[Image from MC10]
Bernadette Mroz suffers from Parkinson’s. When her medication isn’t working, it makes it hard for her to function.
“My world goes into a spin cycle. I cannot function mentally, emotionally or physically,” said Mroz in a University of Rochester news release.
While a cure may not be available anytime soon, researchers are hoping to use wearable sensors to help control Mroz’s and other Parkinson’s sufferers’ tremors and memory lapses by fine tuning their medications using data collected from sensors.
Mroz was part of a University of Rochester clinical study that had her wear a total of 5 sensors – one on each limb and one on her heart. Each sensor measured acceleration in 3 directions 30 times a second for 46 hours. In total, there were 25 million data points over an almost 2 day period.
The data from the sensors helped physicians make a choice on what treatment would be best for her and allowed them to make decisions regarding her medications.
“Instead of treating all patients as averages, which none of us are, we will be able to customize treatment based on individual data,” said Gaurav Sharma, professor of electrical and computer engineering at the University of Rochester and one of the collaborators on the study.
Sharma and neurologist Ray Dorsey used BioStampRC sensors developed by the biomedical healthcare analytics company MC10, whose CEO Scott Pomerantz is a Rochester graduate and supporter of the study.
Kathik Dinesh, a graduate student in Sharma’s lab, created machine learning algorithms to make the massive amounts of data more understandable to physicians. The algorithms correlate signals gathered from the sensors and convert them to signal features that help measure coordination and tremor intensity. The machine learning techniques also use clustering and classification to categorize the different attributes of the different stages of Parkinson’s between those with the disease and those without the disease. Machine learning can also indicate whether a patient has taken their medication and whether the medication is working as it should.
“We’ve just scratched the surface in terms of the depth of data we have to work with,” Sharma said.
Machine learning is supposed to make it easier for physicians and caregivers to read and determine what is going on.
“If you tell a physician you have to look at two GB of data to figure out what’s going on with your patient, you don’t have a chance,” Sharma said. “But if you can present the data in easily digestible plots and visualizations, the physician can comprehend it and act on it.”
“This is a two-way conversation,” Sharma said. “It’s not like I can sit in my office and come up with the best way to do this. I have to ask the physician, ‘What are the attributes that would be most relevant to you and what would be the presentation of data that would make the most sense?'”
Parkinson’s disease is a chronic and progressive movement disorder. Approximately 1 million people have Parkinson’s in the U.S. alone. It is commonly characterized by tremors of the hands, arms legs, jaw and face. Those with Parkinson’s also exhibit signs of slowness of movement, rigidity or stiffness of the limbs and trunk or impaired balance and coordination, according to the Parkinson’s Disease Foundation.
The Rochester researchers hope that one day patients could go to a pharmacy and receive the patches and place them on their bodies two days before a doctors appointment to be able to have more accurate measurements. Currently, doctors and other movement disorder specialists observe the tremors of a Parkinson’s disease sufferer and rates their symptoms on a scale of 0 to 4.
Participants in the Rochester study have to mail their patches back to the researchers – for now. MC10 is developing a new line of sensors that will be as obtrusive as a temporary tattoo and will wireless transmit data to a patient’s smartphone and a secure database where it can be further analyzed. It will also allow patients to monitor their symptoms from home.
[Want to stay more on top of MDO content? Subscribe to our weekly e-newsletter.]
