A high performing algorithm aims to translate paralyzed patients' thoughts into real-time prosthetic movements.

Blackrock Neurotech (Salt Lake City, UT, USA) has signed an exclusive license for Columbia University’s (New York, NY, USA) neural decoder MINT. The license comes on the heels of Blackrock's recent announcement to commercialize its MoveAgain platform, an implantable brain-computer interface (BCI) that restores function in patients impaired by spinal cord injuries and disabilities caused by ALS and paralysis. MINT, which recently demonstrated exemplary performance in the EvalAI competition, will be integrated into the MoveAgain software, enabling patients' thoughts to be translated more optimally into prosthetic movements.

Unlike most other neural decoding algorithms, MINT's predictive models require almost no tradeoff between performance and training time, as they are trained in under a minute. Following training, a patient with electrodes implanted in their brain imagines movement. The electrodes read the electrical firing patterns of their neurons and send the signals back through the MINT algorithm to translate the thoughts into precise movements for prosthetics, cursors, wheelchairs, vehicles and other external devices. The results are highly accurate.

For purposes of clinical translation, an advantage of MINT is that it can be trained on relatively little data and in very little time - so it can be up and running quickly. Furthermore, the algorithm is highly economical, with the ability to run on minimal computational power and simple hardware. Finally, its performance enables real-time decoding and translating of movements that mimic the speed of an able-bodied person. Blackrock has exclusively licensed the MINT algorithm, which was initially developed with Blackrock's neural recording technology. Blackrock plans to release the software with its MoveAgain platform in the fall of 2022.

"An advantage of MINT is the effortless and accurate translation of neural state to behavior," said Mark Churchland, principal investigator at Columbia's Zuckerman Institute, Associate Professor in Columbia's Department of Neuroscience and Co-Director of the Grossman Center for the Statistics of Mind. "Once you know the state, you know the behavior."

"An important function of our MoveAgain interface is the ability to generate precise movement from neural data - and MINT enables this extremely quickly," said Marcus Gerhardt, CEO of Blackrock Neurotech. "MINT is a high-performance algorithm that is both simple and accurate. When integrated into our system, it will enable incredibly powerful thought-to-prosthetic performance for our patients." 

http://www.gzjiayumed.com/en/index.asp