Wrist Rigidity Evaluation in Parkinson’s Disease: A Scoping Review

(1) Background: One of the main cardinal signs of Parkinson’s disease (PD) is rigidity, whose assessment is important for monitoring the patient’s recovery. The wrist is one of the joints most affected by this symptom, which has a great impact on activities of daily living and consequently on quality of life. The assessment of rigidity is traditionally made by clinical scales, which have limitations due to their subjectivity and low intra- and inter-examiner reliability. (2) Objectives: To compile the main methods used to assess wrist rigidity in PD and to study their validity and reliability, a scope review was conducted. (3) Methods: PubMed, IEEE/IET Electronic Library, Web of Science, Scopus, Cochrane, Bireme, Google Scholar and Science Direct databases were used. (4) Results: Twenty-eight studies were included. The studies presented several methods for quantitative assessment of rigidity using instruments such as force and inertial sensors. (5) Conclusions: Such methods present good correlation with clinical scales and are useful for detecting and monitoring rigidity. However, the development of a standard quantitative method for assessing rigidity in clinical practice remains a challenge.


Introduction
Parkinson disease (PD) is one of the most common neurodegenerative disorders, affecting 2-3% of the population above 65 years old [1]. PD is characterized by the presence of Lewy bodies containing α-synuclein in dopaminergic neurons in the substantia nigra of the midbrain [2]. The resulting loss of dopaminergic neurons impairs voluntary and involuntary movement, as well as autonomic [3]. This accumulation of α-synuclein becomes more widespread in the brain during PD progression [4]. Motor disorders, slow initiation of voluntary movements, and increasing reduction in the speed and amplitude of repetitive activities (bradykinesia) are clinical signs of PD. Rigidity, resting tremor, or 1.
First, the title and abstract of the articles were read and analysed considering the inclusion criteria. This analysis was performed by two investigators. 2.
Next, the selected studies were read in full, performing a second filtering, and those that did not fit the inclusion criteria were excluded. This selection was performed by two researchers.

3.
Finally, data and characteristics were extracted, in addition to the quality evaluation of the selected studies.
The results are presented in Figure 1. Eliminating the duplicates, we have found 294 studies from 8 databases. From those, screening the studies by title and abstract, we evaluated whether the study met the inclusion criteria, that is, in addition to citing, it measured wrist rigidity in individuals with PD. Since, several articles cited wrist rigidity but the focus was to measure in another joint, as elbow [29]. Thus, we excluded 234 studies. In the third step, the full-text screening of the 59 studies was performed, and 28 studies were included for analysis. The exclusion of the 32 studies occurred for population criteria (3), outcomes (24), and study design (5) (Figure 1).

Data Analysis
Two authors (Alves C.M. and Rezende A.R.) extracted data from the selected articles. Disagreements between reviewers were resolved by consensus. If no consensus could be reached, a third reviewer (Marques I.A.) would decide. The information from the selected studies was organized in tables by method of evaluation, with each reference corresponding to a study. The information collected were: type of study, year of publication, sample size and characteristics and type of evaluation method.

Description of the Included Studies
The selected studies include 15 case series, 4 cross-sectional studies, 7 case-control and 2 cohort studies. The year of publication of the selected studies ranged from 1988 to 2021. Figure 2 depicts the number of studies published in each decade. There is an increase in wrist rigidity in Parkinson's disease research, particularly in the past decade, when 18 out of the 28 selected studies were published.

Data Analysis
Two authors (Alves C.M. and Rezende A.R.) extracted data from the selected articles. Disagreements between reviewers were resolved by consensus. If no consensus could be reached, a third reviewer (Marques I.A.) would decide. The information from the selected studies was organized in tables by method of evaluation, with each reference corresponding to a study. The information collected were: type of study, year of publication, sample size and characteristics and type of evaluation method.

Description of the Included Studies
The selected studies include 15 case series, 4 cross-sectional studies, 7 case-control and 2 cohort studies. The year of publication of the selected studies ranged from 1988 to 2021. Figure 2 depicts the number of studies published in each decade. There is an increase in wrist rigidity in Parkinson's disease research, particularly in the past decade, when 18 out of the 28 selected studies were published.

Data Analysis
Two authors (Alves C.M. and Rezende A.R.) extracted data from the selected articles. Disagreements between reviewers were resolved by consensus. If no consensus could be reached, a third reviewer (Marques I.A.) would decide. The information from the selected studies was organized in tables by method of evaluation, with each reference corresponding to a study. The information collected were: type of study, year of publication, sample size and characteristics and type of evaluation method.

Description of the Included Studies
The selected studies include 15 case series, 4 cross-sectional studies, 7 case-control and 2 cohort studies. The year of publication of the selected studies ranged from 1988 to 2021. Figure 2 depicts the number of studies published in each decade. There is an increase in wrist rigidity in Parkinson's disease research, particularly in the past decade, when 18 out of the 28 selected studies were published.  In the selected studies, several kinds of sensors were used to assess wrist rigidity in patients with PD ( Figure 3). Indeed, some studies used a combination of sensors to propose a method with greater reliability, but easy to use. The most common combination of sensors was goniometers (articular angles) and force sensors, as observed in five of the selected studies [13,[30][31][32][33]. An important point is that the results of the articles reviewed cannot be directly compared with each other, since there is great variability in the methods used by each group.
Healthcare 2022, 10, x FOR PEER REVIEW 5 of 16 In the selected studies, several kinds of sensors were used to assess wrist rigidity in patients with PD ( Figure 3). Indeed, some studies used a combination of sensors to propose a method with greater reliability, but easy to use. The most common combination of sensors was goniometers (articular angles) and force sensors, as observed in five of the selected studies [13,[30][31][32][33]. An important point is that the results of the articles reviewed cannot be directly compared with each other, since there is great variability in the methods used by each group.

Methods Evaluating Muscular Signals
Eight studies were included, which are summarized in Table 1. In three different studies, EMG was used in conjunction with goniometers (articular angles) and force sensors [34][35][36], two [31] used with force sensor only and one [37] with inertial sensors.
Although not all studies present correlation results with the clinical scales, electromyography (EMG) has been widely used to study the muscle physiology of patients with PD, showing good correlation with clinical scales that assess rigidity. Some studies have found a highly significant correlation between normalized stretch-related EMG and the degree of rigidity. Viscous damping constant had the best correlation (=0.77 ± 0.22) with clinical measures and the greatest number of limbs with significant positive correlation (57%) [35]. Clinical and device assessments that used EMG were significantly correlated, although the relationship was not strong (Spearman's rho = 0.382, p = 0.002) [36].
Two studies used the Myotonometer, a device that measures the biomechanical and viscoelastic properties of muscle [38,39]. The stretches were done manually by the examiner in both studies. Natural oscillation data are obtained using this method, which measures acceleration and "voltage" parameters. This technique estimates muscle tone (Hz) and rigidity (N/m) [40]. Both studies did not present significant positive correlations between clinical rigidity scores and viscoelastic stiffness values (before PWM-s: ρ = 0.12; p = 0.3; and after PWM-s: ρ = 0.284; p = 0.1).

Methods Evaluating Muscular Signals
Eight studies were included, which are summarized in Table 1. In three different studies, EMG was used in conjunction with goniometers (articular angles) and force sensors [34][35][36], two [31] used with force sensor only and one [37] with inertial sensors. During the "off" state, one examiner flexed and extended the wrist joint for 10 times and the measurements were taken for one second.
Effective deep brain stimulation and increased rigidity can significantly change viscoelastic rigidity in the resting muscles in patients with PD.
[34] 2016 Case series 18 PD patients EMG, articular angle and force sensor Participants were tested in the "on" and "off" states. Rigidity was quantified by rigidity work score and slopes of the moment-angle plots during flexion and extension.
The presentation of rigidity used in the determination of diagnosis, treatment and prognosis in PD will not be affected by the continuous nature of the movement trajectory used during clinical assessment.
[35] 2014 Case series 8 PD patients EMG, articular angle and force sensor Rigidity was evaluated during DBS surgery. The therapist performed six cycles of flexion and extension movement imposed on the wrist.
The system is capable of quantifying intra-operative stiffness and is useful for adjusting the position of the electrodes and the stimulation parameters. The study was carried out in "off" state. Rigidity was measured through Myotonometer. The examiner flexed and extend the joint 10 times.
The study supports the use of myotonometry for objective quantification of rigidity. The tool may be helpful for optimizing DBS settings in PD. [31] 2004 Case series 6 PD patients EMG and force sensor The motor moved the wrist joint in movement of flexion and extension. Patients were evaluated in "on" and "off" state.
The shortening reaction was reduced after the medication.
[37] 1992 Cross-sectional 9 PD patients and 10 healthy subjects EMG and inertial sensors EMG was recorded during the flexion and extension movements of the wrist joint. The patients were evaluated in "on" and "off" states.
EMG indicated that mild and moderate rigidity was consistently associated with increased stretch-related activity compared with non-rigid conditions.
Although not all studies present correlation results with the clinical scales, electromyography (EMG) has been widely used to study the muscle physiology of patients with PD, showing good correlation with clinical scales that assess rigidity. Some studies have found a highly significant correlation between normalized stretch-related EMG and the degree of rigidity. Viscous damping constant had the best correlation (=0.77 ± 0.22) with clinical measures and the greatest number of limbs with significant positive correlation (57%) [35]. Clinical and device assessments that used EMG were significantly correlated, although the relationship was not strong (Spearman's rho = 0.382, p = 0.002) [36].
Two studies used the Myotonometer, a device that measures the biomechanical and viscoelastic properties of muscle [38,39]. The stretches were done manually by the examiner in both studies. Natural oscillation data are obtained using this method, which measures acceleration and "voltage" parameters. This technique estimates muscle tone (Hz) and rigidity (N/m) [40]. Both studies did not present significant positive correlations between clinical rigidity scores and viscoelastic stiffness values (before PWM-s: ρ = 0.12; p = 0.3; and after PWM-s: ρ = 0.284; p = 0.1).

Methods Evaluating Capture of Movement
Eight studies were included, which are summarized in Table 2. Inertial sensors were used in four studies [41][42][43][44], in combination with EMG in one study [37], and combined with force sensors in two other studies [15,45]. The study [46] utilized the UPDRS. Three studies applied the evaluation during the DBS surgery [41,43,44], and applied a classification model to differentiate rigid and non-rigid states. The system was used during the DBS surgery. The rigidity was assessed by imposition of passive wrist flexion.
The polynomial models were tested for a larger number of patients. An accuracy of 78% and a rigidity classification error of 3.5% was achieved.
[45] 2017 Case series 4 PD patients Inertial sensors and force sensor The hand was passively moved by the examiner, who had a moment sensor attached to his hand. Moment arm was measured using a ruler.
The PowerGlove measured a difference in off-vs. on-medication condition in all tasks in the patients with most of its outcome parameters.
[44] 2016 Case series 17 PD patients Inertial sensors The system was used during the DBS surgery. The rigidity was assessed by imposition of passive wrist flexion.
The system correctly classified 82.0% of the signals (mean error of 3.4%), which supports the reliability of this solution.
[ The therapist moved the patient's wrist in flexion and extension while the signals were collected.
The device presented 89% of specificity, 82% sensitivity and great portability.
Despite the vast majority of studies using IMU sensor not showing correlation values with clinical scales, some devices were able to correctly classify 83.9% of the evaluated signals, compared to UPDRS [41]. Mann-Whitney rank order comparisons revealed that the median rigidity score for clinically classified rigid patients was significantly higher than the median control (U = 17, p < 0.0001) or the median rigidity score of the non-rigid patient (U = 15, p = 0.0001) [32].

Methods That Used Mechanized and Force Sensors
Seven studies were included, which are summarized in Table 3. In studies of Xia et al. [13,31] the stretching velocity was controlled at 50 • /s. Teräväinen et al. [33] proposed a comparison between different velocities and found that higher velocities were the most sensitive for detecting parkinsonian rigidity. Force sensors alone were in two studies [20,47], and combined with electromyography (EMG) analysis in three studies [16,48,49]. Powell et al. and Xia et al. [16,49] also performed in different velocities and found greater rigidity at higher speeds. PD subjects were assessed in the "on" and "off" states. The torque perturbation was applied five times, thus leading to an overall duration of the session equal to 105 s.
The device allowed to successfully estimate the rigidity, and allowed to discriminate both Healthy subjects from PD subjects, and PD subjects in "off" condition from PD subjects in "on" condition.
[47] 2019 Case series 5 PD patients Force sensors It was measured at the "on" and "off" states. The wrist was evaluated 3 times without and 3 times with the Froment maneuver.
A minimum-maximum limit forming study was done for rigidity values between 0-5 using both Froment rigidity values' scores. [30] 2018 Case control 18 healthy and 4 PD patients Force sensor The wrist joint was positioned in alignment with the centre of rotation of the plate. Ten repetitions of flexion followed by extension were conducted.
The device has sufficient accuracy and sensitivity to measure the interaction torque at the wrist joint and to differentiate PD rigidity from normal muscle tone.
[48] 2011 Case control 12 subjects with PD and 8 controls

Force sensor and EMG
A servomotor executed the flexion and extension of the wrist. EMG and torque were collected. The patients were evaluated in "on" and "off" states.
Subjects showed a higher resistance in "off" compared to healthy control. The medication reduced difference in torque resistance between controls and PD patients. [49] 2009 Case series 12 PD subjects Force sensor and EMG The servomotor executed the movements of flexion and extension and the velocity was constant. EMG was recorded at the flexors and extensors. Patients were evaluated in "on" and "off" state.
Torque resistance scores were more strongly correlated with the EMG ratios. The highest correlation was found between the torque resistance score and EMG during the extension at high velocity in "off" state.
[13] 2006 Cross-sectional 12 PD patients and 7 controls Force sensor Flexion and extension movements were carried out by a servomotor at a constant velocity. Patients were evaluated in "on" and "off" states.
The rigidity was more readily elicited in extension movements. Compared with controls the scores were higher for patients in "on" state. The individual's hand was oscillated passively. The objective rigidity score was expressed in Newton-meter-degrees.
The rigidity was more pronounced at faster movement velocities in patients with PD.
Most of the studies did not show comparison results with the clinical scales. However, it was observed that the p value obtained in the evaluation of the differences between PD OFF and PD ON using a device (p = 0.016) is lower than that obtained by the MDSUPDRS Part III score (p = 0.031). This may lead to the hypothesis of a greater sensitivity of the equipment in relation to the clinical score [30]. Spearman's correlation analysis showed mild to moderate correlations between device scores and UPDRS rigidity scores in the off-medication state (q = 0.482, p = 0.111 for the slowest movement; q = 0.578, p = 0.049 for the faster movement). In Medication Status, there was a low correlation between device and clinical degrees of rigidity for slower movement (q = 0.324, p = 0.305) and no correlation was found for rapid movement (q = 0.248, p = 0.437) [20].

Ordinal Scales
Five studies were included, which are summarized in Table 4. Despite the vast range of sensors utilized, five studies [50][51][52][53] employed ordinal scales to assess individuals. Although these scales are widely used in clinical settings, they have several drawbacks, such as relying on the therapist's experience, resulting in subjective and unreliable assessment. One study applied a new technique using quantitative digitography [54], in which the person performed movements on a keyboard and the application force was measured, at the end they were able to compare rigidity and tremor parameters with the UPDRS. There were significant associations between the rigidity sub-score and release slope (ρ = −0.43, p = 9.78 × 10 −9 ).

Applicability of the Methods in Clinical Practice
An extremely important factor in the application of new rigidity evaluation methods is their suitability in clinical practice and their applicability in different clinical scenarios.
In addition, it is necessary to analyse the methods that can be applied at home, since, even in times of a pandemic like COVID-19, this type of evaluation is important. In this sense, methods that can be adapted to be applied at home are more interesting.
Analysing the articles found in this review, the applicability of the methods in different clinical scenarios was evaluated, such as: outpatient appointments, infirmary recuperation evaluation, intraoperative DBS targeting, telemedicine and home evaluation. Figure 4 shows the applicability of the methods found, in different scenarios. It can be noted that there are still few devices that can be implemented in telemedicine, while the vast majority can be used in outpatient appointments. [53] 1988 Case series 20 PD subjects Ordinal scale wrist flexion and extension movements at varying speeds.
presented 80% of agreement and a weighted Kappa value of 0.636.

Applicability of the Methods in Clinical Practice
An extremely important factor in the application of new rigidity evaluation methods is their suitability in clinical practice and their applicability in different clinical scenarios. In addition, it is necessary to analyse the methods that can be applied at home, since, even in times of a pandemic like COVID-19, this type of evaluation is important. In this sense, methods that can be adapted to be applied at home are more interesting.
Analysing the articles found in this review, the applicability of the methods in different clinical scenarios was evaluated, such as: outpatient appointments, infirmary recuperation evaluation, intraoperative DBS targeting, telemedicine and home evaluation. Figure 4 shows the applicability of the methods found, in different scenarios. It can be noted that there are still few devices that can be implemented in telemedicine, while the vast majority can be used in outpatient appointments.

Discussion
This scoping review included 28 studies investigating the most used methods for wrist rigidity evaluation in individuals with Parkinson's disease. The most studies presented quantitative methods associated with UPDRS, which is a scale commonly used for this assessment in clinical practice [55]. It can also be noticed a greater concentration

Discussion
This scoping review included 28 studies investigating the most used methods for wrist rigidity evaluation in individuals with Parkinson's disease. The most studies presented quantitative methods associated with UPDRS, which is a scale commonly used for this assessment in clinical practice [55]. It can also be noticed a greater concentration of the selected studies in the past 10 years, which denotes an increasing demand for objective methods that present more reliable results [56]. One possible explanation for this increase in publications is the need for therapists to seek out and develop alternative methods that are intuitive and easy to interpret [56]. Furthermore, they must present data that allows for a more accurate tracking of the patient's progress.
Due to the subjective nature of clinical assessments and the small group of patients, it would be difficult to expect a direct correlation between UPDRS values and devices that measure rigidity. The low sensitivity and high variability of the clinical assessment may also explain the relatively modest correlation between the mechanical and clinical assessments of rigidity. Due to differences in the design of measuring devices, the studies are not directly comparable, however, the methods appear to allow researchers to objectively describe changes in stiffness in patients with PD. Most studies presented instruments with the possibility of obtaining objective results for detecting and quantifying rigidity compared to healthy individuals. In this sense, such instruments can be considered as useful tools in the assessment of wrist rigidity in people with PD.
The neural and non-neural components in determining rigidity are still being debated, even though both contribute to the unique nature of Parkinsonian rigidity. Non-neural components correspond to the mechanical properties of muscle fibers, whereas neural components, such as shortening reflexes, are inherent in pathophysiology [13,31,49]. These later aspects can be detected by EMG and force sensors [16,48,49]. Studies that used a combination of such sensors exhibited better results, with a good correlation with scales used in clinical practice. Furthermore, EMG and force measures are more representative of the neurological source of the motor symptoms [57].
On the other hand, other studies used inertial sensors and goniometers (joint angles) to assess wrist rigidity [41][42][43][44]. Inertial sensors are currently used for angular velocity measurement in rigidity assessment in PD [58]. The results in the cited studies indicate that the viscosity and elasticity of the muscle in the rigidity condition are higher than in the healthy condition (relaxed state) [42]. In this sense, the combination of data from inertial sensors and joint angles can detect the on-off fluctuations of parkinsonian rigidity. Range of motion is an indirect measure of the progression of rigidity in Parkinson's disease. In most cases, rigidity is accompanied by a decrease in joint range of motion [59].
Regarding the protocols used in the studies, some used different stretching speeds for the rigidity evaluation [16,33,37,49,52]. Some approaches suggest that parkinsonian rigidity is considered velocity-independent, in contrast to spasticity, which is velocitydependent [52,60]. However, more recent studies have raised questions about the accuracy of this point of view [6,16]. These studies concluded that rigidity depends on the angular velocity and articular amplitude of the applied mobilization and Parkinsonian rigidity is modulated by the amplitude and rate of muscle stretch. The differences between the experimental protocols leads us to different results, while some studies such Powell et al. and Xia et al. [16,49], applied and analysed different stretching speeds, others such as Meara et al. [52] applied different speeds but did not compare the results. These findings may help to understand the biomechanical foundations and physiological characteristics of rigidity and may assist in the development of new methods to assess clinical rigidity in Parkinson's disease.
The way in which passive extension/flexion is performed is a critical aspect of objective rigidity assessment. Servomotor manipulation was used in some studies [13,20,30,48,49], while manual manipulation was used in others [15,38,45,51]. When done manually, it is difficult to ensure that the therapist is stretching at the appropriate speed. As a result, studying the relationship between speed and rigidity is difficult.
A parameter of great relevance in this research is the incorporation of these devices in clinical practice, and its usability in different clinical scenarios. Although most of the devices found have not been applied in clinical practice, it is mentioned where they could be used. According to the analyses carried out, regarding the applicability of these methods, most devices can be applied in outpatient appointments, and few in the other scenarios. One of the reasons for this is the size of the devices presented, which often makes it difficult to carry and mount them in other environments, outside the clinical offices. In addition, to have a better view of these factors, it is necessary to perform some usability tests, in addition, the complexity of the proposed devices must be evaluated [61].
Another relevant parameter is the suitability of these devices for use in telemedicine, that is, for use by patients at home without the physical presence of therapists. Recent studies discuss the limitation of measuring rigidity through remote systems, since it is a parameter that must be evaluated personally, through physical contact with the patient. In this review we found three papers [20,43,54] that discuss this usability, and after some validation tests, the technology proposed by them could be implemented remotely as long as the patient was trained with the device.
Some knowledge gaps were found in this review. Comparative studies were not performed analysing the different instruments compiled in this review. It would be interesting to analyse the diagnostic accuracy to decide which of these methods is the gold standard for identifying stiffness in PD. Despite a recent increase in the number of studies developing new technologies for the quantitative assessment of rigidity, it has not yet been discussed how to incorporate them into clinical practice. More tests should be performed in clinical practice, to understand how these new devices would be incorporated during clinical evaluation. Another important aspect is the evaluation protocol, most articles apply passive flexion and extension of the wrist, but there is no consensus on velocity applied. It is also necessary to determine whether assessments should be performed in the "on" or "off" period of medication.

Main Limitations and Clinical Implications
The present study has some limitations. The quality of the studies was not examined as these factors are not included in a scoping review. This should be verified in future studies. Another limitation was that the review contained only studies that evaluated wrist rigidity, not considering other joints. This choice was made since there are few review articles focused on this articulation. Furthermore, the wrist is an essential component in performing the activities of daily living (ADLs) associated with a high quality of life. Furthermore, this study can help in the development of new technologies focused on the evaluation of wrist rigidity. Another limitation was the lack of quantitative data comparing the proposed methods with the clinical scales, which made it impossible to carry out a more complete statistical analysis of the methods.
Rigidity is one of the most common symptoms of Parkinson's disease, and the wrist joint is frequently affected. It is a joint that directly affects the ability to perform everyday tasks and the quality of life of people with the disease. The diagnosis and follow-up of patients rely heavily on the assessment of rigidity in clinical practice. When this evaluation is done quantitatively, it is possible to track patient treatment more accurately and test new medicines with greater safety. However, the primary concern is how to implement these new technologies into clinical practice. Thus, additional tests with existing technologies are required, with an emphasis on applications that are simple to use. Additionally, the results must be reliable and simple to interpret.

Conclusions
The main findings of this study showed that there is an increase in the number of studies focused on objective methods for assessing wrist rigidity in Parkinson's disease in the last few years. It was possible to observe an increasing search for methods that provide reliable results and that consider all aspects of rigidity. Several types of medical devices have been proposed, but none has been clinically validated or subjected to an industrial validation procedure. Therefore, additional analysis using existing methods is required.
It can be observed that none of the proposed equipment was incorporated into clinical practice. Probably due to the ease of applying clinical scales, and also the lack of more precise studies with these new technologies. It is necessary to apply the proposed methods in a larger number of individuals and carry out tests during procedures performed in the clinic. Another aspect that should be better determined is the protocol to be applied in this evaluation.

Data Availability Statement:
The datasets generated in the current study are not publicly available due to the ethical restrictions preventing public sharing of data. A non-identified set may be requested after approval from the Review Board of the Institution. Requests for the data may be sent to the corresponding author.