2. Application Areas
Nocturnal Hypokinesia |
Facial expression |
|
Handwriting |
Breath |
|
Saccades |
Cortical activity |
|
Figure 1. The number of research publications between 2008 and 2021 that measure each sign and symptom.
Wearable and mobile technology can help medical practitioners assess the progression of PD patients by tracking their symptoms. These symptoms, referred to as cardinal PD features, are divided into motor symptoms and non-motor symptoms, as shown in
Table 1. The motor symptoms are the movement impairments due to PD. They have traditionally been associated with tremor, rigidity, bradykinesia, and postural instability. Additionally, abnormalities in limb movements, such as hand rotation, finger tapping, and arm angle, have been regarded as motor symptoms. In contrast, non-motor symptoms include a large variety of cardinal features, such as sleep disturbance, cognitive activity, fatigue, dementia, and psychiatric impairments, as listed in the second column of
Table 1. Finally, mixed symptoms, such as speech and swallowing difficulties, involve a combination of motor and non-motor activities. To show the prominence of these symptoms and justify their inclusion in
ourthe analysis,
Figure 1 shows the number of articles that measure the different motor and non-motor symptoms. Gait abnormalities, including FoG, are the most common symptoms measured, followed by tremor. Movement problems, such as bradykinesia and dyskinesia, are also common symptoms in PD patients. Many research articles use modern measurement techniques to monitor these symptoms.
Our Previ
ew and previous surveys of the PD literature
[27][29] show that monitoring motor and non-motor symptoms can be helpful in four primary application areas in the PD research context:
diagnosis, prognosis and monitoring, predicting response to treatment, and rehabilitation.
The rest of this section reviews these areas.
PD diagnosis relies on neurologists’ clinical assessment of motor and non-motor symptoms
[30]. To this end, neurologists observe patients while they perform specific tasks and assign them a score according to one of the standard scales: the previously mentioned UPDRS
[22], its updated version, the Movement Disorder Society-sponsored revision of the UPDRS (MDS-UPDRS)
[23], or the Hoehn and Yahr scale
[24]. The clinical information derived from these rating scales is subjective. Hence, it leads to inter-rate variability and also intra-rate variability
[27][29].
Moreover, as previously mentioned, the equipment used to supplement a clinical assessment tends to consist of expensive imaging tools, including MRI, SPECT, and PET. Because early and accurate diagnosis is essential, medical practitioners hope to augment current techniques with objective and cost-effective alternatives enabled by wearable and mobile technology. Therefore, recent research has examined the possibility of using wearable sensors and other portable technology to diagnose PD. These devices can provide objective PD-diagnosis measures that help standardize assessments
[31][32][33]. Many researchers have also been able to use wearable sensors and other devices to distinguish PD patients from healthy controls in lab experiments
[31][32][34].
In summary, wearable and mobile technology is used in the following sub-categories of PD diagnosis:
-
Early diagnosis of patients with PD;
-
Detecting PD symptoms in people with untreated PD;
-
Distinguishing PD patients from either healthy controls or patients with non-PD neurological disorders;
-
Distinguishing PD-related symptoms from similar symptoms not caused by PD (e.g., distinguishing PD tremors from essential tremors).
3. Prognosis and Monitoring the Severity of Symptoms
Predicting the patient’s future condition (prognosis) and assessing the severity of the disease, including current symptoms, are two activities that depend primarily on clinicians’ judgments and patients’ diary- or memory-based feedback. Clinicians’ judgments are subjective
[27][29], while patients’ diary entries and memories are limited by compliance and recall bias
[25][35][36]. Because clinical approaches that rely on these sources of information may not be completely reliable, the objective remote monitoring of PD symptoms holds immense promise for assessing the disease progression, evaluating symptom severity, and monitoring PD patients in unsupervised environments. To address these issues, recent work on PD prognosis and monitoring has focused on the following areas:
-
Home-based or remote monitoring of PD patients;
-
Evaluating the PD progression in a diagnosed patient;
-
Evaluating the severity of PD symptoms for a diagnosed patient.
4. Predicting Response to Treatment
To measure the efficacy of a treatment, clinicians rely on patients’ diary entries and unrecorded memories, both of which, as noted, can be subjective and unreliable. This problem has motivated researchers to measure the effects of treatments, including the effectiveness—and the side effects—with which medications suppress PD symptoms. To this end, researchers studying the ability of practitioners to accurately predict a patient’s response to treatment have been addressing the following issues:
-
Measuring the effects of such treatments as deep brain stimulation on the suppression of patients’ symptoms over time;
-
Measuring the intended effects of medications on patient symptoms;
-
Measuring the unintended effects (i.e., the side effects) of medications on patients (e.g., levodopa-induced dyskinesia).
5. Rehabilitation
Physiotherapy and other rehabilitation techniques are among the most common treatments for PD and, indeed, for movement disorders in general. As with medications, it is crucial to assess the efficacy of rehabilitation techniques. Additionally, it has been observed that cues and feedback are beneficial in assisting a PD patient. A rehabilitation system can provide auditory, visual, or haptic cues to facilitate a patient’s movement. Such a system can be explicitly used for gait training and assessing limb movements. Vibration-based actuators and audio feedback help patients suffering from rigidity, FoG, tremors, and other symptoms. These approaches can also help patients break out of a freeze and even suppress symptoms. Mobile technologies that target therapy and rehabilitation can be divided into the following sub-categories:
-
Audio, visual, or haptic cue for gait or movement training;
-
Sensory feedback to suppress a symptom such as FoG or tremor.
Motor Signs and Symptoms |
Non-Motor Signs and Symptoms |
Mixed Signs and Symptoms |
Gait |
EEG abnormalities |
|
Limb movements |
Cognitive activity |
|
EMG abnormalities |
Depression |
|
FoG |
Dementia |
|
Tremor |
Heart rate |
Speech topics |
Activities of Daily Living (ADL) |
Emotions |
Swallowing |
Bradykinesia and Dyskinesia |
Fatigue |
|
Posture |
Sleep topics |
|
Balance |
Blinking |
|
Nocturnal Hypokinesia |
Facial expression |
|
Handwriting |
Breath |
|
6. Technology in Parkinson’s Disease Research
O
ur review and other surveys of the PD literature
[27][29][37][38] reveal that researchers use a number of device types and technologies in the application areas
presented in the previous section. Using the current literature,
weresearchers classify these devices into the following eight categories based on the device form factor and sensing modalities.
6.1. Wearable Devices
Many recent PD-oriented approaches focus on wearable devices for health monitoring since they help medical practitioners record patients’ activities and symptoms. Wearable devices are ideal monitoring technologies for two main reasons: they are not limited to a specific location and can be easily integrated into patients’ clothes
[39]. The most commonly used wearable devices contain sensors, such as IMUs. The outputs of these sensors can facilitate analyses of body movement, gait, and symptoms, such as tremors.
OurThe study covers the following wearable technologies:
-
IMUs with integrated accelerometers, gyroscopes, and magnetometer sensors;
-
Insole force- or pressure-based sensors that can measure the ground reaction force (GRF);
-
Wearable devices that use sensors such as EEG and EMG to measure neural responses and muscle activities;
-
Clothing-integrated sensors, such as strain or accelerometer sensors, that measure hand tremors;
-
Other wearable devices, including smart glasses and smart hats, that can record patients’ emotions and other specific parameters.
Saccades |
Cortical activity |
|
Table 1. The signs and symptoms assessed by technology in the current literature. The signs and symptoms assessed by technology in the current literature.
Motor Signs and Symptoms |
Non-Motor Signs and Symptoms |
Mixed Signs and Symptoms |
Gait |
EEG abnormalities |
|
Limb movements |
Cognitive activity |
|
EMG abnormalities |
Depression |
|
FoG |
Dementia |
|
Tremor |
Heart rate |
Speech topics |
Activities of Daily Living (ADL) |
Emotions |
Swallowing |
Bradykinesia and Dyskinesia |
Fatigue |
|
Posture |
Sleep topics |
|
Balance |
Blinking |
|
6.2. Biopotential Devices
These devices can measure the electrical signals (biopotentials) generated by
ourthe body’s physiological processes. This category includes electroencephalogram, electrocardiogram, magnetoencephalogram, and electrooculogram devices. They can be either standalone devices, such as a 16-lead EEG system, or wearable devices with a single-lead EMG sensor. When a biopotential device is integrated into a wearable device, the resulting device falls under both the wearable and biopotential-device categories.
6.3. Cueing Devices
Devices that give feedback or cues to patients to rectify their walk or assist with other movements fall under this category. Headphones or speakers can deliver auditory cues, and visual cues can appear on a screen, smart glass, or even in a virtual-reality environment. Vibration sensors and electrical stimulators also can give sensory feedback to patients.
6.4. Optical Motion Tracker
These devices track user motion with either radio frequency or optical signals. For example, motion-capturing systems such as Microsoft Kinect
[40] and Vicon 3D
[41] use structured light for analysis. Similarly, radar-based systems use radio-frequency signals to monitor the motion of subjects
[42][43]. A system of multiple IMUs, even camera-based 3D setups, can also capture a patient’s movement.
6.5. Audio Recording
Many PD patients experience issues with swallowing or speech, as shown in
Table 1. Audio recording devices can analyze and monitor symptoms related to swallowing and speech. For instance, microphones and smartphones record speech tasks or patients’ voices for the subsequent diagnostic analysis
[34][44].
6.6. Video Recording
Video-recording devices are helpful for PD monitoring because gait and motor issues are among the most common PD symptoms. For instance, video cameras record patients’ movements at home or during laboratory experiments, and practitioners can use the recordings to spot symptoms or to corroborate or disprove predictions based on other devices
[6][20][45].
6.7. Force/Pressure
Force or pressure plates measure the force exerted by the patients’ feet while walking. Therefore, they help measure the gait quality of the patient. Force or pressure sensors are also integrated into gait mats to measure the gait quality of a patient.
6.8. Smartphone
Researchers have used accelerometers, gyroscopes, magnetometers, GPS, and other sensors on smartphones for various analytical objectives. Smartphone applications can also record such matters as patients’ emotional displays and drug dosages. Additionally, smartphone screens can record handwriting, microphones can record speech, and cameras can record movement.
6.9. Other
Other devices have also helped practitioners assess PD patients.
WResearche
rs classify these devices that do not fall into the above categories as “other.” For example, digitized tablets can serve as smart screens for handwriting assessment. PD patients can write on these screens during spiralography exams. Likewise, smart pens can record hand movements during writing. Solutions based on virtual or augmented reality have also started becoming popular in PD applications.