3
$\begingroup$

Could you map out all the connections in someones brain just by probing every neuron and tracking them as they send signals to each other (as opposed to every synapse, a much less ambiguous but far harder exercise)

The technology level is 20 minutes into the future. We already have self-assembling diamond-coated copper wires and that human cells actually eat some kinds of nanowires. (we're assuming said mad scientist already has a big enough computer to record all the data, just record and map, no-one's getting their mind uploaded to the cloud anytime soon)

If someone could get enough of these into someone else's head so that there's at least one probe in the nucleus of every neuron, would they be able to work out which neurons were connected where by tracking nerve impulses as they move around the brain?

Improve this question
$\endgroup$
5
  • 2
    $\begingroup$ There is ~86 billion neurons in there, good luck getting the first step to work. The second step of actually recording which neuron is connect to which one even if you had wires connected to all of them isn't something that humans have thought about because we would never thought to be able to do the first step. $\endgroup$
    – A. C. A. C.
    Commented Nov 24, 2017 at 23:10
  • 1
    $\begingroup$ Another problem is that the connections aren't simple: any particular neuron might be connected to thousands of others, so that 86 billion becomes 86 trillion connections, each of which can be weighted according to what it has done in the recent past. $\endgroup$
    – jamesqf
    Commented Nov 25, 2017 at 3:03
  • $\begingroup$ So imagine such a detector exist, it must consume a certain amount of energy however tiny to pinpoint just 1 neuron. I'm saying it would take a long long time to probe every single neuron in a healthy brain without cooking it ever so slightly. Also the signal must not be weaker than the noise generated by the brain activity.😝 $\endgroup$
    – user6760
    Commented Nov 25, 2017 at 4:56
  • $\begingroup$ There's a very nice answer here about the difficulties of mapping something that small. quora.com/… $\endgroup$
    – userLTK
    Commented Nov 25, 2017 at 15:31
  • $\begingroup$ @user6760 I wouldn't worry about cooking someone's brain too much, one source placed a single neuronal spike at ~1 nanoamp (I had to look that up) while modern electronics can measure currents about a million times smaller in the femtoamp range so the neuron wouldn't even notice. $\endgroup$
    – Samwise
    Commented Nov 26, 2017 at 6:23

3 Answers 3

Reset to default
4
$\begingroup$

It might work, if the blasted brain wasn't changing all the time!

Our neurons are constantly adjusting their synapses. During the night, synapses are destroyed by glial cells. During the day, neurons rebuild them. This happens at a quite surprising rate. Studies have shown substantial synapse remodeling in mice on the order of 4-5 days!

As a result, you would never get a snapshot of the brain that "held still" long enough to map completely. Any given neuron only fires about once every 6 second, so you won't get very much data before you have to seriously consider the possibility that new synapses have grown during that time.

The brain is not a stationary creation. It should be expected that its ability to function is meaningfully affected by these new synapses as it responds to input. Thus, a static mapping would not be as helpful as one would like.

It would, however, be quite pretty.

Hippocampus from "Mind Illuminated"

Improve this answer
$\endgroup$
3
  • $\begingroup$ I was thinking more along the lines of a continuous process where you would be correlating the measurements all the time, probably not a big enough computer today, but that's ok that can be done later, data is cheap, even if processing it isn't $\endgroup$
    – Samwise
    Commented Nov 26, 2017 at 6:29
  • $\begingroup$ If you'd like, I can write another answer to cover the continuous process argument, but if I summarize: everything we know about the brain suggests that it is a chaotic system. This means that it is very sensitive to initial conditions. Trying to correlate those numbers over a long time would be brutally difficult. Basically: "Here's 2 billion devices. If any one of them changes state, the results are going to completely decorrelate your scan. And by the way, many of them turn over weekly." $\endgroup$
    – Cort Ammon
    Commented Nov 26, 2017 at 16:05
  • $\begingroup$ Now there is a process which you can use to stabilize chaotic systems to measure them via repeated impulses, or via continuous response (think: electrode causing neurons to fire, rather than just listening to them). However, I don't believe a human would survive this process. It strikes me as unlikely that one could stabilize the brain without losing all that makes it important to us. $\endgroup$
    – Cort Ammon
    Commented Nov 26, 2017 at 16:06
4
$\begingroup$

Not yet.

In 2008 I wrote an overview/roadmap of what we knew back then about whole brain emulation, the scanning and computational reconstruction of brains. The main scheme I assumed was a slice-and-dice scanning method where the tissue is mapped on the sub-cellular level. In Apprendix E I gave the reasons for why I did not explore non-destructive scanning methods in detail.

However, I had a long debate with Robert Freitas Jr. about using nanofibers, manufactured using mature nanotechnology, to connect every neuron more or less to external read-out devices. The details of the sensing can be read in section 4.8.6 and subsequent pages, and the fiber network is described in section 7.3.1 of his book Nanomedicine Vol I. I think there is nothing physically impossible about this scheme. However, I think the medical engineering challenges are worse than he thought: beside needing fairly advanced nanomachinery it also needs to interact with a dynamic, soft environment that is pretty sensitive and responds. Robert argued back; basically, see Nanomedicine Vol IIA for his assessment, especially section 15.3.6.5.

My view is that we will use destructive methods long before we can do the actual read-out using electrodes. That requires a lot of technology development, both of nanotechnology but also handling a messy biological environment. But I do not think it is impossible, just hard and slow.

As a further twist, the synthetic biology people are getting involved in the game by thinking about "molecular tickertape", where signals are recorded on DNA inside cells, and "DNA bar codes" where the cell connectivity is recorded. Of course, the current vision they have for reading this involves putting the brain in a blender and then sequencing the DNA pieces...

Improve this answer
$\endgroup$
2
  • $\begingroup$ Interesting background you have there! Maybe you could look over this answer . $\endgroup$
    – JDługosz
    Commented Nov 26, 2017 at 1:49
  • $\begingroup$ That paper makes for quite interesting reading, especially the part about the Automatic Tape-Collecting Lathe Ultramicrotome (as someone who builds histology equipment, a 40nm cut is pretty amazing). $\endgroup$
    – Samwise
    Commented Nov 26, 2017 at 7:11
0
$\begingroup$

No

While scientists have successfully, for example, created self-assembling copper wires, that is only the beginning of the "we have that technology" curve. That technology is a long, long, long way (as in 50-100 years away) from being a common-in-our-factories mature technology. Since 20 minutes into the future no corporation can use the example technology, it is unreasonable to assume it is a justification for "why can't we do this."

Your solution for mapping the human brain is a "probe in the nucleus of every neuron." If self-assembling wire is 50-100 years in the future, putting probes on neurons is more like 1,000+ years in the future. A far cry from 20 minutes.

I wish I had better news... sorry....

However

If we throw away the "this afternoon" time requirement, then fundamental answer is "yes." If we could place a probe at the nucleus of every neuron and record the data (that's one whomping accurate GPS, btw), then we could map the brain and, theoretically, know everything about that person: from what their memories are to how they process information, etc.

If you're creating a story with this idea as its basis, "far-future technology today" is a fairly common sci-fi theme. The entire Terminator franchise is based on the idea (basically, A Connecticut Yankee in King Authur's Court) with the solution being to pull future technology back in time. Don't let my answer dissuade you. This can be a cool technology in a good story, even one set 20 minutes into the future.

Improve this answer
$\endgroup$
2
  • $\begingroup$ I wasn't planning on making this a 'common' procedure, the story has the whole mapping spread out over many years, it's just that they start now (and need an aircraft hangar of computers for processing). I was just curious if there was anything stopping someone from starting something like this now. $\endgroup$
    – Samwise
    Commented Nov 26, 2017 at 6:32
  • $\begingroup$ As I said, don't let anything I said stop you. Just because people like me, practiced in the art, can see the hole doesn't mean we're not delighted to suspend our disbelief. It's a great idea. Run with it! $\endgroup$
    – JBH
    Commented Nov 26, 2017 at 6:48

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .