INNOVATION ARTICLES THE IDEA SUBMISSION PORTAL FROM MEDTRONIC

July 2019 article image

A GLOVE THAT COULD PREDICT THE ONSET OF PARKINSON’S DISEASE 


Lewis Packwood
July 2019

Prof. Filippo Cavallo and his colleagues at the Biorobotics Institute of Scuola Superiore Sant’Anna in Pontedera, Italy, have developed a glove that could help to predict the onset of Parkinson’s disease up to seven years in advance1. Parkinson’s disease affects an estimated seven million to ten million people worldwide2, and it manifests as the loss of dopamine neurons in the brain, affecting the motor system. In many cases, by the time an individual is diagnosed, they will have already lost a substantial proportion of their dopamine neurons – but early diagnosis and treatment could help to slow the progression of the disorder.

 

Motor-skill deterioration

Prof. Cavallo’s original intention when developing the glove, now known as the SensHand, was to monitor the decay in motor skills that is associated with Parkinson’s progression3. He was intrigued by “the possibility to have a device that allows the neurologist to assess accurately and objectively the motor performance, so they can better follow the progression and the evolution of the disease.” One aim was to track the effect on motor skills of different treatments, such as different dosages of the Parkinson’s drug levodopa.

Prof. Cavallo wanted a device that could be used to measure part three of the UPDRS (Unified Parkinson's Disease Rating Scale) clinical protocol, “which includes a number of physical exercises that neurologists use to evaluate and assess performance.” In particular, he wanted to track the fine motor skills of the fingers. “Most of the wearable sensors available in the literature are sensors that are related to upper or lower limbs,” he says, “but nothing was available for accurately measuring the kinematics of fingers.”

This led him to develop SensHand, which features sensors and components commonly used in smartphones in combination with customized electronics. The ‘glove’ is formed from three sensor-containing rings that sit on the ends of the subject’s fingers and thumb. The rings are in turn connected by wires to an electronic bracelet. The idea is for the subject to perform everyday activities while wearing the glove, then Prof. Cavallo’s team can extract a series of kinematic parameters from the gathered data using algorithms that they have developed. “Using these sensors,” says Prof. Cavallo, “we were able to collect … more than 600 parameters for each experimental session.”

 

Detecting early signs of disease

An exciting development was the realization that SensHand might be able to predict the onset of Parkinson’s disease. “We actually discovered another important use of the device,” says Prof. Cavallo, “which is related to using measurements of motor skills and artificial intelligence algorithms to classify different levels of motor skills. In this sense, we tried to use this device not on Parkinson’s disease patients, but on subjects at risk of Parkinson's. We realized that in this group of people, the device was able to recognize a different level of motor performance and was able to create a sort of classification between the Parkinson patients, subjects at a different level of risk of Parkinson's and completely healthy subjects.”

This led to a study in which Prof. Cavallo and his team recruited 30 individuals with Parkinson’s disease, 30 healthy subjects and 30 people at risk of developing Parkinson’s disease4. The latter group was composed of people who had shown signs of losing their sense of smell – hyposmia – which commonly occurs in people with Parkinson’s. 

The results showed that the SensHand measurements could be used to distinguish between healthy subjects and Parkinson’s patients with around 95% accuracy, and it could pick out patients with hyposmia from Parkinson’s patients and healthy subjects with around 80% accuracy.  Prof. Cavallo says that although using the SensHand alone probably wouldn’t be enough to confidently provide an early diagnosis of Parkinson’s disease, this simple technology could be used to select patients for further testing, such as brain scans, “which you cannot do to everybody, because they are very expensive”.

 

Prof. Filippo Cavallo and his colleagues at the Biorobotics Institute of Scuola Superiore Sant’Anna in Pontedera, Italy, have developed a glove that could help to predict the onset of Parkinson’s disease up to seven years in advance1. Parkinson’s disease affects an estimated seven million to ten million people worldwide2, and it manifests as the loss of dopamine neurons in the brain, affecting the motor system. In many cases, by the time an individual is diagnosed, they will have already lost a substantial proportion of their dopamine neurons – but early diagnosis and treatment could help to slow the progression of the disorder.

 

Motor-skill deterioration

Prof. Cavallo’s original intention when developing the glove, now known as the SensHand, was to monitor the decay in motor skills that is associated with Parkinson’s progression3. He was intrigued by “the possibility to have a device that allows the neurologist to assess accurately and objectively the motor performance, so they can better follow the progression and the evolution of the disease.” One aim was to track the effect on motor skills of different treatments, such as different dosages of the Parkinson’s drug levodopa.

Prof. Cavallo wanted a device that could be used to measure part three of the UPDRS (Unified Parkinson's Disease Rating Scale) clinical protocol, “which includes a number of physical exercises that neurologists use to evaluate and assess performance.” In particular, he wanted to track the fine motor skills of the fingers. “Most of the wearable sensors available in the literature are sensors that are related to upper or lower limbs,” he says, “but nothing was available for accurately measuring the kinematics of fingers.”

This led him to develop SensHand, which features sensors and components commonly used in smartphones in combination with customized electronics. The ‘glove’ is formed from three sensor-containing rings that sit on the ends of the subject’s fingers and thumb. The rings are in turn connected by wires to an electronic bracelet. The idea is for the subject to perform everyday activities while wearing the glove, then Prof. Cavallo’s team can extract a series of kinematic parameters from the gathered data using algorithms that they have developed. “Using these sensors,” says Prof. Cavallo, “we were able to collect … more than 600 parameters for each experimental session.”

 

Detecting early signs of disease

An exciting development was the realization that SensHand might be able to predict the onset of Parkinson’s disease. “We actually discovered another important use of the device,” says Prof. Cavallo, “which is related to using measurements of motor skills and artificial intelligence algorithms to classify different levels of motor skills. In this sense, we tried to use this device not on Parkinson’s disease patients, but on subjects at risk of Parkinson's. We realized that in this group of people, the device was able to recognize a different level of motor performance and was able to create a sort of classification between the Parkinson patients, subjects at a different level of risk of Parkinson's and completely healthy subjects.”

This led to a study in which Prof. Cavallo and his team recruited 30 individuals with Parkinson’s disease, 30 healthy subjects and 30 people at risk of developing Parkinson’s disease4. The latter group was composed of people who had shown signs of losing their sense of smell – hyposmia – which commonly occurs in people with Parkinson’s. 

The results showed that the SensHand measurements could be used to distinguish between healthy subjects and Parkinson’s patients with around 95% accuracy, and it could pick out patients with hyposmia from Parkinson’s patients and healthy subjects with around 80% accuracy.  Prof. Cavallo says that although using the SensHand alone probably wouldn’t be enough to confidently provide an early diagnosis of Parkinson’s disease, this simple technology could be used to select patients for further testing, such as brain scans, “which you cannot do to everybody, because they are very expensive”.

 

WE TRIED TO USE THIS DEVICE NOT ON PARKINSON’S DISEASE PATIENTS, BUT ON SUBJECTS AT RISK OF PARKINSON'S.

Problems and progress

Prof. Cavallo says that the hardest thing to get right in the SensHand was the wearability – finding a material for the finger rings that would keep them from slipping in order to provide accurate measurements, but that would also be comfortable enough to wear for long periods. After trying out many different plastics using a 3D printer to make prototypes, the team eventually settled on using silicone for the rings to provide the required level of softness and rigidity.

Overall, the project has involved a fair amount of interdisciplinary collaboration, says Prof. Cavallo. “There has been strong cooperation between biomedical engineers and neurologists; in particular, our university collaborated with the Apuane Hospital in Massa, Tuscany.” And now the next step is launching a pilot study of the SensHand in the clinic. “In terms of workload, I think that we need one year,” says Prof. Cavallo.

In the meantime, Prof. Cavallo is confident he can make the SensHand even more accurate: “Of course, we can improve the algorithm, so we will do more and more research.”

References

1

Early diagnosis of Parkinson’s disease: wearable sensors and artificial intelligence may predict the onset of disease faster than current diagnostic tools. Sant’Anna Magazine, https://www.santannapisa.it/en/news/early-diagnosis-parkinsons-disease-wearable-sensors-and-artificial-intelligence-may-predict

2

Parkinson’s Disease Statistics. Parkinson’s News Today, https://parkinsonsnewstoday.com/parkinsons-disease-statistics/

3

Interview with Filippo Cavallo, June 2019.

4

Cavallo F, et al. Upper limb motor pre-clinical assessment in Parkinson's disease using machine learning. Parkinsonism Relat Disord. 2019; https://doi.org/10.1016/j.parkreldis.2019.02.028.