
At the Ronald Reagan UCLA Medical Center, a patient with Parkinson’s disease sat at a desk in front of a computer, pen in hand. But this was no ordinary pen. You can’t use it to write on paper. Instead, it’s loaded with a special magnetic ink that can record streams of current, undetectable to the naked eye, but loaded with neurological meaning.
The patient traced spirals, loops, and the letters “M-E-G-P-E-N,” while the pen transformed each flick of the wrist and tremble of the hand into electric signals. That information traveled to a neural network (an AI), trained to detect the subtle irregularities in motion that may betray one of the world’s fastest-growing brain disorders.
With more than 10 million people affected globally and no definitive test, Parkinson’s disease often remains hidden until symptoms grow too disruptive to ignore. The UCLA team behind the study believes they may have built something truly great: a 3D-printed pen, low-cost and power-free, that might be able to detect early Parkinson’s in a patient’s handwriting.
The Science Inside the Stylus

Most diagnostic tools for Parkinson’s rely on clinical observation or expensive biomarker analysis requiring lumbar punctures or specialized scans. This new device aims to change that.
At its core, the pen merges two technologies: a soft, magnetoelastic tip and an ink reservoir filled with ferrofluid — tiny magnetic particles suspended in liquid. As the user writes, the pen’s tip deforms under pressure, changing its magnetic field. Simultaneously, the ferrofluid ink swirls and shifts. Together, these movements induce a voltage in a coil inside the pen barrel through a physical principle known as the magnetoelastic effect.
As a result, each twitch and tremor generates a unique electrical signal.
“We are using the handwriting-generated electrical signal to quantify the tremor during [writing],” Prof. Jun Chen, the senior author of the study and a bioengineer at UCLA, told The Guardian. “It is very cost-effective and fully accessible for lower-income countries.”
The pen’s tip, made of silicone and neodymium particles, behaves like a flexible magnet. The ink doesn’t need to mark a surface. It just needs to move — whether the pen is pressed on paper or swirling in midair. And this, researchers argue, is key.
<!– Tag ID: zmescience_300x250_InContent_3
–>
While traditional digital pens mostly track the trace, what matters here is the movement. This new device picks up on the tremor itself, not just the handwriting, according to Gary Chen, the study’s lead author and a doctoral student in Chen’s lab.
Training the Pen to Recognize Disease
In their pilot study, researchers enrolled 16 people — 13 healthy participants and 3 with Parkinson’s. Each person completed three tasks using the pen: drawing wavy lines, spirals, and writing six capital letters (M-E-G-P-E-N). They did this both on paper and in the air.
The team fed the resulting electrical signals into several types of machine learning models. A convolutional neural network performed best, correctly distinguishing Parkinson’s patients from healthy participants with 96.22% accuracy.
This AI wasn’t reading the shapes on the page. It was analyzing signal features — spikes, peaks, and minute irregularities that corresponded to tremors, stiffness, and slowness. Patients with Parkinson’s showed distinctive patterns, including “minor peaks” during writing that were not seen in healthy controls.
Importantly, the model showed promise even with a small dataset. That suggests a scalable future where crowdsourced data might help refine diagnostics.
Private, At-Home Testing?
The diagnostic pen is designed with mass production in mind. All its parts — from the 3D-printed casing to the replaceable ink reservoir — can be manufactured at low cost. Its components are resilient to temperature, perspiration, and prolonged use. The signals remain stable even after 10,000 writing cycles.
It’s also privacy-conscious. Unlike video-based motion tracking systems, this pen doesn’t record identifiable data. No names, no faces — just hand movements.
Becky Jones, research communications manager at Parkinson’s UK, welcomed the development. “While this study is very small, involving just three people with Parkinson’s, it offers a new way of thinking about diagnosis by measuring changes in handwriting, which can be an early symptom,” she told The Guardian.
That doesn’t mean it’s perfect. The authors acknowledge limitations. The current study had a small sample size, and all participants with Parkinson’s had motor symptoms in their dominant hand. It’s unclear how well the pen will work in people with subtler symptoms or tremors caused by other conditions. Further studies will need to address language differences, handedness, and the ability to distinguish Parkinson’s from similar disorders.
“We view it as a very promising technology,” Gary Chen told Spectrum IEE. “But as we indicate in our paper, our current study has some shortcomings.”
Still, the implications are broad. The team envisions future pens with built-in data storage and wireless transmission to smartphones or cloud databases. That could enable long-term, passive monitoring of motor symptoms — not just for diagnosis, but for tracking disease progression or treatment efficacy.
With the global Parkinson’s population projected to double by 2050, low-cost and accessible tools like this may become essential. The World Health Organization has warned of growing disparities in neurological care. In some low-income countries, there are as few as 0.03 neurologists per 100,000 people.
“A quantitative, low-cost, and accessible method for Parkinson’s disease diagnosis in large populations remains an unmet clinical need,” the authors wrote.
This diagnostic pen could help fill that void — not with ink, but with data captured in every shaky stroke.
The findings appeared in the journal Nature Chemical Engineering.