Machine Learning to Identify Locations of Brain Injury


Reference #: 01291

The University of South Carolina is offering licensing opportunities for Using Biomarkers to Aid Stroke Treatment



Many individuals experience impairments following a brain injury. Brain scans (MRI, CT, fMRI) can help provide a prognosis for recovery, guide treatment, and help determine the ideal location for brain stimulation.

Clinicians who apply transcranial stimulation cannot do so without first seeing where in the brain a patient has damage. This is a major problem, as the tools used for imaging the lesion location (MRI or CT) are very expensive, and are therefore are an obstacle to use of electrical brain stimulation in clinical rehabilitation settings. This invention will allow clinicians to estimate the location of brain damage, and to then administer electrical transcranial brain stimulation without the use of MRI or CT.


Invention Description:

This invention proposes that one can infer the size and location of brain lesions by using a battery of behavioral tests, as well as other biomarkers (e.g. age at injury). This algorithm can then be used to estimate where in the brain a patient has damage, and derive the benefits of brain scanning without requiring the expense, exclusion criteria, and other complications associated with brain imaging.


Potential Applications:

The initial area of focus is the ~1 million people with stroke—induced language impairments (aphasia), and the invention can be applied to other forms of impairment and brain injury.


Advantages and Benefits:

This invention would allow stroke survivors to get better prognosis and treatment without requiring the expenses and challenges faced by brain imaging. There is a huge variability in brain imaging tools, making the findings hard to standardize. By creating a standardized and optimized set of behavioral tests (as well as other simple measures), information can be quickly provided to clinicians. For example, a rapid baseline test could allow the clinician to choose the optimal location for brain stimulation.


Patent Information:
Title App Type Country Serial No. Patent No. File Date Issued Date Expire Date Patent Status
Machine Learning to Identify Locations of Brain Injury Utility United States 16/217,152 10,916,348 12/12/2018 2/9/2021   Issued
For Information, Contact:
Technology Commercialization
University of South Carolina
Julius Fridriksson
Chris Rorden
© 2024. All Rights Reserved. Powered by Inteum