Subtitles section Play video
This episode of Real Engineering is brought to you by Curiosity Steam. Sign up today and
get free access to my new Logistics of D-Day series on Nebula.
Over the last decade Elon Musk has become one of the most famous men on the planet.
Revolutionising the banking, automotive, rocket, and energy industries in a relatively short
period of time. His reputation for disrupting established industries has elevated his status
to some sort of tech jesus for many, and with his latest venture, Neurolink, Musk appears
to trying to take that status to the next level. Neurolink is, to me, Musk's most
fascinating venture yet. With the goal of developing technologies to unearth the mysteries
of our most vital organ, the brain.
We have decoded our DNA and even discovered methods to selectively edit it. We have invented
tiny devices that can be implanted into the body to correct our heartbeat. We can take
organs from donors and transfer them to those in need. We can perform total joint replacements
and artificially grow skin from stem cells.
But the brain remains a mystery in many ways, with little to no options for intervention
when malfunctions occur. We have only scratched the surface of this organs operation, and
to me, it's one of the final great frontiers of science.
If you pay attention to just the headlines of mainstream science publications, this technology
will seem like Musk is trying to create cyborg humans.
Where healthy people will voluntarily get biomedical implants to augment their brain
function, but that's just Musk using his tech jesus status to generate hype for his
latest business venture. In reality Neurolink is so much more than something much more meaningful,
but perhaps less exciting for the average person. This technology could help accelerate
our exploration of the brain, and help people with severe brain malfunctions and injuries
to live happier and longer lives.
To understand what Neurolink is trying to do. We must first look technologies Neuralink
is looking to improve on and get a basic understanding of how the nervous system works. For that
I will pass you over to Stephanie from our new channel Real Science:
We have many different kinds of receptors in our body to gather information about the
world around us. Take the hair cells of your inner ear. They are activated when vibrated
by sound and the cochlea, the snail shaped organ in your inner ear, is shaped in a way
to allow different portions of it to be activated by different frequencies, thanks to the differing
stiffness of the basilar membrane along the length of the cochlea. [1]
This means the base of the cochlea, closest to the oval window connected to the outer
ear is sensitive to high frequencies up to 20,000 hertz. And as we descend deeper into
the snail shaped sensory organ, lower frequencies begin to vibrate the hair cells until we reach
the apex of the cochlea where frequencies as low as 20 hertz can be detected.
When activated, these hair cells send electrical impulses through the auditory nerve to your
brain for interpretation. The exact process of interpretation is insanely complicated
and beyond the knowledge of man, as there are thousands of neurons involved that gradually
branch out as they travel to their final destination.
But thanks to our understanding of the signal input stage we can actually just bypass the
ear as a sensory organ altogether and artificially stimulate the nervous system to allow the
deaf to hear. This is exactly what cochlear implants do Seeing videos like this is quite
possibly the most heart-warming thing on the internet. Children who have never heard the
sound of their mother's voice suddenly able to hear for the first time. Their smiles would
make anyone see the value in this technology.
So how does this work? The device consists of a microphone and a sound processor, which
in turn generates electrical signals to send to an electrode array which is actually inserted
directly into the cochlea where it can directly stimulate the nerves of the inner ear with
electrical impulses. [2]
This bypasses both the hair cells of the inner ear and the sound transmitting structures
of the outer ear, and so it can help people who have malfunctions in these parts to ear.
An astounding technology, but it does not require any implantation of medical devices
into the brain, as Neurolink plans to do. It simply activates the nervous system at
its input stage. Creating a technology which could say, activate the auditory cortex directly
to allow us to hear is a whole other ball game.
Current technology on this side of things is highly invasive. Take braingate. This implantable
device consists of about 256 electrodes which can both read and stimulate neural activity.
This is exactly the function Neuralink is working to improve on. This lady is doing
something amazing. This medical implant was placed on the surface of her brain at the
motor cortex, where it records the activity of the neurons in that area. [3] The data
from those records were then used to effect a mouse cursor which has allowed her to type
and use a computer, despite having no movement in her limbs. The researchers took this a
step further and began using the neural records to allow another woman to control the movement
of a robotic arm. [4] This is the exact technology Neuralink is
seeking to improve upon, and there is a lot to be improved upon.
The first issue with Utah Array is the material properties of the electrodes. These electrodes
are like stiff and sharp needles, which allows them to penetrate into the brain and record
the internal activity, but this causes problems with the bodies immune response. [5]
This is the first part of Neurolinks plans to improve this technology by making these
electrodes much smaller.
The Utah Array's electrodes vary from about 0.03 millimeters at their tip to about 0.1
at their base [5].Neurolink threads are much much smaller at about 0.004 to 0.006 millimetres.[6]
Side by side that looks something like this.
Making the threads thinner allows them to affect a smaller portion of the brain, making
them less likely to affect nerve function or to puncture blood vessels, but perhaps
more critically makes the threads more flexible. Allowing them to move with the brain as it
jiggles around in the skull.
This is actually a huge problem. The tissue in the brain is very soft and elastic. If
you have stiff needle like electrodes fixed in place, the brain will simply deform around
them. This causes scar tissue to form around the needle which over time will block the
needles ability to read brain activity through the scar tissue
Matching the electrodes's material properties to the brains as close as possible will allow
the electrode to move and deform with the brain, and thus decrease this scar tissue
formation and extend the life of electrodes. A vital design parameter from medical implants.
So neurolink has moved away from these stiff silicon electrodes [7] and created thinner
flexible gold electrodes coated in a conductive biocompatible thin film polymer. [6]
But electrodes like this come with their own issues. Their small size and flexibility makes
them very difficult for even the skilled hand of a surgeon to insert, so Neuralink has also
developed a robotic electrode inserter to lend a helping hand.
The robot comes with a suite of camera and light modules to allow the robot to accurately
insert the threads.. The robot uses a needle to advance the electrode thread to the desired
depth in the brain before retracting and leaving the thread behind. This robot on average could
insert an electrode thread in a little over a minute even when the surgeon performed manual
adjustments to avoid blood vessels.
Neuralink's white paper put particular emphasis on this ability as the breaking of the blood
brain barrier is suspected to be a key driver in the brain inflammatory response, which
again can cause scarring and reduce the electrodes function.
It's important to note that Neuralink isn't the first company to create thin film polymer
electrodes [8], but with this robot and their work on streamlining the manufacturing process
for mass production has put Neuralink in a strong position to create a viable medical
device for sale. They have also increased the channel count significantly.
The Utah Array electrode array can reach a max channel count of 256 channels. Whereas
this prototype system, which Neuralink surgically implanted in a rat and successfully recorded
from has 96 electrode threads, each containing 32 electrodes, for a total of 3072 channels
to read from.
This is a very important design parameter as more data equals more control.
This journal paper titled “Learning to control a brain-machine interface for reaching and
grasping by primates” details an experiment where researchers implanted a brain-machine
interface into the brain of macaque monkeys.[9] They trained the monkeys to complete a task
on a screen using a small hand held controller. They recorded the monkeys motor cortex neural
activity during this training and mapped a robot arm to match his hand movements. They
confirmed that the more neurons they could record from the higher the probability of
the robotic arm matching the monkeys actual arm movements. Note this footage is from a
later 2008 study where the researchers actually trained the monkeys to feed themselves. [10]
So if we can record from more channels, we can expect to achieve higher accuracy and
later as the technology progress we can perform more complicated tasks. Perhaps instead of
controlling a cursor or robot arm, we can fit exo-skelatons to paralysed patients to
allow them to walk. However we have one last and significant technology challenge before
that can ever be considered.
We somehow need to get this data out of the brain. The electrodes record analog data from
the brain which first needs to be amplified as neural signals are very faint with voltages
as low as 10 microvolts, noise then needs to be filtered out and finally the analogy
signal is converted to binary data. This reduction to simple bits is vital, as we somehow need
to transfer this data to a computer outside of the head. Installing a processing board
inside the brain is simply not an option.
Looking at the utah array we can see there is a lot left to be desired. The electrodes
themselves require a connector which bears an uncanny resemblance to the headjack from
the matrix. When the researchers wanted to use the brain machine interface they had to
plug these massive neuroport blocks to the connector which feed the data to a huge amplifiers
and signal processing.
Neurolink is trying to fit the amplification and data filtering step inside the onboard
processors. This is their prototype board which they fitted
into the rat. Here the electrode threads fed into 12 custom built microchips each capable
of processing 256 channels of data, equalling the 3072 channels coming from the threads.
However this prototype system simply used a USB C port for both power and data transfer.
Which again is going to require an ugly port breaking the skin.
This isn't just a cosmetic issue. It's a massive open wound in the bodies first line
of defense for infection and it leads straight to the bodies most valuable organ. It's
simply not an option for a commercial product. So Neuralinks next technological challenge
is to develop a method to both power and transfer data to these implantable devices. Elon made
some off handed comments about this during his presentation about this:
“And the interface to the chip is wireless so you have no wires poking out of your head.
Very, very important. So you it's basically bluetooth to your phone. We'll have to watch
the App Store updates for that one make sure we don't have a driver issue. Uhhhm updating
…..”
Good one Elon.
Okay, so beyond joking about people's brain implants potentially having driver issues.
This comment, in typical Elon fashion, is a little misleading. Bluetooth doesn't actually
have the bandwidth needed to transfer this much data, so an alternative method will be
needed to transfer it from the device to outside the skin.
The neuralink whitepaper does not shed much light on this specific part of their plans
here, but they did present very briefly in their presentation that their first planned
product consists of four of their N1 chips. 3 will be implanted in the motor cortex for
control and 1 will be implanted in the somatosensory cortex for sensor feedback. These will feed
data to an inductive charging and data transfer coil under the skin behind the ear, which
will then transfer the data to a wearable computer and charger worn behind the ear.
This device will probably perform some further data processing before transferring the simplified
data through bluetooth to a phone where it will allow the user to control a cursor on
the phone or a computer.
Neuralink stated ambitious goal of beginning human trials this year in order to begin the
long and difficult task of receiving FDA approval. If they managed to get the food and drug association
approval for a commercial product, this would be a major leap forward for the treatment
of injuries resulting in paralysis. Potentially transforming the life of the hundreds of thousands
of people living with paralysis, allowing them to complete simple tasks like controlling
their computers without the help of a carer. While I don't see healthy people using these
kinds of devices anytime soon, as the implanting any device into the body never mind the brain
will always be a massive risk. It's certainly a plausible scenario that future humans could
elect to have a device like this implanted.
There is a legitimate worry that machine learning and artificial intelligence is going to pose
an existential threat to human society in the near future and you can learn more about
it by watching The A.I. Race on curiosity stream.
CuriosityStream has thousands of documentaries from some of the world's best documentary
makers and for just 2.99 a month, you will also get access to our new streaming service
Nebula. Nebula is a streaming service created by creators for creators.
Curiosity Stream has been a massive supporter of this channel and has even funded our new
channel Real Science as well as our new Logistics of D-Day series which is an exclusive Original
series for Nebula.
So far I have covered the reasons for choosing Normandy as a landing location and the deception
tactics the Allies used to stop the German's from discovering their plans. Next I will
be exploring how the Allies managed to punch a whole in the wall of Fortress Europe to
allow the thousands of ships, planes and men of the invasion force to stream through in
a single day.
With Nebula Original series like this, ad free videos, downloadable episodes on iOS,
with an android app on the way, and no algorithm presenting you with endless clickbait distractions,
Nebula is THE best way to watch you favourite educational YouTubers like Wendover Productions,
Polymatter, Lessons from the Screenplay, Lindsay Ellis, Kento Bento and you get it all for
free by signing up to Curiosity Stream.
As always, thanks for watching and thank you to all my Patreon supporters. If you would
like to see more from me the links to my instagram, twitter, subreddit and discord server are
below.