The Way You Draw Can Reveal Parkinson’s Disease With Up to 99% Accuracy, Study Reveals

Researchers have developed an artificial intelligence system that can identify Parkinson’s disease based on drawing patterns with remarkable precision. Scientists at Siksha ‘O’ Anusandhan University in India created the breakthrough diagnostic tool by analyzing how participants completed simple drawing exercises using specially equipped pens that tracked hand movements and pressure.

The study involved 66 participants, including 31 individuals diagnosed with Parkinson’s and 35 healthy controls, who traced spiral and meandering line patterns. The smart pens recorded multiple data points including grip strength, pressure, tilt, and acceleration while simultaneously capturing the visual appearance of the drawings. When researchers fed this dual information into advanced machine learning systems, the AI achieved detection rates of 98.95 percent accuracy for meander patterns and 97.74 percent for spiral patterns.

The system’s success stems from its ability to identify both visual irregularities in the drawn shapes and subtle motor control differences in hand coordination. Parkinson’s disease, a fast-growing neurological condition affecting movement worldwide, currently relies heavily on clinical assessments for diagnosis, creating challenges for early detection. This non-invasive screening approach offers potential for more accessible and affordable testing.

While researchers acknowledge their limited sample size doesn’t fully represent the disease’s complexity across larger populations, they view the findings as a significant step forward. The team suggests that with further validation through larger clinical trials, their framework could eventually support earlier diagnosis and remote neurological screening for Parkinson’s and similar conditions.

Originally reported by Sciencealert →