Researchers in Belgium have discovered that Parkinson’s disease may consist of several hidden subtypes rather than one single disorder.
Using artificial intelligence and machine learning, scientists identified different biological forms of Parkinson’s disease that respond differently to treatments.
The findings, published in Nature Communications, could eventually help doctors develop more personalized therapies for patients.
Parkinson’s disease is a progressive brain disorder that affects movement and many other body functions. It develops when nerve cells involved in movement and coordination become damaged over time.
Common symptoms include tremors, muscle stiffness, slower movement, balance problems, and difficulty walking. As the disease worsens, some patients may also experience memory problems, mood changes, sleep disorders, and cognitive decline.
Although Parkinson’s disease has been studied for decades, doctors still do not fully understand why it develops differently in different people.
Some patients decline slowly over many years, while others experience faster progression. Some respond well to medications, while others gain limited benefit. These differences have puzzled scientists for a long time.
Researchers have also discovered many different genes linked to Parkinson’s disease. However, understanding how these different genetic mutations affect the disease has been extremely difficult.
The new study was led by scientists from VIB and KU Leuven, including Professor Patrik Verstreken and researcher Dr. Natalie Kaempf.
Instead of starting with fixed assumptions about how specific mutations should behave, the researchers allowed artificial intelligence to search for patterns directly from experimental data.
The team studied fruit fly models carrying mutations in 24 genes associated with Parkinson’s disease. Fruit flies are widely used in neuroscience because their nervous systems share many important similarities with humans.
Researchers monitored the flies’ behavior over long periods of time and collected large amounts of data.
The information was then analyzed using machine learning algorithms.
Machine learning is a form of artificial intelligence that allows computers to detect hidden patterns in complex data. It is increasingly being used in medicine because it can uncover relationships too complicated for traditional analysis.
The machine learning system identified two major groups of Parkinson’s disease that could be divided further into five smaller subgroups.
This means that patients who appear similar clinically may actually have very different biological problems occurring inside their brain cells.
Professor Verstreken explained that Parkinson’s symptoms may look similar on the surface, but underneath, the molecular mechanisms driving the disease can differ greatly.
This discovery may help explain why many Parkinson’s treatments fail to work consistently across all patients.
If different patients have different biological forms of the disease, then a single treatment may not be effective for everyone.
The researchers tested this idea by giving experimental compounds to different Parkinson’s subgroups in the animal models.
The results were striking.
A drug that successfully improved symptoms in one subgroup did not work in another subgroup. This suggests that future therapies may need to be tailored to the patient’s specific disease subtype.
Researchers believe this could eventually lead to precision medicine for Parkinson’s disease.
Precision medicine aims to customize treatment based on a person’s genetics and biological characteristics instead of using the same therapy for all patients.
The scientists now hope to identify biomarkers linked to each Parkinson’s subtype. Biomarkers are measurable biological signs that may help doctors determine which type of disease a patient has.
If successful, doctors may one day use genetic testing or biological markers to match patients with treatments most likely to help them.
The findings could also improve drug development. Pharmaceutical companies have struggled for years to develop successful Parkinson’s treatments because patients with different underlying disease mechanisms were often grouped together in clinical trials.
Separating patients into more precise subgroups could make future clinical trials more effective and improve the chances of finding successful therapies.
The study also demonstrates how artificial intelligence may transform disease research.
Traditional medical research often relies on scientists starting with a theory and testing it step by step. In this study, researchers instead used an unbiased machine learning approach that allowed hidden patterns to emerge naturally from the data.
This approach uncovered structures within Parkinson’s disease that might never have been recognized using conventional methods.
Researchers believe the same strategy may also help scientists better understand other complex diseases caused by multiple genes or environmental factors.
Conditions such as Alzheimer’s disease, autism, diabetes, and some psychiatric disorders may also contain hidden biological subtypes that respond differently to treatments.
At the same time, researchers caution that the findings are still at an early stage.
Most of the current work was performed in animal models rather than human patients. More studies are needed to confirm whether the same subtypes exist clearly in people and whether personalized therapies can improve patient outcomes.
Even so, experts say the findings are highly important because they challenge the traditional idea that Parkinson’s disease is one single disorder.
Instead, the disease may actually be a collection of related conditions that require different treatment approaches.
The discovery could eventually lead to earlier diagnosis, better-targeted medications, and more effective therapies for millions of people living with Parkinson’s disease worldwide.
If you care about Parkinson’s disease, please read studies that Vitamin B may slow down cognitive decline, and Mediterranean diet could help lower risk of Parkinson’s.
For more health information, please see recent studies about how wheat gluten might be influencing our brain health, and Olive oil: a daily dose for better brain health..
The findings were published in Nature Communications.
Source: KU Leuven.
