data extraction was extended by additionally collecting the sensor or system name, the device used, the modality (direct measurement of clinical sign, surrogate marker?), the test setting (clinic, lab or home, free-living), and data on the validity, reliability and feasibility (see below) of this biomarker. We paid particular attention to longitudinal studies reporting serial biomarker measurements, as this gives insight in the sensitivity of a biomarker as disease progression biomarker.
Stage 3: Critical appraisal and analysis
In multiple author meetings, critical appraisals were conducted for the proposed top 5 of digital biomarkers per non-motor symptom. In these sessions, validity, reliability, and feasibility of all biomarkers was evaluated by members with clinical expertise in prodromal and manifest PD, as well as with expertise in sensor development, validation and application of technology in PD. Three patient researchers (RvdH, HM, SM) were part of the critical appraisal process to voice the patient experience with special regard to feasibility.
We evaluated the following variables:
-
Validity of the biomarker versus the (gold-)standard, with a specific focus on discriminant and criterion validity.
-
Reliability, based on reported test-retest reliability.
-
Feasibility, based on the following subdomains: user friendliness (including stigmatization), compliance, implementation and practicality (how easily is it deployed in the participant’s context), based on both quantitative variables and qualitative interpretation. This analysis was conducted together with patient researchers, informed by a roadmap for implementation of patient-centered digital outcomes172.
-
Sensitivity to disease progression, as measured by serial measurements in a longitudinal study, if applicable.
Based on these results, a grade for the stage of development within a specific symptom category was made. The stage of development was based on the Technological Readiness Level (TRL), as defined by the most recent European Union definitions which we adapted to fit the field of digital biomarkers and medical monitoring devices in general (Fig. 2)187. Developmental stages range from development or proof of concept (TRL1, lowest), to public availability or commercialization applied in a longitudinal study (TRL9, highest) via several early and late validation stages in either lab setting or free-living conditions. We assigned a TRL per non-motor symptom or sign separately for the prodromal phase and for the clinically manifest phase. Finally, using a structured process in which both performance metrics, TRL and applicability in PD were assessed, all authors agreed on a final top 3 of digital biomarkers per non-motor symptom to be presented in one main table. The remainder of the included studies was added to Supplementary Table 2.
Technological readiness level (TRL)190 for digital biomarkers in clinical trials.
Data availability
All data generated or analyzed during this study are included in this published article and its supplementary materials. A reference manager file is available from the authors.
References
McFarthing, K. et al. Parkinson’s Disease Drug Therapies in the Clinical Trial Pipeline: 2023 Update. J. Parkinsons Dis. 13, 427–439 (2023).
Goetz, C. G. et al. Movement Disorder Society‐sponsored revision of the Unified Parkinson’s Disease Rating Scale (MDS‐UPDRS): scale presentation and clinimetric testing results. Mov. Disord. 23, 2129–2170 (2008).
Evers, L. J. W., Krijthe, J. H., Meinders, M. J., Bloem, B. R. & Heskes, T. M. Measuring Parkinson’s disease over time: The real-world within-subject reliability of the MDS-UPDRS. Mov. Disord. 34, 1480–1487 (2019).
Artusi, C. A. et al. Implementation of Mobile Health Technologies in Clinical Trials of Movement Disorders: Underutilized Potential. Neurotherapeutics 17, 1736–1746 (2020).
Warmerdam, E. et al. Long-term unsupervised mobility assessment in movement disorders. Lancet Neurol. 19, 462–470 (2020).
Zhan, A. et al. Using Smartphones and Machine Learning to Quantify Parkinson Disease Severity: The Mobile Parkinson Disease Score. JAMA Neurol. 75, 876–880, (2018).
Burq, M. et al. Virtual exam for Parkinson’s disease enables frequent and reliable remote measurements of motor function. NPJ Digit. Med. 5, 65 (2022).
Bloem, B. R., Post, E. & Hall, D. A. An Apple a Day to Keep the Parkinson’s Disease Doctor Away? Ann. Neurol. 93, 681–685 (2023).
Lipsmeier, F. et al. Reliability and validity of the Roche PD Mobile Application for remote monitoring of early Parkinson’s disease. Sci. Rep. 12, 12081 (2022).
Mirelman, A. et al. Arm swing as a potential new prodromal marker of Parkinson’s disease. Mov. Disord. 31, 1527–1534 (2016).
Schalkamp, A. K., Peall, K. J., Harrison, N. A. & Sandor, C. Wearable movement-tracking data identify Parkinson’s disease years before clinical diagnosis. Nat. Med., https://doi.org/10.1038/s41591-023-02440-2 (2023).
Postuma, R. B. et al. MDS clinical diagnostic criteria for Parkinson’s disease. Mov. Disord. 30, 1591–1601 (2015).
Martinez-Martin, P., Rodriguez-Blazquez, C., Kurtis, M. M., Chaudhuri, K. R. & Group, N. V. The impact of non-motor symptoms on health-related quality of life of patients with Parkinson’s disease. Mov. Disord. 26, 399–406 (2011).
Berg, D. et al. MDS research criteria for prodromal Parkinson’s disease. Mov. Disord. 30, 1600–1611 (2015).
Martinez-Fernandez, R., Schmitt, E., Martinez-Martin, P. & Krack, P. The hidden sister of motor fluctuations in Parkinson’s disease: A review on nonmotor fluctuations. Mov. Disord. 31, 1080–1094 (2016).
FDA. Patient-Focused Drug Development Guidance Series for Enhancing the Incorporation of the Patient’s Voice in Medical Product Development and Regulatory Decision Making (FDA, 2023).
FDA. Framework for the Use of Digital Health Technologies in Drug and Biological Product Development (FDA, 2023).
van Wamelen, D. J. et al. Digital health technology for non-motor symptoms in people with Parkinson’s disease: Futile or future? Parkinsonism Relat. Disord. 89, 186–194 (2021).
Merola, A. et al. Technology-based assessment of motor and nonmotor phenomena in Parkinson disease. Expert Rev. Neurother. 18, 825–845 (2018).
Sigcha, L. et al. Deep learning and wearable sensors for the diagnosis and monitoring of Parkinson’s disease: A systematic review. Expert Syst. Appl. 229, 120541 (2023).
Darweesh, S. K. et al. Trajectories of prediagnostic functioning in Parkinson’s disease. Brain 140, 429–441 (2017).
McGregor, S. et al. The use of accelerometry as a tool to measure disturbed nocturnal sleep in Parkinson’s disease. npj Parkinsons Dis. 4, 1 (2018).
Madrid-Navarro, C. J. et al. Validation of a Device for the Ambulatory Monitoring of Sleep Patterns: A Pilot Study on Parkinson’s Disease. Front Neurol. 10, 356 (2019).
Mirelman, A. et al. Tossing and Turning in Bed: Nocturnal Movements in Parkinson’s Disease. Mov. Disord. 35, 959–968 (2020).
O’Dowd, S. T. et al. Longitudinal assessment of sleep in an incident Parkinson’s disease cohort. Mov. Disord. 31, S117 (2016).
Breen, D. P. et al. Sleep and circadian rhythm regulation in early parkinson disease. JAMA Neurol. 71, 589–595 (2014).
Carlson, C. et al. Bed-based instrumentation for unobtrusive sleep quality assessment in severely disabled autistic children. Annu Int Conf. IEEE Eng. Med Biol. Soc. 2016, 4909–4912 (2016).
Li, S. & Chiu, C. A Smart Pillow for Health Sensing System Based on Temperature and Humidity Sensors. Sensors 18, https://doi.org/10.3390/s18113664 (2018).
Filardi, M. et al. Objective rest-activity cycle analysis by actigraphy identifies isolated rapid eye movement sleep behavior disorder. Eur. J. Neurol. 27, 1848–1855 (2020).
Kemlink, D., Perinova, P., Dusek, P., Ruzicka, E. & Sonka, K. Actigraphic screening for the rapid eye movement sleep behavior disorder in Czech population. Sleep. Med. 40, e155 (2017).
Louter, M., Arends, J. B., Bloem, B. R. & Overeem, S. Actigraphy as a diagnostic aid for REM sleep behavior disorder in Parkinson’s disease. BMC Neurol. 14, 76 (2014).
Moerman, C. et al. Towards a handy screening tool for REM sleep behaviour disorder: RDBAct algorithm from wrist actigraphy data. J. Sleep Res. 29, https://doi.org/10.1111/jsr.13181 (2020).
Raschella, F., Scafa, S., Puiatti, A., Martin Moraud, E. & Ratti, P. L. Actigraphy Enables Home Screening of Rapid Eye Movement Behavior Disorder in Parkinson’s Disease. Ann. Neurol., https://doi.org/10.1002/ana.26517 (2022).
Stefani, A. et al. Screening for idiopathic REM sleep behavior disorder: Usefulness of actigraphy. Sleep 41, https://doi.org/10.1093/sleep/zsy053 (2018).
Naismith, S. L., Rogers, N. L., Mackenzie, J., Hickie, I. B. & Lewis, S. J. The relationship between actigraphically defined sleep disturbance and REM sleep behaviour disorder in Parkinson’s Disease. Clin. Neurol. Neurosurg. 112, 420–423 (2010).
Kemlink, D., Perinova, P., Dusek, P., Ruzicka, E. & Sonka, K. Actigraphic differences in the rapid eye movement sleep behavior disorder patients. J. Sleep. Res. 27, 51 (2018).
Bolitho, S. J. et al. Objective measurement of daytime napping, cognitive dysfunction and subjective sleepiness in Parkinson’s disease. PLoS One 8, e81233 (2013).
Kotschet, K. et al. Daytime sleep in Parkinson’s disease measured by episodes of immobility. Parkinsonism Relat. Disord. 20, 578–583 (2014).
Kundinger, T., Sofra, N. & Riener, A. Assessment of the Potential of Wrist-Worn Wearable Sensors for Driver Drowsiness Detection. Sensors 20, https://doi.org/10.3390/s20041029 (2020).
Leng, Y. et al. Excessive daytime sleepiness, objective napping and 11-year risk of Parkinson’s disease in older men. Int J. Epidemiol. 47, 1679–1686 (2018).
Memon, A. et al. Relationship between subjective and objective measures of sleepiness in Parkinson’s disease. In Movement Disorders. Conference: 1st Pan American Parkinson’s Disease and Movement Disorders Congress. Miami, FL United States 32, https://doi.org/10.1002/mds.26972 (2017).
Huang, S., Li, J., Zhang, P. & Zhang, W. Detection of mental fatigue state with wearable ECG devices. Int. J. Med. Inform. 119, 39–46 (2018).
Zeng, Z. et al. Nonintrusive Monitoring of Mental Fatigue Status Using Epidermal Electronic Systems and Machine-Learning Algorithms. ACS Sens. 5, 1305–1313 (2020).
Jeklin, A. T. et al. The association between heart rate variability, reaction time, and indicators of workplace fatigue in wildland firefighters. Int. Arch. Occup. Environ. health 94, 823–831 (2021).
Tseng, V. W. S., Valliappan, N., Ramachandran, V., Choudhury, T. & Navalpakkam, V. Digital biomarker of mental fatigue. npj Digital Med. 4, https://doi.org/10.1038/s41746-021-00415-6 (2021).
Aqajari, S. A. H. et al. Pain Assessment Tool With Electrodermal Activity for Postoperative Patients: Method Validation Study. JMIR mHealth uHealth 9, e25258 (2021).
Atee, M., Hoti, K., Parsons, R. & Hughes, J. D. Pain Assessment in Dementia: Evaluation of a Point-of-Care Technological Solution. J. Alzheimers Dis. 60, 137–150 (2017).
Jacobson, N. C. & O’Cleirigh, C. Objective digital phenotypes of worry severity, pain severity and pain chronicity in persons living with HIV. Br. J. Psychiatry 218, 165–167 (2021).
Wallen, M. B., Hasson, D., Theorell, T., Canlon, B. & Osika, W. Possibilities and limitations of the Polar RS800 in measuring heart rate variability at rest. Eur. J. Appl Physiol. 112, 1153–1165 (2012).
Gamelin, F. X., Berthoin, S. & Bosquet, L. Validity of the polar S810 heart rate monitor to measure R-R intervals at rest. Med Sci. Sports Exerc 38, 887–893 (2006).
Nunan, D. et al. Levels of agreement for RR intervals and short-term heart rate variability obtained from the Polar S810 and an alternative system. Eur. J. Appl Physiol. 103, 529–537 (2008).
Stone, J. D. et al. Assessing the Accuracy of Popular Commercial Technologies That Measure Resting Heart Rate and Heart Rate Variability. Front. Sports Act. Living 3, 585870 (2021).
Kunkels, Y. K., van Roon, A. M., Wichers, M. & Riese, H. Cross-instrument feasibility, validity, and reproducibility of wireless heart rate monitors: Novel opportunities for extended daily life monitoring. Psychophysiology 58, e13898 (2021).
Sagl, G. et al. Wearables and the Quantified Self: Systematic Benchmarking of Physiological Sensors. Sensors 19, https://doi.org/10.3390/s19204448 (2019).
Schuurmans, A. A. T. et al. Validity of the Empatica E4 Wristband to Measure Heart Rate Variability (HRV) Parameters: a Comparison to Electrocardiography (ECG). J. Med. Syst. 44, https://doi.org/10.1007/s10916-020-01648-w (2020).
Niwa, F., Kuriyama, N., Nakagawa, M. & Imanishi, J. Circadian rhythm of rest activity and autonomic nervous system activity at different stages in Parkinson’s disease. Auton. Neurosci. 165, 195–200 (2011).
Barbosa, M. P., da Silva, N. T., de Azevedo, F. M., Pastre, C. M. & Vanderlei, L. C. Comparison of Polar(R) RS800G3 heart rate monitor with Polar(R) S810i and electrocardiogram to obtain the series of RR intervals and analysis of heart rate variability at rest. Clin. Physiol. Funct. Imaging 36, 112–117 (2016).
Santarelli, L. et al. Development of a Novel Wearable Ring-Shaped Biosensor. Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. 2018, 3750–3753 (2018).
Nardelli, M., Vanello, N., Galperti, G., Greco, A. & Scilingo, E. P. Assessing the Quality of Heart Rate Variability Estimated from Wrist and Finger PPG: A Novel Approach Based on Cross-Mapping Method. Sensors 20, https://doi.org/10.3390/s20113156 (2020).
Pandian, P. S. et al. Smart Vest: wearable multi-parameter remote physiological monitoring system. Med Eng. Phys. 30, 466–477 (2008).
Sarhaddi, F. et al. A comprehensive accuracy assessment of Samsung smartwatch heart rate and heart rate variability. PLoS One 17, e0268361 (2022).
Kinnunen, H., Rantanen, A., Kentta, T. & Koskimaki, H. Feasible assessment of recovery and cardiovascular health: accuracy of nocturnal HR and HRV assessed via ring PPG in comparison to medical grade ECG. Physiological Meas. 41, 04NT01 (2020).
Reali, P., Tacchino, G., Rocco, G., Cerutti, S. & Bianchi, A. M. Heart Rate Variability from Wearables: A Comparative Analysis Among Standard ECG, a Smart Shirt and a Wristband. Stud. Health Technol. Inform. 261, 128–133 (2019).
Bloem, B. R. et al. The Personalized Parkinson Project: examining disease progression through broad biomarkers in early Parkinson’s disease. BMC Neurol. 19, 160 (2019).
Darweesh, S. K. L. ClinicalTrials.gov. Slow-SPEED-NL: Slowing Parkinson’s Early Through Exercise Dosage-Netherlands (Slow-SPEED-NL), 2024.
Park, S. M. et al. A mountable toilet system for personalized health monitoring via the analysis of excreta. Nat. Biomed. Eng. 4, 624–635 (2020).
Molavi, B., Shadgan, B., Macnab, A. J. & Dumont, G. A. Noninvasive Optical Monitoring of Bladder Filling to Capacity Using a Wireless Near Infrared Spectroscopy Device. IEEE Trans. Biomed. Circuits Syst. 8, 325–333 (2014).
El Helou, E. et al. Mobile sonouroflowmetry using voiding sound and volume. Sci. Rep. 11, 11250 (2021).
Schultz, R. E. Smartphone App for In-home Uroflowmetry. Urol. Pr. 9, 524–530 (2022).
Baker, L. B. et al. Skin-interfaced microfluidic system with personalized sweating rate and sweat chloride analytics for sports science applications. Sci. Adv. 6, https://doi.org/10.1126/sciadv.abe3929 (2020).
Nyein, H. Y. Y. et al. A wearable patch for continuous analysis of thermoregulatory sweat at rest. Nat. Commun. 12, 1823 (2021).
Gil, B., Anastasova, S. & Yang, G. Z. A Smart Wireless Ear-Worn Device for Cardiovascular and Sweat Parameter Monitoring During Physical Exercise: Design and Performance Results. Sensors 19, https://doi.org/10.3390/s19071616 (2019).
Sim, J. K., Yoon, S. & Cho, Y. H. Wearable Sweat Rate Sensors for Human Thermal Comfort Monitoring. Sci. Rep. 8, 1181 (2018).
Ummel, J. D. et al. Kick Ring LL: A Multi-Sensor Ring Capturing Respiration, Electrocardiogram, Oxygen Saturation, and Skin Temperature(1). Annu Int Conf. IEEE Eng. Med Biol. Soc. 2020, 4394–4397 (2020).
Madrid-Navarro, C. J. et al. Multidimensional Circadian Monitoring by Wearable Biosensors in Parkinson’s Disease. Front Neurol. 9, 157 (2018).
Kim, H. et al. Smart Patch for Skin Temperature: Preliminary Study to Evaluate Psychometrics and Feasibility. Sensors 21, https://doi.org/10.3390/s21051855 (2021).
van Marken Lichtenbelt, W. D. et al. Evaluation of wireless determination of skin temperature using iButtons. Physiol. Behav. 88, 489–497 (2006).
Ibrahim, B. & Jafari, R. Cuffless Blood Pressure Monitoring from an Array of Wrist Bio-Impedance Sensors Using Subject-Specific Regression Models: Proof of Concept. IEEE Trans. Biomed. Circuits Syst. 13, 1723–1735 (2019).
Narasimhan, R., Parlikar, T., Verghesel, G. & McConnell, M. V. Finger-Wearable Blood Pressure Monitor. Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. 2018, 3792–3795 (2018).
Nachman, D. et al. Comparing blood pressure measurements between a photoplethysmography-based and a standard cuff-based manometry device. Sci. Rep. 10, 16116 (2020).
Sola, J. et al. Validation of the optical Aktiia bracelet in different body positions for the persistent monitoring of blood pressure. Sci. Rep. 11, 20644 (2021).
Gaurav, A., Maheedhar, M., Tiwari, V. N. & Narayanan, R. Cuff-less PPG based continuous blood pressure monitoring: a smartphone based approach. Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. 2016, 607–610 (2016).
Rachim, V. P. & Chung, W. Y. Multimodal Wrist Biosensor for Wearable Cuff-less Blood Pressure Monitoring System. Sci. Rep. 9, 7947 (2019).
Mark, E. B. et al. Ambulatory assessment of colonic motility using the electromagnetic capsule tracking system. Neurogastroenterol. Motil. 31, e13451 (2019).
Axelrod, L., Axelrod, S., Navalgund, A. & Triadafilopoulos, G. Pilot Validation of a New Wireless Patch System as an Ambulatory, Noninvasive Tool That Measures Gut Myoelectrical Signals: Physiologic and Disease Correlations. Digestive Dis. Sci. 66, 3505–3515 (2021).
Su, A., Gandhy, R., Barlow, C. & Triadafilopoulos, G. Utility of high-resolution anorectal manometry and wireless motility capsule in the evaluation of patients with Parkinson’s disease and chronic constipation. BMJ Open Gastroenterol. 3, e000118 (2016).
Maqbool, S., Parkman, H. P. & Friedenberg, F. K. Wireless capsule motility: comparison of the SmartPill GI monitoring system with scintigraphy for measuring whole gut transit. Dig. Dis. Sci. 54, 2167–2174 (2009).
Pehlivan, M. et al. An electronic device measuring the frequency of spontaneous swallowing: digital phagometer. Dysphagia 11, 259–264 (1996).
Hadley, A. J., Krival, K. R., Ridgel, A. L., Hahn, E. C. & Tyler, D. J. Neural network pattern recognition of lingual-palatal pressure for automated detection of swallow. Dysphagia 30, 176–187 (2015).
O’Brien, M. K. et al. Advanced Machine Learning Tools to Monitor Biomarkers of Dysphagia: A Wearable Sensor Proof-of-Concept Study. Digital Biomark. 5, 167–175 (2021).
Jayatilake, D. et al. Smartphone-Based Real-time Assessment of Swallowing Ability from the Swallowing Sound. IEEE J. Transl. Eng. Health Med. 3, https://doi.org/10.1109/JTEHM.2015.2500562 (2015).
Kantarcigil, C. et al. Validation of a Novel Wearable Electromyography Patch for Monitoring Submental Muscle Activity During Swallowing: A Randomized Crossover Trial. J. Speech Lang. Hearing Res. 63, 3293–3310 (2020).
Li, P. et al. Circadian disturbances in Alzheimer’s disease progression: a prospective observational cohort study of community-based older adults. Lancet Healthy Longev. 1, e96–e105 (2020).
Targa, A. D. S. et al. The circadian rest-activity pattern predicts cognitive decline among mild-moderate Alzheimer’s disease patients. Alzheimer’s Res. Therapy 13, https://doi.org/10.1186/s13195-021-00903-7 (2021).
Tranah, G. J. et al. Circadian activity rhythms and risk of incident dementia and mild cognitive impairment in older women. Ann. Neurol. 70, 722–732 (2011).
Rogers-Soeder, T. S. et al. Rest-Activity Rhythms and Cognitive Decline in Older Men: The Osteoporotic Fractures in Men Sleep Study. J. Am. Geriatrics Soc. 66, 2136–2143 (2018).
Dodge, H. H., Mattek, N. C., Austin, D., Hayes, T. L. & Kaye, J. A. In-home walking speeds and variability trajectories associated with mild cognitive impairment. Neurology 78, 1946–1952 (2012).
Leng, Y., Redline, S., Stone, K. L., Ancoli-Israel, S. & Yaffe, K. Objective napping, cognitive decline, and risk of cognitive impairment in older men. Alzheimers Dement. 15, 1039–1047 (2019).
Khan, T. & Jacobs, P. G. Prediction of Mild Cognitive Impairment Using Movement Complexity. IEEE J. Biomed. Health Inform. 25, 227–236 (2021).
Gunn, D. G., Naismith, S. L., Bolitho, S. J. & Lewis, S. J. Actigraphically-defined sleep disturbance in Parkinson’s disease is associated with differential aspects of cognitive functioning. J. Clin. Neurosci. 21, 1112–1115 (2014).
Akl, A. et al. Clustering Home Activity Distributions for Automatic Detection of Mild Cognitive Impairment in Older Adults. J. Ambient Intell. Smart Environ. 8, 437–451 (2016).
Dawadi, P. N., Cook, D. J. & Schmitter-Edgecombe, M. Automated Cognitive Health Assessment From Smart Home-Based Behavior Data. IEEE J. Biomed. Health Inf. 20, 1188–1194 (2016).
Alberdi, A. et al. Smart Home-Based Prediction of Multidomain Symptoms Related to Alzheimer’s Disease. IEEE J. Biomed. Health Inf. 22, 1720–1731 (2018).
Akl, A., Taati, B. & Mihailidis, A. Autonomous unobtrusive detection of mild cognitive impairment in older adults. IEEE Trans. Biomed. Eng. 62, 1383–1394 (2015).
Sacco, G. et al. Detection of activities of daily living impairment in Alzheimer’s disease and mild cognitive impairment using information and communication technology. Clin. Interventions Aging 7, 539–549, (2012).
Lussier, M. et al. Smart Home Technology: A New Approach for Performance Measurements of Activities of Daily Living and Prediction of Mild Cognitive Impairment in Older Adults. J. Alzheimers Dis. 68, 85–96 (2019).
Jeon, S. Y. et al. Circadian rest-activity rhythm and longitudinal brain changes underlying late-life cognitive decline. Psychiatry Clin. Neurosci. 77, 205–212 (2023).
Antonsdottir, I. M. et al. 24 h Rest/Activity Rhythms in Older Adults with Memory Impairment: Associations with Cognitive Performance and Depressive Symptomatology. Adv. Biol., e2300138, https://doi.org/10.1002/adbi.202300138 (2023).
Hooghiemstra, A. M., Eggermont, L. H., Scheltens, P., van der Flier, W. M. & Scherder, E. J. The rest-activity rhythm and physical activity in early-onset dementia. Alzheimer Dis. Assoc. Disord. 29, 45–49 (2015).
Sevil, M. et al. Discrimination of simultaneous psychological and physical stressors using wristband biosignals. Comput. Methods Programs Biomed. 199, https://doi.org/10.1016/j.cmpb.2020.105898 (2021).
Jacobson, N. C., Lekkas, D., Huang, R. & Thomas, N. Deep learning paired with wearable passive sensing data predicts deterioration in anxiety disorder symptoms across 17–18 years. J. Affect. Disord. 282, 104–111 (2021).
Di Matteo, D. et al. Smartphone-Detected Ambient Speech and Self-Reported Measures of Anxiety and Depression: Exploratory Observational Study. JMIR Form. Res 5, e22723 (2021).
Opoku Asare, K. et al. Predicting Depression From Smartphone Behavioral Markers Using Machine Learning Methods, Hyperparameter Optimization, and Feature Importance Analysis: Exploratory Study. JMIR mHealth uHealth 9, e26540 (2021).
Rykov, Y., Thach, T. Q., Bojic, I., Christopoulos, G. & Car, J. Digital Biomarkers for Depression Screening With Wearable Devices: Cross-sectional Study With Machine Learning Modeling. JMIR Mhealth Uhealth 9, e24872 (2021).
Narziev, N. et al. STDD: Short-Term Depression Detection with Passive Sensing. Sensors 20, https://doi.org/10.3390/s20051396 (2020).
Saeb, S., Lattie, E. G., Kording, K. P. & Mohr, D. C. Mobile Phone Detection of Semantic Location and Its Relationship to Depression and Anxiety. JMIR Mhealth Uhealth 5, e112 (2017).
Choi, J., Lee, S., Kim, S., Kim, D. & Kim, H. Depressed Mood Prediction of Elderly People with a Wearable Band. Sensors 22, https://doi.org/10.3390/s22114174 (2022).
Saeb, S., Lattie, E. G., Schueller, S. M., Kording, K. P. & Mohr, D. C. The relationship between mobile phone location sensor data and depressive symptom severity. PeerJ 4, e2537 (2016).
Schutz, N. et al. A Sensor-Driven Visit Detection System in Older Adults Homes: Towards Digital Late-Life Depression Marker Extraction. IEEE J. Biomed. Health Informatics., https://doi.org/10.1109/JBHI.2021.3114595 (2021).
Huang, M. et al. Association of Depressive Symptoms with Sleep Disturbance: A Co-twin Control Study. Ann. Behav. Med. 15, https://doi.org/10.1093/abm/kaab040 (2021).
Zhang, Y. et al. Relationship Between Major Depression Symptom Severity and Sleep Collected Using a Wristband Wearable Device: Multicenter Longitudinal Observational Study. JMIR mHealth uHealth 9, e24604 (2021).
Mulryan, P., Affonso, S. & Sullivan, A. Sleep and depression in Parkinson’s disease: Investigating the relationship between sleep and depression using a combination of subjective and objective sleep assessment methods. Ann. Neurol. 88, S196 (2020).
Coutts, L. V., Plans, D., Brown, A. W. & Collomosse, J. Deep learning with wearable based heart rate variability for prediction of mental and general health. J. Biomed Inf 112, https://doi.org/10.1016/j.jbi.2020.103610 (2020).
Dogrucu, A. et al. Moodable: On feasibility of instantaneous depression assessment using machine learning on voice samples with retrospectively harvested smartphone and social media data. Smart Health 17, https://doi.org/10.1016/j.smhl.2020.100118 (2020).
Palmius, N. et al. Detecting bipolar depression from geographic location data. IEEE Trans. Biomed. Eng. 64, 1761–1771 (2017).
Braund, T. A. et al. Smartphone Sensor Data for Identifying and Monitoring Symptoms of Mood Disorders: A Longitudinal Observational Study. JMIR Ment. Health 9, e35549 (2022).
Lee, D. et al. Changes in the Circadian Rhythm of High-Frequency Heart Rate Variability Associated With Depression. J. Korean Med Sci. 38, e142 (2023).
Siddi, S. et al. The usability of daytime and night-time heart rate dynamics as digital biomarkers of depression severity. Psychol. Med 53, 3249–3260 (2023).
Hayashi, M., Takeshima, M., Hosoya, T. & Kume, Y. 24-Hour Rest-Activity Rhythm in Middle-Aged and Older Persons with Depression. Int. J. Environ. Res. Public Health 20, https://doi.org/10.3390/ijerph20075275 (2023).
de Feijter, M., Kocevska, D., Ikram, M. A. & Luik, A. I. The bidirectional association of 24-h activity rhythms and sleep with depressive symptoms in middle-aged and elderly persons. Psychol. Med. 53, 1418–1425 (2023).
Whitehead, D. L., Davies, A. D., Playfer, J. R. & Turnbull, C. J. Circadian rest-activity rhythm is altered in Parkinson’s disease patients with hallucinations. Mov. Disord. 23, 1137–1145 (2008).
Godkin, F. E. et al. Feasibility of a continuous, multi-sensor remote health monitoring approach in persons living with neurodegenerative disease. J. Neurol., 1–14, https://doi.org/10.1007/s00415-021-10831-z (2021).
Keogh, A., Dorn, J. F., Walsh, L., Calvo, F. & Caulfield, B. Comparing the Usability and Acceptability of Wearable Sensors Among Older Irish Adults in a Real-World Context: Observational Study. JMIR mHealth uHealth 8, e15704 (2020).
Areia, C. et al. Wearability Testing of Ambulatory Vital Sign Monitoring Devices: Prospective Observational Cohort Study. JMIR mHealth uHealth 8, e20214 (2020).
Saif, N. et al. Feasibility of Using a Wearable Biosensor Device in Patients at Risk for Alzheimer’s Disease Dementia. J. Prev. Alzheimers Dis. 7, 104–111 (2020).
Silva de Lima, A. L. et al. Feasibility of large-scale deployment of multiple wearable sensors in Parkinson’s disease. PLoS One 12, e0189161 (2017).
Meinders, M. J., Donnelly, A. C., Sheehan, M. & Bloem, B. R. Including People with Parkinson’s Disease in Clinical Study Design and Execution: A Call to Action. J. Parkinsons Dis. 12, 1359–1363 (2022).
Schneider, R. et al. Feasibility of using a smartphone application for the objective evaluation of Parkinson’s disease. Mov. Disord. 32, 973 (2017).
Seelye, A. et al. Feasibility of In-Home Sensor Monitoring to Detect Mild Cognitive Impairment in Aging Military Veterans: Prospective Observational Study. JMIR Form. Res. 4, e16371 (2020).
Betthauser, L. M. et al. Mobile app for mental health monitoring and clinical outreach in veterans: Mixed methods feasibility and acceptability study. J. Med. Internet Res. 22, https://doi.org/10.2196/15506 (2020).
Heinzel, S. et al. Update of the MDS research criteria for prodromal Parkinson’s disease. Mov. Disord. 34, 1464–1470 (2019).
Otaiku, A. I. Association of sleep abnormalities in older adults with risk of developing Parkinson’s disease. Sleep 45, https://doi.org/10.1093/sleep/zsac206 (2022).
Bugalho, P. et al. Polysomnographic predictors of sleep, motor and cognitive dysfunction progression in Parkinson’s disease: a longitudinal study. Sleep. Med. 77, 205–208 (2021).
Dijkstra, F., de Volder, I., Viaene, M., Cras, P. & Crosiers, D. Polysomnographic Predictors of Sleep, Motor, and Cognitive Dysfunction Progression in Parkinson’s Disease. Curr. Neurol. Neurosci. Rep. 22, 657–674 (2022).
Bugalho, P. et al. Heart rate variability in Parkinson disease and idiopathic REM sleep behavior disorder. Clin. Auton. Res 28, 557–564 (2018).
Yang, J. H. et al. Association of heart rate variability with REM sleep without atonia in idiopathic REM sleep behavior disorder. J. Clin. Sleep. Med. 17, 461–469 (2021).
Suzuki, M. et al. Wearable sensor device-based detection of decreased heart rate variability in Parkinson’s disease. J. Neural Transm. 129, 1299–1306 (2022).
Haapaniemi, T. H. et al. Ambulatory ECG and analysis of heart rate variability in Parkinson’s disease. J. Neurol. Neurosurg. Psychiatry 70, 305–310 (2001).
Speelberg, D. H. B. et al. Prodromal Cognitive Deficits and the Risk of Subsequent Parkinson’s Disease. Brain Sci. 12, https://doi.org/10.3390/brainsci12020199 (2022).
Roheger, M., Kalbe, E. & Liepelt-Scarfone, I. Progression of Cognitive Decline in Parkinson’s Disease. J. Parkinsons Dis. 8, 183–193 (2018).
Frazzitta, G., Ferrazzoli, D., Folini, A., Palamara, G. & Maestri, R. Severe Constipation in Parkinson’s Disease and in Parkinsonisms: Prevalence and Affecting Factors. Front Neurol. 10, 621 (2019).
Yoo, S. W. et al. Delayed orthostatic hypotension in Parkinson’s disease. NPJ Parkinsons Dis. 7, 37 (2021).
Fedor, S. et al. Wearable Technology in Clinical Practice for Depressive Disorder. N. Engl. J. Med. 389, 2457–2466 (2023).
Cheng, P. G. et al. Psychologist in a Pocket: Lexicon Development and Content Validation of a Mobile-Based App for Depression Screening. JMIR Mhealth Uhealth 4, e88 (2016).
Ozkanca, Y. et al. Depression Screening from Voice Samples of Patients Affected by Parkinson’s Disease. Digital Biomark. 3, 72–82 (2019).
Robin, J., Xu, M., Kaufman, L. D. & Simpson, W. Using Digital Speech Assessments to Detect Early Signs of Cognitive Impairment. Front. Digit Health 3, 749758 (2021).
Konig, A. et al. Objective measurement of gait parameters in healthy and cognitively impaired elderly using the dual-task paradigm. Aging Clin. Exp. Res. 29, 1181–1189 (2017).
Pagano, G. et al. Trial of Prasinezumab in Early-Stage Parkinson’s Disease. N. Engl. J. Med. 387, 421–432 (2022).
Torrado, J. C. et al. Digital phenotyping by wearable-driven artificial intelligence in older adults and people with Parkinson’s disease: Protocol of the mixed method, cyclic ActiveAgeing study. PLoS One 17, e0275747 (2022).
Dorsey, E. R. et al. Deep Phenotyping of Parkinson’s Disease. J. Parkinsons Dis. 10, 855–873 (2020).
Sung, H. Y. et al. The prevalence and patterns of pharyngoesophageal dysmotility in patients with early stage Parkinson’s disease. Mov. Disord. 25, 2361–2368 (2010).
Barnes, J. & David, A. S. Visual hallucinations in Parkinson’s disease: a review and phenomenological survey. J. Neurol. Neurosurg. Psychiatry 70, 727–733 (2001).
Winge, K., Skau, A. M., Stimpel, H., Nielsen, K. K. & Werdelin, L. Prevalence of bladder dysfunction in Parkinsons disease. Neurourol. Urodyn. 25, 116–122 (2006).
Mueller, C. A. et al. A self-administered odor identification test procedure using the “Sniffin’ Sticks. Chem. Senses 31, 595–598 (2006).
Elhanbly, S. M., Abdel-Gawad, M. M., Elkholy, A. A. & State, A. F. Nocturnal penile erections: A retrospective study of the role of RigiScan in predicting the response to sildenafil in erectile dysfunction patients. J. Adv. Res. 14, 93–96 (2018).
Taore, A., Lobo, G., Turnbull, P. R. & Dakin, S. C. Diagnosis of colour vision deficits using eye movements. Sci. Rep. 12, 7734 (2022).
Ozgur, O. K., Emborgo, T. S., Vieyra, M. B., Huselid, R. F. & Banik, R. Validity and Acceptance of Color Vision Testing on Smartphones. J. Neuroophthalmol. 38, 13–16 (2018).
He, R. et al. Olfactory Dysfunction Predicts Disease Progression in Parkinson’s Disease: A Longitudinal Study. Front Neurosci. 14, 569777 (2020).
Iravani, B., Arshamian, A., Schaefer, M., Svenningsson, P. & Lundstrom, J. N. A non-invasive olfactory bulb measure dissociates Parkinson’s patients from healthy controls and discloses disease duration. NPJ Parkinsons Dis. 7, 75 (2021).
Taylor, K. I., Staunton, H., Lipsmeier, F., Nobbs, D. & Lindemann, M. Outcome measures based on digital health technology sensor data: data- and patient-centric approaches. NPJ Digit Med. 3, 97 (2020).
Manta, C., Patrick-Lake, B. & Goldsack, J. C. Digital Measures That Matter to Patients: A Framework to Guide the Selection and Development of Digital Measures of Health. Digit Biomark. 4, 69–77 (2020).
Espay, A. J. et al. A roadmap for implementation of patient-centered digital outcome measures in Parkinson’s disease obtained using mobile health technologies. Mov. Disord. 34, 657–663 (2019).
Stephenson, D., Badawy, R., Mathur, S., Tome, M. & Rochester, L. Digital Progression Biomarkers as Novel Endpoints in Clinical Trials: A Multistakeholder Perspective. J. Parkinsons Dis. 11, S103–S109 (2021).
Evers, L. J. W., Peeters, J. M., Bloem, B. R. & Meinders, M. J. Need for personalized monitoring of Parkinson’s disease: the perspectives of patients and specialized healthcare providers. Front. Neurol. 14, 1150634 (2023).
Martinez-Martin, N., Insel, T. R., Dagum, P., Greely, H. T. & Cho, M. K. Data mining for health: staking out the ethical territory of digital phenotyping. NPJ Digit Med. 1, https://doi.org/10.1038/s41746-018-0075-8 (2018).
Berger, A. M. et al. Methodological Challenges When Using Actigraphy in Research. J. Pain. Symptom Manag. 36, 191–199 (2008).
Maetzler, W., Klucken, J. & Horne, M. A clinical view on the development of technology-based tools in managing Parkinson’s disease. Mov. Disord. 31, 1263–1271 (2016).
Roussos, G. et al. Identifying and characterising sources of variability in digital outcome measures in Parkinson’s disease. NPJ Digit Med. 5, 93 (2022).
Djeldjli, D., Reguig, F. B. & Maaoui, C. A robust photoplethysmographic imaging for contactless heart and respiratory rates measurement using a simple webcam. Int. J. Med. Eng. Inform. 13, 224–236 (2021).
Grossman, P., Spoerle, M. & Wilhelm, F. H. Reliability of respiratory tidal volume estimation by means of ambulatory inductive plethysmography. Biomed. Sci. Instrum. 42, 193–198 (2006).
Sano, E. et al. Developing a novel mobile wearable tool of measuring real-time respiratory pattern and thoracic motion non-invasively. Respirology 24, 104 (2019).
Sneddon, G. et al. Cardiorespiratory physiology remotely monitored via wearable wristband photoplethysmography: Feasibility and initial benchmarking. Thorax 73, A197 (2018).
Yilmaz, G. et al. A Wearable Stethoscope for Long-Term Ambulatory Respiratory Health Monitoring. Sensors 20, https://doi.org/10.3390/s20185124 (2020).
Sheikh, M., Qassem, M. & Kyriacou, P. A. Wearable, Environmental, and Smartphone-Based Passive Sensing for Mental Health Monitoring. Front. Digit Health 3, 662811 (2021).
Gonzalez-Robles, C. et al. Outcome Measures for Disease-Modifying Trials in Parkinson’s Disease: Consensus Paper by the EJS ACT-PD Multi-Arm Multi-Stage Trial Initiative. J. Parkinsons Dis. 13, 1011–1033 (2023).
Chaudhuri, K. R. et al. The metric properties of a novel non-motor symptoms scale for Parkinson’s disease: Results from an international pilot study. Mov. Disord. 22, 1901–1911 (2007).
EU. Horizon 2020 Work Programme 2014-2015 (EU, 2014)
Elhanbly, S. & Elkholy, A. Nocturnal penile erections: the role of RigiScan in the diagnosis of vascular erectile dysfunction. J. Sex. Med. 9, 3219–3226 (2012).
Ortiz-Tudela, E., Martinez-Nicolas, A., Campos, M., Rol, M. A. & Madrid, J. A. A new integrated variable based on thermometry, actimetry and body position (TAP) to evaluate circadian system status in humans. PLoS Comput. Biol. 6, e1000996 (2010).
Mankins, J. C. Technology readiness levels. White Pap. 6, 1995 (1995).
Acknowledgements
J.M.J.D. is supported by a Therapeutic Pipeline Grant of the Michael J. Fox Foundation (Grant ID MJFF-019201). A.M. receives grants from the Department of Defense of the Israeli Ministry of Justice, the Michael J Fox Foundation, and the EU Joint Program—Neurodegenerative Disease Research. Sirwan Darweesh was supported in part by a Parkinson’s Foundation- Postdoctoral Fellowship (PF-FBS-2026) and a Veni award (09150162010183). Bastiaan Bloem has received research support from Biogen, Cure Parkinson’s, Davis Phinney Foundation, Edmond J. Safra Foundation, Gatsby Foundation, Hersenstichting Nederland, Horizon 2020, IRLAB Therapeutics, Maag Lever Darm Stichting, Michael J Fox Foundation, Ministry of Agriculture, Ministry of Economic Affairs & Climate Policy, Ministry of Health, Welfare and Sport, Netherlands Organization for Scientific Research (ZonMw), Not Impossible, Parkinson Vereniging, Parkinson’s Foundation, Parkinson’s UK, Stichting Alkemade-Keuls, Stichting Parkinson NL, Stichting Woelse Waard, Topsector Life Sciences and Health, UCB, Verily Life Sciences, Roche and Zambon.
Author information
Authors and Affiliations
Contributions
J.J.D., A.M., B.R.B.: conceptualization – J.J.D., R.B., E.P. M.S.C.M. R.H., J.V., S.M., H.M.: data curation, formal analysis, investigation, review & editing. – J.M.J.D., R.B., E.P., M.S.C.M.: original draft writing – J.M.J.D., R.B., A.M., S.K.L.D., L.E., B.R.B.: design, methodology, review & editing.
Corresponding authors
Ethics declarations
Competing interests
B.R.B. has received honoraria from serving on the scientific advisory board for Abbvie, Biogen, UCB, and Walk with Path (paid to the institute); has received fees for speaking at conferences from AbbVie, Zambon, Roche, GE Healthcare, and Bial (paid to the institute). A.M. receives consulting fees from Clexio; receives payment for lectures from Biogen; is on the advisory committee of the Michael J Fox Foundation; and is chair of the advisory board of the Michael J Fox Foundation. The other authors have nothing to disclose.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Janssen Daalen, J.M., van den Bergh, R., Prins, E.M. et al. Digital biomarkers for non-motor symptoms in Parkinson’s disease: the state of the art. npj Digit. Med. 7, 186 (2024). https://doi.org/10.1038/s41746-024-01144-2
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41746-024-01144-2
