Nathan Higgins

Willett and team (Jaimie Henderson Lab at Stanford) and Sean Metzger and team (Edward Chang Lab at UCSF) recently published two cutting edge demonstrations of a brain computer interface speech neuroprosthesis. The users trained recurrent neural networks that enabled translation of brain activity into speech generation that approached normal speech ~ 150 wpm. It is Nobel prize-level science and represents a continuation of progress that provides immense hope to people suffering from paralysis. I have been asked by many people how the results should be interpreted from a clinical translation perspective. Nick Ramsay and Nathan Crone recently reviewed the results and wrote a measured piece in Nature (nature.com/articles/d41586-0…) which reads as a reasonable interpretation. They wrote: “Two BCIs represent a great advance in neuroscientific and neuroengineering research, and show great promise in boosting the quality of life of individuals who have lost their voice as a result of paralysing neurological injuries and diseases.” They highlighted some key points that may impact clinical translation: 1. “High-bandwidth recordings were taken from hundreds of electrodes, which had to be connected to external amplifiers through a ‘pedestal’ that penetrates the skin, which is cosmetically unappealing.” 2. “Highly skilled researchers were actively involved in the operation of the reported BCIs, which remain too complicated for caregivers to operate in home settings without extensive training and maintenance.” 3. “MEAs…signals tend to be unstable and require frequent updating of speech-decoding models…might be limited by degradation of the electrode materials and tissue encapsulation of the devices. ECoG electrodes…can elicit superficial tissue reactions.” So, demonstrations in experimental settings continue to soar in potential and raise expectations. While nonetheless brilliant, it is worth remembering that the unmet patient need is not necessarily met by an Olympian race the top of supra-human functioning. As Ramsay and Crone put it: “Even a basic BCI implant that allows the user to select letters or icons in assistive-technology software provides them with considerable benefits and satisfaction in daily life.” To meet the needs of patients, clinical translation requires a fully implanted, low training, low troubleshooting, generalisable, platform-agnostic solution that can be implanted in a highly accessible clinical setting. This means an invisible, simple and scalable solution that does not require caregiver intervention as the first priority for clinical translation...and critically…conducting a pivotal clinical trial. Only once we have solved these fundamental challenges obstructing clinical translation we can set our sights on the lofty potential and promise of BCI. Moore’s law is clearly in effect with BCI. Such work from Henderson and Chang’s labs will continue to raise the bar and point to what is possible.

Phillip and Rodney are two incredible people fighting ALS. We got them on a group chat. Except I was the only one using my hands to type. ted.com/talks/tom_oxley_a_br… – at Manhattan, NY