Cummings and Resnik Receive MPower Seed Funding

Two faculty members in the University of Maryland Institute for Advanced Computer Studies (UMIACS) have received seed funding to advance interdisciplinary research focused on Parkinson’s disease, hearing disorders and antibiotic-resistant bacteria.

Michael Cummings (left in photo), a professor of biology, and Philip Resnik (right), a professor of linguistics, are active in three of the 17 projects recently chosen to split $3 million in seed funding from the University of Maryland Strategic Partnership: MPowering the State, known as MPower.

The MPower initiative was launched in 2012 to foster research, scholarship and innovation between the state’s top public research institutions, the University of Maryland, College Park, and the University of Maryland, Baltimore.

MPower’s competitive seed grant program—there were 52 proposals submitted this year—offers teams of faculty researchers grants of between $49,000 to $250,000. The money is intended to jumpstart high-impact research in areas that are of critical importance to the state of Maryland and the nation.

Cummings is principal investigator of a project that will use machine learning algorithms to help analyze mobility data from people suffering symptoms of Parkinson’s disease, a progressive nervous system disorder that affects nearly one million people in the United States.

He is joined on the project by Rainer von Coelln, M.D., an assistant professor of neurology at the University of Maryland School of Medicine.

The cross-institutional team will assess the severity of symptoms from 300 Parkinson’s disease patients in comparison to 50 control subjects who do not have Parkinson’s. All participants will wear sensors able to quantify and analyze their movements in their day-to-day activities or in specialized tasks they are asked to perform. Data to be analyzed includes symptoms like slowness, stiffness and shaking, mental health issues like anxiety/depression, and so-called autonomic symptoms like bladder dysfunction and dizziness.

The severity of these symptoms—and their rate of progression—often varies between patients. The captured sensor data analyzed by sophisticated algorithms could help clinicians quickly identify which Parkinson’s patients need more aggressive treatment protocols to help prevent their rapid deterioration.

Cummings is also involved in a second MPower-funded project being led by Matthew Goupell, a professor of hearing and speech sciences at UMD, and Ronna Hertzano, M.D., an otolaryngologist surgeon-scientist at the University of Maryland School of Medicine.

In this project, the researchers are developing innovative tools to effectively query and analyze hearing loss data for research purposes.

They plan to adapt a cloud-based tool developed at University of Maryland to support audiological clinical research data visualization and analysis. Ultimately, this could result in an advanced and intuitive clinical informatics tool that would assist in the implementation of multi-site clinical studies for hearing disorders.

Resnik is collaborating on an MPower seed grant with Katherine E. Goodman, an assistant professor of epidemiology and public health at the University of Maryland School of Medicine.

Their joint UMD/UMB team is interested in addressing the spread of antibiotic-resistant bacteria that can pose a grave challenge in U.S. hospitals.

One concern, the researchers say, is that colonized (i.e., “silent” carrier) patients often go undetected, transmitting antibiotic-resistant bacteria to other patients. Moreover, the colonized patients themselves are at a significantly higher risk of developing antibiotic-resistant infections, where mortality rates can exceed 50%.

But wide-scale screening for antibiotic-resistant bacterial colonization remains impractical for most U.S. hospitals. As a result, hospitals can miss critical opportunities to identify colonized patients early, when it is still possible to prevent what are often devastating patient outcomes.

The MPower-funded team is pursuing an innovative strategy: using state-of-the-art natural language processing and machine learning techniques to analyze the language in electronic health records, automatically detecting pre-admission exposures that might be found in a patient’s clinical notes.

Those notes—showing someone arriving at the hospital from a nursing home, for example—can often yield important information on strong risk factors for carrying antibiotic-resistant bacteria.

The research team aims to lay foundations for automated technology that will detect these high-risk patients in a much more targeted and cost-effective way than is currently available.

This is the second round of MPower seed funding for Resnik, who in 2016 was awarded a grant with UMB Professor of Psychiatry Deanna Kelly to develop computational models that help identify symptomatic changes in people suffering from schizophrenia or depression. The 2016 MPower grant was followed by an $842,000 National Science Foundation award in 2021.