同步公司可提供无需脑部手术的脑机接口技术

彭博晨光是您清晨获取资讯的最佳途径，直接在播客订阅中即可收听。

Bloomberg Daybreak is your best way to get informed first thing in the morning right in your podcast feed.

嗨。

Hi.

我是凯伦·莫斯科。

I'm Karen Moscow.

我是内森·黑格。

And I'm Nathan Hager.

每天清晨，我们都会早起制作最新一期的《彭博晨光》美国版。

Each morning, we're up early putting together the latest episode of Bloomberg Daybreak US edition.

这是您每日十五分钟的播客，涵盖全球新闻、政治和国际关系的最新动态。

It's your daily fifteen minute podcast on the latest in global news, politics, and international relations.

每天早上收听《彭博晨光》美国版播客，获取重要新闻故事及其所需背景。

Listen to the Bloomberg Daybreak US edition podcast each morning for the stories that matter with the context you need.

在苹果、Spotify或您常用的任何播客平台都能找到我们。

Find us on Apple, Spotify, or anywhere you listen.

欢迎来到《医疗先锋》系列的另一期节目。

Welcome to another episode of the Vanguard of Healthcare series.

我是马特·亨里克森，彭博智库的医疗技术分析师，彭博智库是彭博有限合伙企业的内部股票研究平台。

My name is Matt Henriksen, the medical technology analyst at Bloomberg Intelligence, which is the in house equity research platform of Bloomberg LP.

今天我们很高兴邀请到汤姆·奥克斯利，他是Synchron公司的创始人兼首席执行官，这是一家开发脑机接口技术的私营医疗公司。

We are pleased to have with us today Tom Oxley, founder and CEO of Synchron, a privately held healthcare company that is developing brain computer interface technology.

汤姆，感谢你今天加入我们。

Tom, thank you for joining us today.

谢谢邀请。

Thank you for having me.

那么，我们不妨从你的职业经历谈起，是什么促使你创立了Synchron。

And so why don't we just start with the steps in your career path that led you to founding Synchron.

好的。

Sure.

我在澳大利亚医学院学习内科医学，专攻经颅磁刺激技术并获得荣誉学位

So Australia Medical School Internal Medicine, honors in transcranial magnetic stimulation

嗯。

Mhmm.

神经内科住院医师培训，博士学位，这要追溯到2011年公司初创时期，当时获得了DARPA和美国国防部在澳大利亚的资助

Neurology residency, PhD, which was the start of the company back in 2011, DARPA funding, US defense funding back in Australia

嗯。

Mhmm.

中风领域，后来我个人通过临床中风研究项目来到纽约，进而完成了神经介入专科培训。

Stroke, then personally clinical stroke fellowship got me over to New York to then do a neurointerventional fellowship.

好的。

Okay.

神经介入领域在过去十年间因血栓切除术和Solitaire技术实现了大规模扩张。

And so neurointervention has has expanded massively over the last ten years, with thrombectomy and solitaire.

是的。

Yeah.

我们的愿景是希望神经介入技术能进入功能治疗领域，就像心脏学科从介入发展到电生理那样。嗯。

The vision was was that neurointervention would make its way into the functional domain, and we'd see, you know, in the same way cardiac went from intervention to EP Mhmm.

这将在神经领域实现。

That would happen in neuro.

随着我早期临床生涯的发展，脑机接口（BCI）作为一种理论逐渐浮现，它似乎是神经介入技术与解决BCI开颅手术挑战之间的自然交汇点。

And then BCI kind of emerged as a thesis as I was progressing through my early clinical career where it seemed to be a natural intersection between in neurointerventional techniques, solving the craniotomy challenge of BCI.

嗯。

Mhmm.

这就是起源故事。

And so that's the that's the origin story.

好的。

Okay.

加里，这很有趣，因为你刚才与心脏病学领域做的类比很有意思。

And then, Gary, it's interesting because it's a an interesting analogy you just made with the cardiology space.

为了确保我理解正确，你指的是心脏病学中用于放置支架的微创导管技术，进而转化为能够通过心脏消融传递信号。

Just to make sure I understand this correctly, you're talking about a minimally invasive catheter for cardiology for delivering a stent, translating then into being able to deliver a signal with a cardiac ablation.

这就是该技术在心脏病学类比中的发展路径吗？

Is that kind of how that cardiac analogy progresses in the technology?

医生。

Doctor.

是的。

Yeah.

但电生理导管既能感知也能刺激。

But the EP catheters can do sensing and stim.

那些用于标测的电生理导管，嗯。

The EP catheters that do mapping Mhmm.

它们正在进行某种形式的感知。

Are doing a version of sensing.

所以它们通过扫描心脏区域来创建三维心脏电生理图，然后进行消融治疗。

So they're mapping across a range of the heart to create a three-dimensional electrophysiological map of the heart, then they're doing ablation as therapy.

但在大脑中，感知组件将开启整个脑内数据经济。

But the the sensing component is what in the brain is going to unlock a whole data economy within the brain.

因此我认为，正如我与投资者交流时所说，这正是需要些时间才能让人恍然大悟的关键。

And so I think this is what's as I've spoken to investors, this is what has taken a little bit for the penny to drop.

英伟达实际上是让我明白这一点的那家公司。

NVIDIA is actually the one that made it obvious to me.

英伟达与我们合作时，带着非常清晰的愿景：一旦找到将生成脑数据的设备商业化的机制，嗯。

NVIDIA, we partnered with, came to us with a very clear vision that once you had figured out a mechanism to commercialize a device that generated brain data Mhmm.

那么你就可以基于这些脑数据开发服务于用户的产品。

Then you can build product off the basis of the brain data in service of the of the user.

但这与目前神经科技领域其他植入式神经技术有着截然不同的核心理念。

But that's a very different, thesis to anything else in neuro tech implantable neurotechnology right now.

嗯。

Mhmm.

植入式神经技术行业主要包括人工耳蜗和脑深部刺激术。

The industry of implantable neurotechnology is cochlear, DBS.

嗯。

Mhmm.

你知道吗？

You know?

然后还有Inspire。

And then there's Inspire.

还有颅神经植入设备。

There's cranial nerve implants.

主要设备方面，美敦力现在正在其深部脑刺激闭环系统中进行传感，但这种传感仅服务于深部脑刺激的参数调整。

Primary set so Medtronic is now doing sensing in their closed loop arm of the deep brain stimulation, but that sensing is just in service of the stimulation parameters for the deep brain stimulation.

脑机接口主要是用于传感。

BCI is primary sensing.

好的。

Okay.

所以问题在于，哪种技术能够可扩展地将传感器植入大脑以生成数据流，然后如何利用这些数据流大规模构建产品？

And so the the the question is which technology is going to scalably deliver sensors into the brain which generate data streams, and then how do you then use the data streams to build product at scale?

好的。

Okay.

也许这是讨论脑机接口的一个有趣角度。

And maybe that's an interesting way to talk about BCI or brain computer interface.

对于刚接触这个领域的听众，甚至包括我自己来说，用通俗的话讲，这些平台希望捕捉什么样的信号？

And just, you know, for listeners who are just new to the space and even myself, in kind of layman terms, what are the signals that, these platforms are hoping to, capture?

然后这些信号如何转化为可用于某些特定功能场景的数据集？

And then that how does that get translated into a dataset that can be able to be utilized for, some certain cases of functionality?

大脑由大量神经元组成，嗯。

So the brain is made up of a bunch of, neurons Mhmm.

大脑不同区域分布着不同的神经团块，在人群中这些脑区功能具有相当高的可预测性。

With different blobs all around the different areas of the brain, which across a population have pretty high predictability on what those areas of brain do.

嗯。

Mhmm.

第一波脑机接口的应用领域是运动控制、肢体控制、运动和肌肉控制。

The first wave of BCIs is in the domain of movement control, limb control, motor, muscle control.

嗯。

Mhmm.

如果你观察大脑中控制肌肉运动的部分，会发现存在一种可预测程度的电活动，这种活动可以在频域和时域中被捕捉，本质上就是神经元放电引发电位电压变化。

So if you look at the part of the brain that controls muscle movement, there's a very predictable degree of electrical activity that is can be captured in frequency domain and time domain, basically firing neurons, causing electrical potential voltage changes.

如果能检测到电压变化，并以某种方式标记这些数据，从而知道用户正试图执行某个动作（比如控制某块肌肉），那么你就可以从大脑中分类出这个信号，并将其转化为类似游戏摇杆的控制信号。

If you can detect the voltage changes and you can label that data in a way that you know the user is trying to perform an action, in that case, to control an a muscle, then you can classify the signal out of the brain and use it and turn it into a joystick.

嗯。

Mhmm.

这将代表运动控制功能。

And it would represent the movement control.

因此第一代脑机接口利用的是大脑中代表运动控制的电信号。

So the first wave of BCI is using electrical potentials in the brain that represent movement control.

但随着时间推移，当我们研究不同脑区时，我们将标记并识别出大脑活动的特征，逐步开发出能代表大脑各功能领域的产品。

But over time, going to different brain regions, we're going to, like, label and and figure out the features of brain activity and start to build product that represents all the different domains of brain activity.

它们将覆盖各种不同的大脑状态，解决许多迄今为止无法解决的问题。

So they're gonna cover a whole range of different brain conditions, and solve a whole bunch of problems that have not been able to be solved, up to this point.

这是否像一种正向反馈循环——获得的数据越多，对这些信号的理解就越深入，从而能在特定区域收集更多数据，逐步完善这个数据库？

And is this does this feel like a kind of a positive feedback loop as the more data you get, the better understanding you get of these signals, which allows you then to collect more data in certain areas and just be able to build out that database over time?

是的。

Yes.

所以这个领域的论点是，正在兴起一个名为神经AI的领域。嗯。

So thesis in the field in there's an emerging field called neuro AI Mhmm.

它提出支撑大型语言模型的预训练技术，比如在你不完全理解数据集所有规则的情况下，对大型数据集进行转换器预训练。

Which posits that the, pretraining techniques that have underpinned large language models, like transformer pretraining of large datasets where you don't fully understand all the rules in that dataset.

基于颅内脑数据训练数据集，能让你获得理解大脑运作的基础模型，然后你可以在代表不同认知领域的脑功能各个领域上对该模型进行后调优。

Training a dataset on the basis of intracranial brain data gives you a foundation model of understanding how the brain works, and then you can post tune that model on various domains of brain function that represent different areas of cognition.

因此你可以基于大脑中的传感器扩展来训练基础模型，然后利用它来驱动大脑已具备的一系列活动表现。

So you could train a foundation model on the basis of scaling, sensors in the brain and then use that to drive, performance across a range of activities that the brain already does.

是的。

Yeah.

你之前提到的一个方法是开颅手术，作为植入脑机接口的一种方式。

And so one of the things you mentioned earlier was just, cranotomy as kind of one of the ways to implant the BCI.

也许这是介绍你们NeuroEP的一个切入点。

And maybe this is a way to then introduce ourselves to your NeuroEP.

有哪些不同的方法可以用来捕捉这些信号呢？

What are some of the different methodologies for being able to pick up those signals?

市面上有哪些不同的设备？

What are the different devices out there?

植入这些设备有哪些不同的手术方法？

What are the different procedures to implant those devices?

神经电生理学如何与其它技术区分开来？

And how does neuro EP differentiate itself from that other technology?

所以如果你是...如果你是...

So if you're if you're

作为神经科医生，你可以专攻电生理学，但主要工作领域将是癫痫治疗。

a neurologist, you can do a fellowship in electrophysiology, but you'll primarily be working in the domain of epilepsy.

好的。

Okay.

这方面有个有趣的例子是Neuropace公司的反应性神经刺激产品。

And so maybe that's an interesting one because you've got Neuropace out there with the responsive neurostimulation product.

他们通过传感技术预测、检测、预判癫痫发作并中止发作的能力，是利用电生理参数表征病理的一个范例。

Their ability to do sensing to then predict, detect, forecast seizures, and then abort seizures is an example of the use of electrophysiological parameters that, represent pathology.

嗯。

Mhmm.

然后是帕金森病，闭环功能被用于提升刺激疗法的效果。

And then you've got Parkinson's disease, where closed loop features are being used to, improve, performance, of stimulation.

现在还有脑机接口技术，利用运动控制特征来创建操纵杆，让瘫痪患者能够控制外部设备。

You've now got BCI where the use of features of motor control are used to, create joysticks to let people who are paralyzed control external products.

嗯。

Mhmm.

当我们转向大脑其他区域时，下一个应用领域将是语言。

As we go to different brain the next one would be, speech.

因此下一代脑机接口技术正在推动大脑控制，通过语言的电生理标记来翻译言语。

So the the next wave of BCIs are driving the brain control, the electrophysiological markers of speech and then translating speech.

接着还会涉及情绪、认知和规划等领域。

And then you get into things like, emotion and cognition and planning.

因此，情绪的电生理参数将为脑机接口开辟全新机制，生成非语言交流方式，从而克服一系列相关病症带来的障碍。

And so, the electrophysiological parameters of, emotion would open up a whole new mechanism for BCIs to generate nonverbal communication that could overcome a whole range of conditions where that's a problem.

所有这些应用的共同点在于，大脑运作的电生理参数是共通且保持不变的。

What's common across all of those is that the the electrophysiological parameters of how the brain works are are common and preserved.

所以一旦你构建了解码运动控制的架构，它同样适用于语音。

So once you've built an architecture for decoding movement control, that will work for speech.

它将适用于运动控制。

It will work for motion.

它将适用于导航。

It will work for, navigation.

它也将适用于某些记忆功能。

It will work for some elements of memory.

是的。

Yeah.

这某种程度上又回到了那个正向反馈循环：收集更多数据以更好理解所有不同组件，然后能够在此基础上继续发展。

And that kinda goes back to that positive feedback loop of building collecting more data to better understand all the different components and then being able to build off of that.

我认为那些实现规模化的企业，这个行业将会出现网络效应。

I I think the companies that scale I think there's going to be a network effect in this industry.

嗯。

Mhmm.

我认为无论哪家公司能率先规模化并构建这类基础模型，都将获得难以追赶的优势。

I think whichever companies can scale first and build this type of, foundational model, is going to have an advantage that's gonna be hard to catch.

是的。

Yeah.

此外，另一个不同点是，虽然存在所有这些不同的功能，但用于收集和分析数据以最终实现功能的设备形态也存在显著差异。

And then, yep, the other difference too, there's all these different functionalities, but there's also significantly different form factors that are being used to, you know, collect this data and be able to then, you know, analyze this data then to create the the functionality.

你们的神经电生理设备形态与那些直接植入大脑的其他设备有何不同？

How does your form factor with neuro EP differ from some of the other ones that go directly to the brain?

我很喜欢你称它为神经电生理这个新类别。

I love that you're calling it neuro EP is this new category.

我喜欢这个说法。

I love it.

我们与其他所有可植入脑机接口公司（目前约有七家）之间的核心差异——嗯。

The major difference between us the the the singular difference between us and every other implantable BCI company, and there's about seven at this point Mhmm.

我们不做开颅手术。

Is that we're not performing craniotomies.

好的。

Okay.

开颅手术是一种神经外科手术，需要进入颅骨，通常需要在神经外科中心的手术室进行，因此在全国范围内这些神经外科中心的分布有限，这限制了像Neuropace、人工耳蜗和深部脑刺激这类技术的发展。

So craniotomy is, accessing the skull in a, a neurosurgical procedure, which typically requires an OR in a in a neurosurgical, center, and therefore, somewhat limited access around the country in terms of where those neurosurgical centers are, which has been one of the restrictions in growth of technologies like Neuropace and Cochlear and deep brain stimulation.

没有哪个能达到，比如人工耳蜗，我认为每年大约做4万例。

None none of those like, cochlear, I think, is doing about 40,000 units per year.

我想在美国的深部脑刺激手术大约是1万例。

I think DBS in The US is 10,000.

是的。

Yeah.

这项技术并未如人们预期那样真正起飞。

It hasn't really taken off the way that with the technology, it hasn't taken off the way people were expecting.

不过是非常有效的治疗方法。

Very effective therapies, though.

比如反应性神经刺激技术Neuropace，我认为他们现在每年能达到一千例左右。

Like, responsive neurostimulation, Neuropace, I think they're up to a thousand per a thousand cases per year.

如果对比心脏领域，这些数字与七八十年代开胸手术的数量相当。

So if you look at cardiac, they're kinda similar numbers to what was going on back in the seventies and eighties with thoracotomies for various types of procedures.

而随着介入心脏病学技术的发展，这类手术数量已从数万例增长到数十万例，如今达到每年数百万例。

And now with the advent of interventional cardiology techniques, those numbers have gone from tens of thousands to hundreds of thousands to now millions of procedures per year.

现在甚至在二级医院就能安装心脏起搏器或放置支架。

And you can get a pacemaker or a stent now down at a secondary level hospital.

嗯

Mhmm.

因此市场机会必须非常巨大才能实现大规模普及

So huge proliferation, and and so there needed to be a huge market opportunity.

然后还需要有一条可扩展的实施路径，可能要从教学医院、从手术室走出来

Then there had to be a scalable, pathway that was probably out of a, you know, teaching hospital, out of a out of an OR.

所以我认为这就是神经医学领域即将迎来的变革

And so I think that's what's coming in neuro.

因此，神经介入治疗自2015年血栓切除术取得突破以来，现已蓄势待发，我认为其规模正接近——或许已经达到——数万例，是的。

And so neuro intervention has been now poised with the breakthrough in 2015 of thrombectomy now getting into it's getting close, I think, to hundreds of that, maybe, like Yeah.

正接近十万例左右。嗯。

Getting on about a hundred thousand Mhmm.

是的。

Yeah.

病例。

Cases.

如今神经介入医师数量激增，虽不及心脏介入医师多，但增长曲线确实陡峭。有种观点认为该领域即将饱和，但心脏介入领域当年并未出现这种情况。

So there's been a huge proliferation now of neuro interventionists, not as many as interventional cardiologists, but actually getting that that the slope is high, and kind of a view that, well, there's gonna be now a a saturation of that field, but that's not what happened in cardiac.

随后医疗界转向了下一轮创新，我认为神经介入的下一个重大范式转变将是电生理学领域。

You then moved on to the next kind of innovation, and I think the next wave of the major paradigm shift with neuro intervention is going to be EP.

将会是功能性或电生理学方向的。

It's going to be functional or electrophysiologic.

为什么？

Why?

因为大脑是一个巨大的电器官，我们现在已经解决了进入微小血管甚至更远的问题。

Because the brain is this huge electrical organ, and we've now solved getting to tiny blood vessels, and beyond.

而且新技术正在涌现，将继续突破这些界限。

And there are new techniques coming that are gonna continue to break down the, horizons.

你提供的支架是否相当于开放式手术？如果我们联想到心血管领域，比较开放式心脏手术和支架植入，我们是否应该这样理解Synchron正在做的事情？

Is what you're providing a stent equivalent of what maybe was an open procedure for if we think about cardiovascular and we're thinking about open heart surgery versus a stent, Is that how we should think about what Synchron is doing?

他们提供的是那种脑机接口的支架形式，这样就不需要开颅，只需植入支架？

They're providing a stent form of that BCI that instead of having to open up the brain, you can just deliver that stent?

在我看来正是如此。

Exactly what it looks like to me.

动脉瘤的神经介入治疗也是如此。

And the same was in neuro intervention for aneurysms.

要么是开颅夹闭，要么是血管内治疗。

It was open clipping or it was endovascular.

嗯。

Mhmm.

我认为这正是BCI领域正在发生的变革。

And I think that's what's happening in BCI.

所以现在有一批基于开颅手术的开放式BCI公司，同时也存在介入式方案。

So there's a bunch of, craniotomy based open BCI companies, and then there's an interventional approach.

有个有趣的现象，我们在讨论数据收集和信号捕捉时——

One of the things that's interesting, you know, we're talking about the data collection, the signals that we're picking up.

支架植入后，如何通过动脉（而非静脉）输送的支架来解析这些神经信号？

How does a stent that gets delivered how do you go about deciphering those signals through a stent, It's delivered through an artery, not a vein, right?

这个认知很重要

Always good

。

to know that.

当支架被送入静脉后，如何解析来自大脑的信号？

When it's delivered into the vein, how do you decipher those signals from the brain?

毕竟这些信号需要从大脑经静脉传递后转换。

Because, you know, it's gotta be translated somewhat from the brain through the vein.

你们是如何构建平台和基础模型来检测这些信号的？

How how have you been able to build your platform, your foundational model to be able to, detect those?

信号主要分为三类。

So there are three, types of signal.

侵入式电极。

Penetrating Mhmm.

直接插入大脑的探针。

Needles into the brain.

没错。

Yep.

这就是Neuralink和Paradromics在做的事，深部脑刺激技术。

That's what Neuralink paradromics are doing, deep brain simulation.

还有硬膜下皮层电图。

There's subdural ECOG Mhmm.

就是癫痫检查中使用的网格电极阵列，专门用于癫痫检测的那种。

Which, if you remember, epilepsy workups for the grid arrays, they're called grid arrays used for epilepsy detection.

目前有一些BCI公司正在研发硬膜下技术，即在脑表面放置扁平阵列而不穿透脑组织。

There's a range of there's there's some BCI companies coming up that are doing subdural, so flat arrays on the surface of the brain that don't penetrate.

那就是硬膜下阵列。

That's a subdural array.

然后还有血管内技术。

And then you've got endovascular.

血管内技术中，血管会覆盖在大脑表面。

So the endovascular, the blood vessel then sits on top of the brain.

嗯。

Mhmm.

所以你把传感器植入覆盖在大脑表面的血管壁中。

So you put sensors in the wall of the blood vessel that's sitting on the top of the brain.

这比硬膜下技术多了一层隔离，因为中间隔着血管。嗯。

So it's one step removed from subdural in that there is a blood vessel there Mhmm.

那些微小血管正好覆盖在大脑顶部。

And the little blood vessels sit right on top.

较大的血管壁稍厚一些。

The bigger blood vessels have a bit of a thicker wall.

好的。

Okay.

所以这些血管实际上是与大脑接触的。

So the the the the vessels are are are touching the brain actually.

它们紧贴在大脑表面。

They're sitting on the brain.

是的。

Yeah.

好的。

Okay.

穿透式电极阵列记录的是神经脉冲信号。

So penetrating arrays record spikes.

ECOG硬膜下阵列记录的是ECOG或局部场电位，而血管内技术也有相应的版本。

ECOG subdural arrays record ECOG or local field potentials, and the endovascular has a version of that.

是的。

Yeah.

所以基本上，你可以通过那种方式获取数据。

So, basically, you're able to get that data that way.

我们某种程度上一直在回避那些显而易见的答案。

We've kind of been skating the, and kind of the obvious answers.

但你知道，当你采用这种血管内入路时，它如何消除某些副作用——比如我们讨论过的脑部穿刺针或硬膜下植入的副作用？

But, you know, when you do such an endovascular approach, how does that eliminate some of the side effects of you know, we talked about the needles into the brain or even the subdural.

请详细说明其他方案可能对患者造成的那些副作用。

Just walk us through some of those side effects that those other options may have on those patients.

嗯，开颅手术通常耐受性良好。

Well, think craniotomies are pretty well tolerated.

取决于

Depending on

开颅手术的规模将决定你的恢复期长短。

the size of the craniotomy will dictate how long your recovery would be.

我不认为在医院多住一两天恢复会真正限制这些技术的推广规模。

I don't think you know, spending an extra day or two in hospital to recover is not probably what limits the scale of those techniques.

我认为关键在于报销经济性，以及手术室资源在医疗资源匮乏环境中的可用性。

I think it's the the reimbursement economics and the availability of OR in a very kind of resource, poor environment with ORs and hospitals.

明白了。

Gotcha.

据我所知，DBS手术的等待名单已经排到六到九个月了。

Like, I believe there are wait lists for DBS out to six to nine months.

嗯。

Mhmm.

医院在DBS病例上很难盈利。

Hospitals are struggling to make money on the DBS cases.

是的。

Yeah.

我想你说的就是这些其他病例，归根结底采用血管内介入方法，周转率会高得多，一天内能完成更多患者治疗。

And I think that's what you're saying is that all these other cases, and it goes back to just doing an endovascular approach, the turnover is much higher, and you'll be able to get more patients done in a certain day.

我是说，并不指望你最终能做到，但也许某天你每天能处理几十个病例。

And I mean, not expecting you to do Eventually, some point, probably you can be able to do dozens of patients per day.

想想这些支架植入手术，你一小时就能完成一例支架手术，而开胸手术需要数小时才能完成。

When you think about these stent deliveries, you can do a stent in an hour type of capability versus an open heart surgery is hours and hours to be able to.

正是如此。

Exactly.

没错。

Yeah.

那么你就会触及到其中的经济问题，我认为最接近的类比就是心脏起搏器。

So then you get to they get to the economics of it, and I think the closest corollary I can think of with an endovascular BCI is the cardiac pacemaker.

嗯。

Mhmm.

因为它是一个电极。

Because it's an electrode.

它被放入静脉，需要拍X光片。

It gets put in a vein, takes an X-ray.

你在介入手术室完成它。

You do it in interventional suite.

它会被留在体内。

It's left in.

这是永久性的。

It's permanent.

它装有电池。

It's got a battery.

所以有支持活性植入物的报销历史，但现在你是在介入手术室完成。

So there's a history of reimbursement that supports active implantables, but now you're doing it in an interventional suite.

因此我认为，从经济学角度看，这很可能成为一个庞大的产业。

So I think, I think it it's the economics point to it being a very potential, large industry.

我们何不稍后再讨论报销问题。

Why don't we we can talk about reimbursement in a little bit.

不过，你知道，让我们先从获批路径开始讨论。

But, you know, let's just start with the path to get to approval first.

你们目前正在进行COMMAND早期可行性研究。

You have currently the COMMAND early feasibility study.

这项研究旨在突出什么？

What is that designed to highlight?

研究策略是创建一个稳定的传感平台，然后为用户建立培训方案，启动软件、进行配置，并学习如何在屏幕上导航和选择光标，从而能够控制设备以提高独立性。

The strategy in the studies is to create a stable sensing platform that we can then build a training regimen for the users to turn on the software, onboard, and then learn how to navigate and select a cursor on a screen to, be able to control devices that will improve independence.

嗯。

Mhmm.

因此该领域目前尚未有获批产品。

So there has not yet been an approved product in the field.

所以目前还没有经过验证的BCI独立使用指标。

So and there's no validated metric for independent use of a BCI.

我们正在与FDA合作，试图明确能够代表临床意义疗效的终点指标。嗯。

We're working with the FDA to try to clarify what an endpoint would be that represents, clinically meaningful efficacy Mhmm.

而且这将是可以获得报销的。

And which is going to be reimbursable.

我们真正关注的是改善用户生活中的独立性，这主要针对严重运动障碍患者，如ALS患者、神经退行性疾病患者，他们已丧失操作数字设备的能力。

And the thing that we're really zeroing in on is the improvement of independent in the lives of the users, and this is predominantly severely motor impaired patients such as with ALS, neurodegenerative disease, who've lost the ability to, navigate with their digital, devices.

好的。

Okay.

因此它替代了用手或语音打字或在屏幕上移动光标的功能，我们实际上正以苹果产品为首个获批目标。

So it replaces the use of your hands or your voice to type or to push a cursor around the screen, and we're actually targeting, working with Apple products for the for the first approval.

好的。

Okay.

这确实是件有趣的事，因为你们确实发布了与苹果合作的新闻稿。

And so that's actually one of the interesting things because you guys did have a press release, about your partnership with Apple.

那次合作的具体细节是什么？你们希望患者能够实现哪些功能？

What did the details entail with that, partnership, and what are you hoping to have the patients be able to do?

我们正在使用苹果的无障碍平台。

We are using the Apple accessibility platform Mhmm.

作为一种后端系统来发送摇杆控制信号，从而在苹果设备上移动屏幕光标，因为他们有机制可以接收各种输入信号。

To as a kind of, back end to send out, joystick control signals to push the cursor around the screen on the Apple device because they have a mechanism to enable all sorts of inputs to come in.

因此与苹果合作的这段历程中，我们一直致力于与他们共同开发首个用于BCI产品控制的蓝牙协议。

So part of that journey with Apple has been to, work alongside them developing the first Bluetooth protocol for BCI control of the product.

该协议现已包含在苹果开发者套件中。

So that's now released in the Apple, developer kit.

嗯。

Mhmm.

这很有趣，因为蓝牙配置文件被称为HID（人机接口设备）配置文件。

It's interesting because, Bluetooth profiles, they're called HID, human interface device, HID profile.

好的。

Okay.

其中有一种是鼠标配置文件。

There's a mouse one.

还有键盘配置文件。

There's a keyboard.

以及眼动追踪HID配置文件，现在又新增了脑机接口配置文件。

There's an eye tracking HID profile, and there's now a brain BCI profile.

它成为了开发者的沙盒环境，让BCI公司能够通过预先商定的协议发送信号，嗯。

And it's, it's it's becomes a sandbox for developers, for BCI companies to send signals with a, you know, kind of pre agreed protocol Mhmm.

设备能识别信号来自大脑，而现在使用鼠标和手指会受到限制，因为你无法改变手部输入方式。

Where the device recognizes that it's coming from the brain, and, there's now you know, we're we're you're kinda constrained with the mouse and your fingers because you can't change your inputs from your hand.

但如今BCI公司将能持续创造新的输入方式，并通过蓝牙配置文件作为一种标准化语言来运作。

But now the BCI companies will be able to continually create new inputs and use this profile over Bluetooth to as a kind of a standardized, language to work within.

这非常有趣，而且有望成为新的行业标准。

So it's it's super interesting, and, it's hopefully gonna become a new industry standard.

这是否属于我们之前讨论的学习过程的一部分？就是与FDA沟通的早期可行性研究？

So is this part of the learning process that we're talking about with the, you know, early feasibility study, your conversation with the FDA?

这些反复讨论是否在尝试确定蓝牙功能的最佳方案？

Are these just kind of the back and forth as you try to figure out the best program for the Bluetooth capabilities?

FDA对于这项开发的初步意见和反馈是什么？

What have been the FDA's initial comments and feedback about, this development?

是的。

Yeah.

我认为，要获得商业批准，必须进行一项关键研究来证明安全性和有效性。

I I think, in order to get commercial approval, you have to conduct a pivotal study that demonstrates safety and efficacy.

是的。

Yep.

安全性相对直接明了。

So safety is pretty straightforward.

就像，你知道如何评估安全性。

Like, you you know how to, assess safety.

有效性则不那么明显，因为进入关键阶段前，必须确定一个指标，理解该指标的分布范围，了解哪些患者群体会对该指标产生反应，并确保其经过验证和可重复。

Efficacy is is not as obvious because, to get into a pivotal, you have to have identified a metric, understand the spread of the metric, understand which patient population will respond to that metric, and have it validated and repeated.

所以你提到的反复讨论，是的，这就是我们开发产品的本质过程。

So the back and forth you mentioned, yes, that's been the nature of the is that we build a product.

我们去测试它。

We go and test it.

我们达到一定的功能水平。

We get to a certain level of function.

我们意识到还需要再多做一些工作，这中间会有反复调整。

We realize we have to do a bit more, and there's back and forth.

而且，其中一部分工作涉及提升制造能力，以使平台更加稳定。

And, part of that has been ramping up of the manufacturing to make the, the platform stable.

嗯。

Mhmm.

每次做出微小改动时，尽管我们有数百个组件。

Every time you make a little change, though we've got hundreds of components.

每次做出微小改动，FDA都要求进行大量测试以确保...是的。

Every time you make a little change, the FDA wants a bunch of testing to Yep.

你知道，要确保它是安全的。

You know, make sure it's safe.

所以，你看，如果你看看所有成功的三类公司——比如Inspire、Axonics这些。

So, you know, if you look at if you look at all the class three you know, look at Inspire, look at, Axonics, look at those successful class three companies.

它们都经历过类似Setpoint最近的情况。

They've all gone through a Setpoint more recently.

我认为Setpoint进行了三项可行性研究，是的。

I think Setpoint had three feasibility studies Yeah.

在进入关键试验之前。

In the lead into the Pivotal.

所以我们正在尽快推进。

So we're trying to get there as fast as we can.

我们刚完成第二项研究，现在正在规划第三项。

We've just finished our second study, and we're now planning our third.

我们刚刚在DVT阶段实现了重大设计冻结，相信这将达到商用级系统标准。

We've just hit a pretty major design freeze with the DVT with what we believe will be a commercial grade system.

好的。

Okay.

因此我们对此感到振奋，目前正在明确疗效标准，为进入关键研究做准备。

So we're excited about that, and we're now working towards the clarification of what that efficacy is to then enter into a pivotal study.

其中部分工作包括标准化流程——让患者通过脑数据与设备的交互训练，达到特定的用户熟悉度。

And part of that is standardizing the pipelines with which the patients do the training of the back and forth between the brain data and the device to get to a certain level of familiarity with the user.

我知道当我尝试这样做时，光标会这样移动，它有效且稳定，我现在可以独立使用它了。

I know when I try and do that, the cursor's gonna do this, and it works, and it's stable, and I can just go away and use it on my own now.

是的。

Yeah.

好的。

Okay.

有趣的是你们正逐步接近设计冻结阶段。

And that's interesting that you're kind of getting the steps closer to the the design freeze.

因为，你知道，以我作为股票分析师的背景来看，我写的这些报告可以不断改进、编辑，但我也需要将内容定稿发布出去。

Because, you know, in my mind and then in my background as equity analyst, these reports I write, I could keep improving them, editing them, but I also need to get something published and out the door.

当你谈到这些迭代过程并提到设计冻结时，你们是如何确定‘好，这样已经足够好了’这个节点的？

As you were talking about those iterations and you're talking about getting the design freeze, when did you get to the point where like, okay, this is good enough.

我们可以继续推进这个方案。

We can go ahead with this.

还有，这个决定最终是如何做出的？

And, you know, how is that decision ultimately made?

我是说，在达到那个冻结状态之前，你们需要满足哪些具体条件或标准？

I mean, what was something you what were the, like, the the boxes that you wanted to check off before you got to that, that freeze?

我认为我们花了一些时间来确定那些真正能改变患者生活的使用场景

I I think it took a bit of time for us to identify the use cases in the patient population that were really going to be life changing

嗯。

Mhmm.

而且这些不仅要受到用户喜爱，还要能够获得CMS的报销覆盖

And and which were not just liked by the user, but were going to be covered by CMS.

是的。

Yeah.

所以报销这部分很重要，甚至FDA也公开承认，他们不想批准一个关键性试验和商业产品，如果它最终无法获得报销的话

So the reimbursement part of it and even, you know, the FDA has also freely admitted they're, like, not exactly FDA does not wanna approve a pivotal trial and commercial product that's not, gonna be reimbursable.

是的。

Yeah.

没错。

Yeah.

我想说的是，我们花了一些时间才意识到，可以构建一个我称之为最小可行产品的方案，它能成为定义类别的首个市场批准产品，从而为这个新兴行业设立标准。

I would say it's taken time for us to realize, you know, what I might call a a minimal viable product that we can build that can be a category defining first market approval that's then going to set the bar for what will be a new industry.

我不会在这次播客中透露具体内容，因为目前我们还需要对此保密。

I'm not gonna, admit what that was on on, this podcast because it will keep that close to our chest for the moment.

但这并不显而易见，因为人们总想下棋——是的。

But it's not obvious because people wanna play chess Yep.

玩游戏、上网冲浪。

Play games, surf the Internet.

这些虽然很重要，也是脑机接口的良好应用场景，但它们对生活质量的临床意义提升并不显著。

They're very important things and a good use case for a BCI, but they're not clinically significant improvements in your life.

是的。

Yeah.

所以确实。

So it yeah.

因此，那个（核心应用）究竟是什么并不那么显而易见。

So it it wasn't that obvious with what that might be.

嗯，我是说，任何涉及新领域的事情，都需要时间去了解目标在哪里，甚至在你到达之前。

Well, I mean, just anything with the new frontier and everything, it's gonna take a while to learn what the the goalposts are before you even get there.

是的。

Yeah.

是的。

Yeah.

嗯，我能预见一个庞大的产业，怎么说呢，一条临床护理路径。

Well, what I can say is foresee a big industry, how would I put it, a a clinical care pathway Mhmm.

职业治疗师在其中扮演着至关重要的角色。

Where the occupational therapist plays a really critical role.

好的。

Okay.

国内有二十五万名职业治疗师。

There's a quarter of a million occupational therapists in the country.

嗯。

Mhmm.

他们经常上门服务。

They often go out to the home.

他们协助设备连接工作，在我看来，他们显然是神经科医生和康复医师的天然合作伙伴，后者将负责开具脑机接口的处方。

They help with, device connectedness, and they seem to me to be the obvious partner for the neurologist and rehabilitation physician who will prescribe the BCI.

然后治疗师将作为职业治疗师上门，确保设备正常运行、假肢在家中正常工作并保持连接，同时进行故障排除，并且我认为他们将帮助患者处理持续更新的不同使用场景和训练模式。

And then the therapist will be a OT who will go out and ensure the device is working, the prosthetic is working in the home and is connected and will troubleshoot and will, you know, I think be helping the patient service the continued updates that are going keep coming on the different use cases and the different training paradigms.

这很有趣，因为我关注了一些机械取栓公司。

Which is interesting because I follow some of the mechanical thrombectomy companies out there.

他们在开发治疗急性中风的取栓设备方面做得非常出色，无论是抽吸装置还是支架取栓器。

They've done a great job of creating those thrombectomy devices, either aspiration or the stent retriever, to treat the acute stroke.

但这些公司中有许多曾试图投入资金解决康复业务方面的问题，却未能取得成功。

But a lot of these companies had tried to put in money to try to figure out the rehabilitation side of the business, and they have not been successful with that.

我不确定中风康复是否是你关注的重点，特别是考虑到你曾是神经科医生，但当你谈到康复、职业治疗师和各种疗法时，我觉得有必要提出这个关联点。

I'm not sure if stroke rehabilitation is something that is of note for you, especially because you're a previous neurologist, but I just felt intrigued to bring that out as a connection with when you talk about rehabilitation and the OTs and the therapies and everything.

我很好奇这是否也是你曾考虑过的问题。

I'm curious if that's something that's come across your mind.

传统上，康复治疗一直是一个难以获得报销的极具挑战性的领域。

So rehabilitation has been a very challenging domain to secure reimbursement in traditionally.

是的。

Yep.

康复治疗旨在通过多种方式促进身体的自然恢复。

Rehabilitation therapy is to promote natural recovery of the body in various ways.

因此，这些设备试图证明能加速康复或带来其他方式无法实现的改善。

And so the devices have intended to try to demonstrate speeding up or a, improvement in recovery that wasn't otherwise found.

脑机接口不属于这一类别。

BCI is not in that category.

好的。

Okay.

它属于假体类别。

It's in the prosthetic category.

好的。

Okay.

所以和人工耳蜗一样。

So same as cochlear.

所以人工耳蜗是一种假体。

So cochlear is a prosthetic.

而假体是一种能恢复丧失身体功能的设备。

So and a prosthesis is a device that, restores a lost bodily function.

嗯。

Mhmm.

因此我认为这就是BCI未来可能走的报销途径。

And so I think that's the reimbursement pathway that's gonna be out for in for BCI.

但作为审批的一部分，你必须证明你恢复了哪个身体部位的功能。

But then, you know, as a part of your approval, you're going to have to demonstrate what body part you've restored.

嗯。

Mhmm.

第一波BCI，我们称之为运动功能嗯。

The first wave of BCIs, I we're calling motor Mhmm.