Identifying signs of spontaneous stroke recovery in brain signals
Identifying signs of spontaneous stroke recovery in brain signals lead image
Strokes are highly debilitating, affecting millions of people worldwide every year. While the stroke itself can be deadly, its long-term effects can lead to chronic disabilities and years spent in recovery. Being able to predict the potential for recovery can help guide rehabilitation efforts and improve quality of life for patients recovering from strokes.
Meneghetti et al. studied indicators of stroke recovery potential from extracellular brain signals in mice. Using machine learning, they identified features that could be used to predict the potential of long-term spontaneous motor recovery.
“Most research on post-stroke motor recovery focuses on therapy-induced improvements, often overlooking the brain’s intrinsic capacity for spontaneous recovery,” said author Nicolò Meneghetti. “We wanted to investigate whether early neural activity, recorded independently of any intervention, could offer prognostic insight into this natural recovery process.”
The authors recorded local field potentials, which measure electrical activity in specific brain areas, from mice shortly after inducing a stroke. They extracted multiple features from the data, and evaluated the spontaneous recovery of these mice a month later. Using a machine learning model, the team identified a combination of signal power, signal complexity, and interhemispheric communication as key features that could reliably predict the degree of motor recovery.
The researchers are planning to expand their work and explore additional stroke models and brain regions. They also plan to study whether this data can be collected using less invasive methods, such as electroencephalograms (EEGs) in human patients.
“Our results open the possibility of developing electrophysiological biomarkers for stroke prognosis,” said Meneghetti. “In the future, bedside recordings could potentially be used in the acute phase to inform clinical decisions and personalize rehabilitation strategies as early as possible.”
Source: “Post-stroke spontaneous motor recovery in mice can be predicted from acute-phase local field potential using machine learning,” by Nicolò Meneghetti, Michael Lassi, Verediana Massa, Silvestro Micera, Alberto Mazzoni, Claudia Alia, and Andrea Bandini, APL Bioengineering (2025). The article can be accessed at https://doi.org/10.1063/5.0263191 .
