Data-Collecting Sensors Are Bringing Hope to Parkinson’s Patients
PREDICTIVE

Data-Collecting Sensors Are Bringing Hope to Parkinson’s Patients

The BlueSky lab at the IBM Thomas J. Research Center in Yorktown Heights, NY is built to look just like a residential home.

It has a living room, bedroom, bathroom, kitchen and den. But what separates it from the average living space are the data-gathering sensors placed throughout—in the cabinets, under the chair cushions, on the bed and the doorknobs. They contain magnetometers, gyroscopes and accelerometers that measure the movements of the people living there, down to a squirm or twitch.

Data-Collecting Sensors Are Bringing Hope to Parkinson’s Patients

It might sound a little creepy, but all of this data is necessary to the lab’s purpose: Collect and analyze data about patients suffering from Parkinson’s disease (PD), a progressive motor disorder that affects around 1 million people in the United States. Today, there is no cure for Parkinson’s. Instead, patients receive treatment to manage the myriad symptoms associated with the disease, from uncontrollable physical tremors to amnesia. But with a better understanding of a medicine’s impact on patients—borne of the unprecedented amount of data being collected in the BlueSky house—physicians can improve treatment plans.

What’s more, the model is designed to make the development of future treatments faster and more effective. Overall, the project’s goal is to enhance the quality of life and longevity for those who suffer from the disease.

Laying the Groundwork for Data Gathering

The BlueSky project, led by IBM, got underway in the spring of 2016. Its goal is to develop a method of studying PD that’s ready for testing in a clinical-trial format within three years. While body sensors have been used for diagnosis and treatment of PD patients before, using them in conjunction with ambient sensors is a new concept. Together, the system is designed to pick up even more information about a person’s movements, like their ability to grasp a doorknob or the way they slide a chair out from the table.

BlueSky researchers hope their approach will produce not only more data but also pure data—uninfluenced by biases on the part of the patient or doctor self-reporting the data.

Another goal of the project’s early stages is to train AI to assess PD symptoms the same way a doctor would. The standard test (known in the medical field as MDS-UPDRS3) assesses motor functions like walking, sitting and finger-tapping on a scale of normal (0) to severe (4). Researchers have built code segments, or “primitives,” that correlate with these specific movements to help AI understand this scale. For example, there’s a specific code for wrist pronation (palm facing up) and wrist supination (palm facing down).

Once trained, the AI can take measurements continuously and track improvement or deterioration along the 0-4 scale in real-time. What’s more, rather than having patients do scripted tasks in the examining room, the monitoring happens while they go about their daily lives: buttoning a shirt or twisting off a cap. This method will help researchers gather more holistic, more accurate data about patients.

Continuous Monitoring, Better Treatment

According to researchers on the BlueSky project, this more-precise monitoring of PD patients is particularly important because of the extreme variability of symptoms. The intensity of tremors, rigidity, slowed movement, and impairment of balance and posture can fluctuate daily.

“When [a patient] finally gets their 20 minutes with the neurologist to review their meds schedule and symptoms and changes, what the health professional observes is not necessarily an accurate reflection of how that person’s life is,” says Holly Chaimov, executive director of advocacy group Parkinson’s Resources of Oregon. “This objective data could fill in the rest of the picture.” Armed with more accurate insights, clinicians can better understand the progression of the disease in their patients and make treatment recommendations accordingly.

One day, a phone app might communicate with sensors to provide doctors with streams of data that help them determine how to treat patients. One IBM researcher worked toward that day with some extra motivation, with a family member suffering from the disease. His father’s doctor appointments to check symptoms happen every six months–but with the technology, adjustments of medication could happen day by day or hour by hour.

With this technology, adjustments of medication could happen day by day or even hour by hour.

Hope Beyond Treatment

The BlueSky researchers hope that once they prove their approach can work for PD patients, they can begin to apply it to other diseases, such as Alzheimer’s, where monitoring small changes can lead to more-effective, more-personalized treatments.

If symptoms can be captured in real-time, that could lead to briefer clinical trials with fewer subjects, which could make it quicker and cheaper to develop new therapies. It could also lead to earlier disease diagnosis. Many in the field expect this technology to be helpful in whatever approaches are developed to alter disease course, and to lead to better chances of halting the disease.

And while the word “cure” is rarely uttered in the Parkinson’s community, Rice believed that a shift toward data-gathering technology in medicine will lead to significant advances. With exponentially more (accurate) information about PD and its progression, the medical community will have a much better understanding of the disease. The data alone won’t find a cure, but it may point researchers in the right direction.


IBM Research remains committed to the work and legacy of Dr. Jeremy Rice. Research continues on the Blue Sky project in his memory. For more information, visit https://www.research.ibm.com/healthcare-and-life-sciences/

This content is produced by WIRED Brand Lab in collaboration with Western Digital Corporation.


Get news and stories from Data Makes Possible delivered to your inbox, including exclusive WIRED content:
[rainmaker_form id=”2931″]

 

Leave a Reply