Parkinson’s disease is a debilitating, and still incurable, neurological disease. While there are no reliable blood or laboratory tests for diagnosing Parkinson’s, new research using advanced computer-based screening methods is providing new hope for earlier identification and treatment of the disease. One especially exciting new study (Yang et al., 2022, in Nature Medicine) applies artificial intelligence to detect Parkinson’s based on nocturnal breathing patterns.
Parkinson’s symptoms can be similar to other nervous system disorders, which further complicates diagnosis. It’s also hard for doctors to determine how bad Parkinson’s will be for each patient. Therefore, diagnostic methods which can detect early-stage disease are needed.
New AI testing methods, traditional imaging approaches, and advances in wearable devices and virtual reality are now all being leveraged for Parkinson’s diagnosis and treatment.
What is Parkinson’s disease and the current state of diagnosis?
By many reports, Parkinson’s is the fastest-growing nervous system disease worldwide, with prevalence doubling in the last quarter century. Parkinson’s is caused by the breakdown or death or nerve cells (neurons) in a part of the brain called the substantia nigra. This part of the brain controls movement, and Parkinson’s patients experience abnormalities such as uncontrollable shaking. While medication may lessen the symptoms, Parkinson’s patients struggle with everyday tasks and maintaining quality of life.
There are currently no laboratory tests that can detect Parkinson’s. Neurologists diagnose the disease through a review of the patient’s medical history and a physical exam. They commonly use the MDS-Unified Parkinson's Disease Rating Scale (MDS-UPDRS) and look for symptoms such as shaking of the hands and limbs, stiff muscles, and loss of coordination.
Physicians also have to rely on the patient to self-report Parkinson’s symptoms. Movement difficulty occurs several years after Parkinson’s onset, which means diagnosis in this way often only detects advanced cases of Parkinson’s. Much earlier identification would be ideal.
Doctors may also order imaging tests and prescribe medication to treat Parkinson’s. Improvement of symptoms with the medication can help verify presence of the disease, and rule out other conditions.
Clearly, both diagnosis and treatment are imprecise, making the disease even more elusive, frustrating, and damaging. All these factors underscore the importance of finding more effective forms of handling Parkinson’s, starting with diagnosis.
Exciting new ways of diagnosing Parkinson’s
State-of-the-art technology continues to be researched for all aspects of Parkinson’s. Researchers are increasingly finding hope not just in medical technology such as imaging and laboratory testing, but also, as in Yang et al., 2022, in widely trending areas such as AI, virtual reality, and wearables.
Artificial intelligence
Artificial intelligence (AI) is the development of computer systems that perform tasks that typically require human intelligence, such as visual perception, speech recognition, decision-making, and natural language understanding. By utilizing algorithms, machine learning, and data, AI enables machines to learn, adapt, and evolve autonomously, ultimately increasing their ability to solve complex problems, automate tasks, and augment human capabilities.
Computers can be trained on large amounts of data to make “intelligent” decisions. For example, AI can be trained on sample videos of many people walking to the point where it will be able to evaluate if a person’s walk is unnatural or indicative of something wrong. This type of logic lets scientists and doctors use AI as a powerful tool to detect diseases.
Changes in breathing and lung function have been noted in Parkinson’s patients. Yang et al., 2022 showed how AI can potentially be used to successfully diagnose Parkinson’s and determine how the degree to which the disease will progress. The AI in this study was trained on recordings of night-time breathing patterns from more than 7,000 patients. After training, the AI was then tested on a different set of patient data.
The algorithm could accurately tell Parkinson’s patients from healthy controls. It was able to correctly detect over 80% of Parkinson’s patients. The AI was also effective at correctly detecting healthy patients. Distinguishing Parkinson’s patients apart from healthy controls is a chief measure of the AI’s ability. Importantly, the algorithm could also accurately predict Parkinson’s severity.
Other computer-based approaches have previously been employed for Parkinson’s diagnosis, with varying degrees of success. This includes an approach funded in part by the Michael J. Fox foundation. The approach, called PreciseDx, uses AI to look at patients’ tissue samples for the presence of Lewy bodies, which are a biological marker of Parkinson’s.
Biomarkers
Biomarkers are measurable indicators of normal or abnormal processes in the body. They can be used to diagnose disease, monitor disease progression, and predict the response to treatment. In recent years, there’s been increasing interest in using biomarkers to detect Parkinson's disease.
There are many potential biomarkers for Parkinson's disease, but only a few have been well studied. Among those suggested are changes in the lipid portion of patient blood and the presence of Parkinson’s-specific proteins in blood and brain fluids.
There are also no established protocols on the gold standard for sensitivity and specificity of a new biomarker. This leaves these areas of research in need of attention and breakthroughs in addition to the previous examples.
Imaging techniques
Imaging techniques in the context of Parkinson’s include many traditional brain imaging techniques such as magnetic resonance imaging (MRI) and positron emission tomography (PET). These methods allow doctors to see molecular, structural, and functional changes within a patient’s brain. Many of these changes may indicate Parkinson’s.
The eye is a new imaging target for Parkinson’s diagnoses. During Parkinson’s, physical changes in the eye can occur in the disease’s early stages. Retinal optical coherence tomography (OCT) is not painful or invasive, but it lets clinicians see the structural changes in patients’ eyes. Newer OCT technology may enable both diagnosis of Parkinson’s at earlier stages and prediction of disease severity.
Wearable devices
Wearable devices (wearables) comprise a category of electronic devices that can be worn as accessories, such as smartwatches, fitness bands, and belts. They integrate advanced sensors and connectivity, offering features like fitness tracking, health monitoring, and communication.
They can be used to collect patient data and transmit it wirelessly. Health status data such as blood glucose level, respiration, and heart rate can all be detected using wearable devices. Accuracy of such readings remains a work in progress, but advances have been fast and impressive.
Doctors can potentially diagnose Parkinson’s by combining wearable data with AI. in Yang et al., 2022, a wearable respiration monitor called a breathing belt was used. This belt, worn across the abdomen, collects nighttime breath data from patients. Sensors in the belt detect and record the rise and fall of the abdomen as the wearer breathes. This provides a measure of the smoothness and fullness of each breath. Since breath changes are associated with Parkinson’s and happen at early disease stages, this data can be used diagnostically.
Virtual reality
Virtual reality (VR) is a computer-generated visual (and sometimes audible) simulation of a real environment that viewers/users can enter and interact with. To date, using VR usually requires users to wear special goggles with a screen inside.
Users must wear special equipment to use VR spaces. This equipment usually includes a helmet with a screen inside. Gloves or handheld controllers to track hand movement may also be worn. In many VR setups, cameras which can track the user’s body as they move, are also used.
VR provides another promising option for early and accurate Parkinson’s diagnosis. It lets users to enter customized virtual spaces that can mimic activities of everyday life. Importantly, VR devices allow for precise measurements of hand and head motion. As hand and head tremors are common in Parkinson’s, VR is a useful tool for judging movement problems.
A recent report showed that the use of VR screening before imaging techniques could predict motor changes. Similar studies using VR-based testing to assess Parkinson’s have shown that VR screening can detect changes in brain function among Parkinson’s patients at earlier stages of disease.
VR may also be useful for Parkinson’s treatment and therapy. The ability to perform daily activities in a controlled environment can be used for Parkinson’s treatment. Indeed, studies suggest that VR training may be just as effective as traditional movement training for Parkinson’s patients. In particular, VR training has been demonstrated to enhance hand function and balance in Parkinson’s patients.
A promising future for Parkinson’s diagnosis
The diagnosis of Parkinson's disease currently relies on clinical evaluation and neurological examinations. While there’s no specific test to confirm the disease, ongoing research is exploring the potential of biomarkers, imaging techniques, and as seen in Yang et al. (2022), AI and wearables. This array of technologies give hope for earlier and more accurate detection.
Continued advancements in medical technology and research bring optimism for improved diagnostic tools and treatment options, ultimately leading to better outcomes and quality of life for those living with Parkinson's disease.
What’s your exciting research and how can we help?
Are you using new technology to treat diseases, conquer social issues, and change the world in other exciting ways? We’re here to support you through the entire research cycle, from finding literature gaps to editing to publication and expanding your impact. View our full range of author-support services. No matter what you’re challenge, our experts can assist. See Edanz’s Researcher Services and Editing Services.
Author bios:
|
Ebony Gary, PhD, was the main author of this post. She is a post-doctoral research fellow in the Weiner Laboratory at The Wistar Institute (Philadelphia, PA, USA), where she is part of a team working on novel DNA vaccines and treatments for cancer and emerging infectious diseases. Ebony earned her doctorate from the Drexel University College of Medicine. Her research interests include vaccine design, viral pathogenesis, mucosal immunity, tolerance-breaking cancer immunotherapy approaches, and vaccine-induced immunity in the aged.
|
|
Adam Goulston, PsyD, MBA, MS, MISD, ELS, planned and edited this post. Adam is a US-born, Japan-based science copywriter, editor, and marketer. He is a former in-house Senior Language Editor at Edanz and runs the scientific marketing company Scize Group.
|