Post-stroke spontaneous motor recovery in mice can be predicted from acute-phase local field potential using machine learning

What is it about?

After a stroke, some patients recover better than others, but it’s still difficult to predict early on who will improve the most. In our study, we looked at brain activity in mice shortly after they had a stroke affecting the motor areas of the brain. We recorded brain signals just two days after the stroke and later measured how well the mice regained their motor skills one month later. By analyzing the early brain signals with machine learning, we found that certain patterns (especially from the contra-lesional hemisphere of the brain) could reliably predict how much the mice would recover. Interestingly, the brain's activity before the stroke didn't help in making predictions, suggesting that what happens after the injury is more important for recovery. These results open the door to using early post-stroke brain monitoring to forecast recovery potential and, eventually, to guide treatment decisions in stroke patients.

Featured Image

Why is it important?

Our study shows that the brain’s electrical activity, recorded just two days after a stroke, already holds valuable clues about its natural potential for long-term recovery. Unlike many studies that focus on improvements driven by therapy, we specifically investigated spontaneous motor recovery, i.e., the brain’s ability to reorganize and heal itself without any rehabilitative intervention. Our findings suggest that early patterns of neural activity, especially in the healthy hemisphere, could help predict how well an individual will recover over time.

Perspectives