Predicting stroke recovery motor progress using EEG data

After a stroke, the brain experiences heightened plasticity that can be used to support recovery if a patient stabilizes quickly enough. To take advantage of this, physicians need to understand the healing capacity of the patient, and whether they will be able to regain motor control.

Lassi et al. created a neural network called StrokeRecovNet that uses electroencephalography (EEG) to predict a patient’s future motor recovery. This is presented as a continuously updated score from the Fugl-Meyer Assessment scale, a performance-based impairment index for patients who have had a stroke.

StrokeRecovNet breaks the EEG data into short time windows, then extracts quantitative features from the EEG data that correlate to different brain activities.

“Because our primary goal was to predict recovery as early as possible after stroke, we first tested our model on a dataset of acute stroke patients,” said author Michael Lassi.

However, using only the data from acute stroke patients gave comparable results to previous models, so the researchers trained and tested the model on an additional dataset from subacute stroke patients — patients who are a few days or weeks out of their stroke. This improved their results, showing that these combined datasets could be used to refine healing predictions during the acute phase.

In the future, the researchers want to improve the model by including kinematic and electromyographic data from upper limb movement into the predictive framework, since these both contain important information to help predict recovery.

“We would like to further develop this approach into a flexible and modular framework that can be easily adapted to different types of data, paving the way for more comprehensive and personalized prediction models in stroke rehabilitation,” Lassi said.

Source: “Enhancing upper limb motor recovery prediction after acute stroke using EEG and subacute data,” by Michael Lassi, Stefania Dalise, Luigi Privitera, Nicola Giannini, Michelangelo Mancuso, Valentina Azzollini, Tommaso Ciapetti, Antonello Grippo, Silvestro Micera, Francesca Cecchi, Alberto Mazzoni, Carmelo Chisari, and Andrea Bandini, APL Bioengineering (2026). The article can be accessed at https://doi.org/10.1063/5.0287165 .