Making a neural interface requires accessing neural activity; the traditional approach involves removing part of the skull and placing electrodes onto or into the brain, something widely done today.1 However, drilling a hole through the skull and boring through brain tissue to access the required signals is a heavy lift for most applications. It really requires a lot of reward for the level of risk being taken. For some patients this trade-off makes sense, but it certainly is something that requires careful weighing.
Synchron, founded by Tom Oxley, Nicholas Opie and Rahul Sharma in Australia in 2012, is working on a different, really quite elegant idea: using the blood vessels — the body’s natural road system — to navigate into the brain, leaving the electrodes in the walls of the vasculature rather than penetrating into the sensitive bulk tissue of the brain itself. And they’ve taken this idea all the way into humans, with a two patient first-in-human study last year and a total of four patients currently implanted, with as of yet no serious adverse events. This is a huge accomplishment, and a testament to Synchon’s team.
There are a few obvious objections one might raise to such an approach. First, while a craniotomy has acute risk, once a patient is a few weeks out of surgery the ongoing risk is minimal as long as the skin is healed closed and no through-skin connections remain; you might worry longer about clotting with a device in the blood vessels. Second, while small blood vessels are everywhere, stents are in practice limited to relatively large vessels, constraining the areas they can record from, and they are usually unable to reliably detect single-unit activity.
As it turns out, with the use of routine blood thinners during recovery empirically the risk of stroke is low. In part this is due to the stent being placed into venous rather than arterial flow, but also because it only takes a couple weeks for the implant to be absorbed into the blood vessel wall. Animal studies show 95% endothelialization by 45 days, and preliminary recording data suggests that the process starts almost immediately. Further, many of the most interesting brain areas near term happen to be right by large blood vessels, and specifically the central venous sinus. While it’s true that Synchron can’t detect single neurons, in practice this isn’t a barrier to making products that provide real help to real patients. And, though they haven’t yet shown this, I suspect that they’ll eventually be able to get single units, through either improved electronics or accessing progressively smaller vessels, or both.
There are real advantages to an endovascular approach. First of all, there are only about 3,500 neurosurgeons in the US and only 1502 specialize in functional neurosurgery; contrast that to 10,000 interventional physicians with 2,500 specializing in neurointervention. Additionally, though the balance of equities may justify it, any placement of a device into the parenchyma of the brain is going to kill some cells, and this becomes especially significant when thinking about multiple device upgrade cycles. But to the extent that the stents can be recovered it suggests a safe multi-cycle upgrade path for patients, averting a difficult waiting game of choosing when to start one’s journey with assistive implant technology. The absence of drilling through bone also makes it a plausible candidate for conscious sedation rather then general anesthesia, further simplifying the surgical process and increasing patient access.
With all of this in mind, when I was recently connected to Tom Oxley, Synchron’s CEO, and the opportunity emerged to help in an advisory role, I jumped at the chance to work with them. It’s been amazing to watch this field go from an ad-hoc amalgam of academic labs and piles of MATLAB to the industry it’s rapidly becoming today, with exciting new ideas reaching patients and hundreds of millions of dollars flowing into the space. Synchron is absolutely among the most serious groups working on next-generation neural interfaces, and I’ve been super impressed with Tom and his team.
It’s easy to sit on an advisory board, but given my excitement for what they’re doing, I asked Tom if it would be possible to do something more substantive and invest financially. I’m excited to have the opportunity to ride along on their journey and, while realistically advisors are usually of limited usefulness, to help them succeed in whatever little way I can. In doing so, I’m joining Khosla Ventures, who led Synchron’s Series B, along with Christian Angermayer’s re.Mind Capital, DARPA, and the US Office of Naval Research, among many others, in supporting Synchron’s work. There are easily tens of millions of patients who could significantly benefit from Synchron’s near term technology, and I expect the company’s ambitions to grow with success. Engineering the brain is such a profoundly powerful concept that I expect there to be multiple trillion-dollar companies to emerge in this space as it matures.
Synchron is at an important inflection point, deep in the translation from bench to bedside, and are very actively hiring. If this work sounds interesting to you and aligns with your background, please consider working with them to help get this important technology not just to the first few trial participants now, but eventually the much larger patient populations who so desperately need it!
1. All of Neuralink’s, Blackrock’s, and Paradromics’s devices require cortical access through the skull; as do neuromodulation systems from companies like Boston Scientific, Medtronic, and Neuropace; and also all of the ECoG arrays used for epilepsy mapping and basic research.
2. If you search the American Association of Neurological Surgeons directory for practicing functional neurosurgeons in the US, there are 146 results.