Magnetic 3D-printed pen could help diagnose people with Parkinson's

The Guardian

4 days ago

It won't be much good for taking down notes, but a 3D-printed pen filled with magnetic ink could help identify people with Parkinson's disease, a small pilot study suggests.
More than 10 million people worldwide are thought to be living with Parkinson's, a neurodegenerative disorder with symptoms including tremors, rigidity, slowness of movement and mobility difficulties. While there is no cure, early diagnosis can help those affected access support and treatments earlier.
However, the team behind the new work note diagnostic methods based on observations of motor symptoms are often inefficient and lack objectivity, while those based on biomarkers – such as levels of substances in the cerebrospinal fluid – often involve specialist equipment and highly-trained healthcare professionals.
Now they say they have developed a pen that can capture tell-tale motion signs to determine if an individual has Parkinson's.
'It is very cost-effective and fully accessible for lower income countries,' said Prof Jun Chen, co-author of the study from the University of California, Los Angeles, adding that the system would be linked to a phone app to analyse the results.
Writing in the journal Nature Chemical Engineering, the researchers report how they created a pen containing a soft, silicone tip embedded with magnetic particles. The pen was then loaded with an ink that contained tiny floating particles that were magnetised by the tip.
When the pen is applied to a surface, the magnetic properties of the tip change. This, together with the dynamic movement of the ink during handwriting, produces a voltage in a metal coil within the pen, resulting in current signals, which are recorded.
'We are using the handwriting-generated electrical signal to quantify the tremor during [writing],' said Chen.
The team found signals made when participants drew wavy lines, spirals or writing – both on surfaces and in the air – accurately captured the movements.
They then used a variety of machine learning models – a type of artificial intelligence – to classify handwriting signals from 16 participants, three of whom had Parkinson's disease. The researchers found that, after training, one model was able to distinguish patients with Parkinson's from healthy participants with an average accuracy of 96.22%.
Chrystalina Antoniades, an associate professor of clinical neuroscience at the University of Oxford who was not involved in the work, said people with Parkinson's often developed smaller handwriting than normal – although this was often seen once symptoms of the condition have begun.
But while Antoniades said the pen-based approach was interesting and intriguing, she added that further testing was required, and many other approaches for early diagnosis of Parkinson's were also in development.
'What I always say is that you can't just have one biomarker. This [pen] is diagnosing the problem with handwriting, which is just one of the many symptoms that we see in our patients,' Antoniades said. 'But it can be complementing what we already found, picking up something that might be difficult to see.'
Becky Jones, the research communications manager at Parkinson's UK, also welcomed the work, noting there was still no definitive test for Parkinson's.
'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 said. 'We now need larger, more diverse studies, to better understand the potential of this method and how it might support earlier and more accurate diagnoses in the future.'