The magic of machine learning in medicine

Professor Hanna Suominen is bringing together experts from across different disciplines for the Our Health in Our Hands project. Photo: Eric Byler/ANU

Professor Hanna Suominen is bringing together experts from across different disciplines for the Our Health in Our Hands project. Photo: Eric Byler/ANU

Imagine that tapping your fingers on a phone screen could tell you if you are showing symptoms of Parkinson disease. Or looking through a pair of binoculars could detect tiny patches of dysfunction in your eyes, long before you would ever notice any changes in vision yourself.

These apps and devices may seem to work as if by magic, but there’s a lot of research behind the scenes to make the magic happen.

Professor Hanna Suominen is a leader at Our Health in Our Hands (OHIOH), an initiative at ANU that aims to develop personalised precision medicine for people with chronic health conditions such as Parkinson disease, multiple sclerosis (MS) and diabetes.

She says that scientific research, clever machine learning approaches and careful crafting is required to make such innovative solutions work.

“People need experts to adapt machine learning to their needs,” Suominen says.

“Also, transparent and trustworthy technology evaluations are critical. Additional safety mechanisms are needed when apps are used as decision aids and they are even more important if apps are used remotely at home. Finally, privacy is a priority.

“Technologies we have innovated put your health in your hands. For example, as easily accessible smartphone apps.”

One of the devices co-created by OHIOH research teams is a smartphone app that measures how much a user’s finger tapping pattern differs from, or resembles, those of people with and without Parkinson disease.

The magic of machine learning in medicine | Hanna Suominen | TEDxCanberra

The magic of machine learning in medicine | Hanna Suominen | TEDxCanberra

Sounds simple, right? Not exactly. The foundations of the app rest in data science and pattern recognition.

“Machine learning can classify if the person tapping is living with Parkinson disease — as 10 million people in the world do,” Suominen says.

Another example applying the magic of machine learning to diagnostics is an app that screens a user’s voice for symptoms of Parkinson disease.

“You simply say ‘aaah’ to the microphone in your phone,” Suominen explains.

The app measures a person’s voice and uses machine learning to compare specific biomarkers in the voice with those of people with and without Parkinson disease.

“Biomarkers like these can be quite complicated to compute and even impossible for a human to hear,” Suominen says.

“These ‘magic tricks’ of finger tapping and saying ‘aaah’ into your phone may sound simple, but repeating them regularly, every day, week or month, allows for medical intervention at the right time.”

This approach allows better timing of more burdensome tests of Parkinson disease, for example brain MRIs that require a patient to lie in a tube at the hospital for at least 30 minutes.

“Our app can measure the right information at the right time — when it is convenient to you or when you are concerned about a given sign or symptom. This supports early intervention before dying brain cells begin to cause complications in how people with Parkinson disease feel, think, move or behave,” Suominen says.

“Of course, the aim is not to replace medical experts but rather to support them in measuring signs and symptoms.”

Another device developed and successfully commercialised by an OHIOH research team is the objectiveFIELD Analyser (OFA), which is used to conduct visual field testing.

How wide your eye can see when focusing on a central point is your visual field. Visual field testing is one method of checking vision loss over time.

The OFA, a device which tests for vision loss, was developed by a team at ANU led by Professor Ted Maddess. Photo: Tracey Nearmy/ANU

The OFA, a device which tests for vision loss, was developed by a team at ANU led by Professor Ted Maddess. Photo: Tracey Nearmy/ANU

Many eye and brain disorders, such as diabetic eye damage and MS, cause patchy reductions in visual function. Unlike conventional eye screening devices, OFA is incredibly fast and does not rely on people pressing buttons to signal whether they can see a visual cue or not.

Professor Ted Maddess has been leading the team at ANU developing OFA for 14 years. 

“After years of research and development work, the OFA can now test 40 parts of the two eyes while a person looks into it for just 82 seconds,” Maddess says.

“The easy rapid test means people of all ages can be tested and followed over time, reducing the time burden for clinical practices.”

The actual magic of Suominen’s work at OHIOH is bringing people together across disciplines, professions and areas of expertise.

The ANU initiative involves clinicians and researchers from various fields, including chemistry, computing, engineering, medicine, physics and psychology. Importantly, the team works with people with lived experience who inform research every step of the way, from choosing research topics to designing studies and testing new devices and programs.

“Partnering with health consumers can be extremely rewarding for us all — these lived experience experts, their healthcare providers, us scientists and research translators,” Suominen says.

“Without partnering of this kind, we cannot make the discoveries, collect the data, or convert them into information, knowledge and wisdom.”

Do you or someone you know have lived experience with Parkinson disease, MS or type 1 diabetes? If you would like to help Professor Suominen and her team progress their work, please contact ohioh.management@anu.edu.au.

This article was first published in the ANU Reporter.

If you liked this story please follow us on Google News or subscribe to our FacebookInstagram or Twitter accounts.

Related news

Jasmine Pearson and The Hon Julie Bishop at ANU graduation.

18 Jul 2024

Off the fence: A journey to postgraduate excellence

Research into drug checks has won Jasmine Pearson a Postgraduate Medal for Academic Excellent.

Jin Si Lam at ANU in Canberra, ACT.

16 Jul 2024

Rising Star: Lam wins 2024 University Medal

Jin Si Lam focused on the mechanisms of thrombosis and cancer.

Updated:  29 January 2024/Responsible Officer:  Science Web/Page Contact:  Science Web