Improve Your Recovery After Stroke By Machine-Learning

New approach through machine learning for outcome prediction in stroke can significantly improve its treatment as per a study at the Ecole Polytechnique F�d�rale De Lausanne, published in the journal Brain.

Stroke is defined as a sudden restriction of blood flow to the brain that affects our function. Certain types of stroke can result in devastating or long-term disability to a patient. It has become a common emergency condition with over 1.5 million new cases each year, 3.7 million patients having persistent impairments, and only 15% achieving full recovery in Europe.

‘New approach through machine learning technique �V Connectome helps predict the recovery in stroke. This cutting-edge tool can help identify neuronal network patterns to make high-accuracy predictions on the outcome for stroke patients and improve their treatment strategies.’

The stroke recovery through rehabilitation varies depending upon the different parts of the brain affected. Hence it is important to optimize recovery based on individual courses of recovery.

The present study demonstrated two powerful, cutting-edge tools as a predictive method – connectomes and machine learning. Connectome is a map or "blueprint" of a brain's wiring, that is, the way neurons (brain cells) connect. They are generated by analyzing multiple brain images through magnetic resonance imaging and then reconstructing them based on structural or functional wiring.

Connectome – The Recovery Prediction Tool

The study team analyzed connectomes from 92 patients two weeks after the stroke and tracked their connectome changes up to three months. The patient��s motor impairment was later assessed using a standardized scale to monitor their connection changes during recovery.

The connectome information was then fed into a "support-vector machine", or SVM – a type of machine-learning model that uses examples to map an input onto an output for further recovery interpretation that was validated for its potential.

"This tool can support the prediction of individual courses of recovery early on and will have an important impact on clinical management, translational research, and treatment choice," says Professor Friedhelm Hummel, a neuroscientist and Director of the Defitech Chair for Clinical Neuroengineering at EPFL's School of Life Sciences.

Source: Medindia