Young scientists from the National Research Nuclear University MEPhI, in collaboration with colleagues from the Moscow State Technical University and the Scientific Center of Neurology, have developed a method for noninvasive diagnosis of Parkinson's disease. Now you can learn about this serious disease even before the symptoms appear – tremor, slowness of movement, cognitive impairment. Parkinsonism can be detected not by examinations and tests, but by... light. Nikita Baynaev-Mangilev, a graduate of MEPhI, spoke about the essence of the new method.

The interview was conducted under the heading "Voice of Science".

– Very soon, on April 11, the World Day against Parkinson's Disease will be celebrated. Nikita, how urgent is the problem?

– The date of April 11 was chosen not by chance, but in honor of the birth of the discoverer of the so–called tremor palsy, the English physician James Parkinson. He described this neurodegenerative disease back in 1817. It was not possible to defeat the disease. Moreover, nowadays there is a sharp increase in the number of people with this disease all over the world. Despite the fact that 85% of patients are over 65 years old, there is a form with an early onset (before the age of 40), and even a juvenile form of the disease (with a debut at the age of 20). For example, the famous actor Michael J. Fox, who played the main character in the movie "Back to the Future," was diagnosed with Parkinson's disease at the age of 35. The disease is considered incurable. However, with early diagnosis of Parkinsonism and proper therapy, patients can be helped.

– And you are currently working on a method that will detect the disease at an early stage?

– Yes, my colleagues and I from Bauman Moscow State Technical University and the Scientific Center of Neurology are engaged in non-invasive diagnostic methods for various diseases, in particular, we are currently testing the method for Parkinson's disease.

– What is the essence of this method?

– Our method is based on fluorescence spectrometry of scattered light. We irradiate skin cells at a specific wavelength. Different internal molecules (fluorophores) absorb laser light and emit at a different wavelength. By registering this light, we can determine the presence and relative ratio of fluorophores. It is known that in Parkinson's disease, processes in cells are disrupted. It is believed that the cause of neuronal death is the protein alpha-synuclein, which forms aggregates in Parkinson's disease, destroying cells. Since alpha-synuclein is present in the epidermis and disrupts cell function, the spectral profile of the skin also changes. We can use it to distinguish sick people from healthy people.

– Thanks to your method, can doctors identify the disease, or does machine learning come to the aid of doctors?

– Although the spectral profiles of sick and healthy people differ, it is almost impossible to determine the presence of the disease by eye. This is where machine learning comes to the rescue. Using the output spectra, we teach the model to find differences in these spectra. In addition, we have developed a user-friendly device that will be able to determine the probability of having a disease by itself, so you can't do without a machine learning model.

– Is your method unique and what is its advantage?

– Today there are several methods for determining Parkinson's disease (for example, by the clinical picture), but they are effective mainly in the late stages of the disease, when a significant part of the neurons are already affected. Our method, on the contrary, involves early diagnosis, long before the first symptoms appear. This will allow you to provide timely assistance.

By no means do we claim to replace other methods, but we assume that our method will complement the others. It is non–invasive and very fast - the study using our method takes no more than a minute. This is much faster than the same MRI or transcranial sonography also used to diagnose Parkinson's disease. In addition, due to the availability of components, our solution will be quite cheap, which is also an advantage over other methods.

Fluorescence spectroscopy is already used in medicine. For example, there is a device that monitors blood and lymph flow, evaluates oxidative metabolism using fluorescence spectroscopy, and is designed to monitor changes in the condition of tissues in diabetes. The main difference between our development is that we do not use key points, but compare spectral profiles entirely within the area of interest, which provides more information. Also, our device is focused not only on monitoring, but also on detecting the disease. This is its uniqueness.

– Is there already a prototype of the device?

– Yes, of course. In 2025, we created the first small series of devices. The devices include all the basic functionality that is expected when a full batch is released. At the moment, due to the limited experimental data, this series has been sent to our colleagues from the Scientific Center of Neurology to collect spectra for training the ML model. But even with the existing sample (~130 people), it was possible to obtain an accuracy of more than 80%. We expect that with an increase in the database, we will be able to increase the accuracy to 90% or higher.

– What are the prospects for this project?

– Our main task today is to bring the first series of devices to readiness by collecting a sufficient database and training the model on this data. Although we are currently focusing on Parkinson's disease, in general, this method is applicable not only for it, but also potentially suitable for general screening of skin cells. I believe that the main prospect is to expand the scope of the method in case of diagnosis of various diseases or conditions of the body as a whole. In addition, another direction may be to complicate the ML model by providing additional data from other analyses, for example, a general urine and blood test. This could potentially help improve diagnostic accuracy and the ability to differentiate between different diseases with similar spectral profiles.

Interviewed by Natalia Sysoeva