Brainoid

The new progress in the cynomolgus monkey experiment with brainoids came on Christmas Eve.

Just when everyone was still working on the surface but thinking about their loved ones for the holidays, the cynomolgus monkey experiment at the center of "Nimbus" suddenly made a new discovery.

The previous method of connecting electrodes into a rigid chip and inserting it into the brain of cynomolgus monkeys using Qi Min's idea can indeed complete the memory transfer function similar to that of rats. However, after being connected to a cynomolgus monkey for a long time and causing it to become comatose, problems arose when waking up the cynomolgus monkey. They found that after the "memory" is extracted through the connection with the brain-like body, if the "memory" is not transferred back to the cynomolgus monkey, it will not wake up again and lose the "memory" like the rats, but will not be able to do so anyway. be awakened. And when the interval between the two connections exceeds two hours, the cynomolgus monkey cannot be awakened even if it is connected again, and instead becomes a "plant monkey" from then on.

This made Qi Min realize that from rats to cynomolgus monkeys it was not a linear increase from simple to complex. The increase in the number of brain neurons to a certain extent will definitely lead to qualitative changes. This is a new breakthrough, and if she cannot solve the problem of cynomolgus monkeys becoming "plant monkeys" after being connected, the article may also be questioned by others.

So how to solve the problem that the cynomolgus monkey cannot wake up if it does not connect back after being connected? Qi Min stood in front of the operating table in the clean room and thought for more than ten seconds, and soon came up with a simple guess: If the connection of the corpus callosum interface caused organic changes in the cynomolgus monkey's brain, then reducing " Can "dose" be avoided to some extent?

This "dose" can be understood as the total number of interfaces or the strength of the interfaces. Qi Min discussed with Ms. Zhou Xiaoshan. She felt that since the structure of the interface itself is difficult to replace, it might be feasible to reduce the total number of interfaces.

However, Ms. Zhou Xiaoshan said: "I think this may not be the case. You see, if, as you said, the increase in the number of neurons from rats to cynomolgus monkeys causes a qualitative change, then the interface, as a 'channel' that actually collects information, reflects is the number of effective neurons you actually connect to the brain-like body. If you reduce the interface, the number of effective connections will be reduced. By then you may be able to obtain results similar to those of rats, but they may not reflect the order of magnitude of cynomolgus monkeys. Case."

"This... is really true." Qi Min couldn't help but ask, "If we can't reduce the quantity, how can we reduce the intensity?"

Ms. Zhou Xiaoshan frowned and thought for a while, then suddenly said: "Now this interface is not strictly related to single neurons. If it is regarded as a combination of signals... then if a Gaussian transformation is done on the chip, it may be received when The difference between high-resolution and relatively low-resolution signals is also not that great.”

Since each interface itself does not correspond to the signal of a single neuron, it can be regarded as the sum of multiple neurons. After Gaussian transformation is performed on the received signals of all interfaces, the hotspot distribution of neuron excitation and inhibition can be obtained, and its accuracy can be Greater than the accuracy of the interface itself. According to the principles of information theory, if there are infinite interfaces, but the resolution of the interfaces itself is reduced due to mutual noise interference, under ideal circumstances the hotspot distribution obtained after the same Gaussian transformation should be exactly the same as before, and exactly equal to the real neural network. Meta-activity distribution - If the encoding method is lossless, the information itself is not affected.

Of course, since the number of interfaces they use cannot be infinite, the hotspot distribution obtained after Gaussian transformation-which happens to be equivalent to the information transmitted from the brain-will definitely be slightly different from before, but its essence is also the same. The outgoing brain signal, which is also lossyly compressed, should be no different from before.

Qi Min understood: "Do you think we can use non-insertive electrodes?"

"I'm afraid the non-insert type won't work," said Ms. Zhou Xiaoshan. "On the premise of ensuring signal reception, we must also ensure signal transmission - maybe we can try subcutaneous electrodes that are not inserted into the skull."

The idea is simple and relatively easy to implement. But this experiment is still a bit expensive - it requires a new batch of cynomolgus monkeys, which is not cheap.

——

But it's worth the money spent.

After Qi Min and the technicians worked all night to modify the corpus callosum interface chip until December 26, and conducted experiments for three consecutive days, the cynomolgus monkey experiment with weakened interface strength finally succeeded - albeit in a partial sense.

Using the weakened interface, the cynomolgus monkeys no longer fell asleep after being connected for a long time, but they no longer fell into a coma. And when they were disconnected and tested for memory, these cynomolgus monkeys also completely remembered what they had learned before, unlike the rats that had been connected for a long time and performed the water maze. They seemed to have "forgot" their previous memories.

It can be considered that it seems that the brain-like body cannot extract the memory of the cynomolgus monkey through the weakened interface. But the pattern of data changes in the brainoids does bear some resemblance to previous connections using high-strength interfaces.

Qi Min feels that this may not be a problem of inappropriate interface strength. Through the log of the code that runs the partition to clear the data in the brain-like body, we can see that the data traces generated in the brain-like body when using a low-intensity interface to connect to the cynomolgus brain are similar to the traces generated in a short period of time when using a high-intensity interface. However, when the data was not cleared in time on certain occasions, after a delay of several hours, the traces produced in the brain-like body by using a high-intensity interface to connect the cynomolgus monkey brain would spontaneously "spread" and even cause damage to all areas of the brain-like body. data fluctuations.

This gives people a strange feeling. It's like...cynomolgus monkeys "living" in brainoids.

It suddenly occurred to her that if the brain-like body extracts not the "working memory" of the cynomolgus monkey, but "consciousness", then everything can be explained.

If cynomolgus monkeys really have the so-called "consciousness" - should such a complex primate have it, right? The long-term connection of the high-strength interface imports the cynomolgus monkey's consciousness into the brain-like body. Therefore, if there is no return transmission in a short period of time, the cynomolgus monkey's body will not be able to spontaneously generate new consciousness, thus becoming a "plant monkey". The low-intensity interface may not be able to completely introduce consciousness into the brain-like body, so it only produces some traces when the consciousness is connected to the brain-like body, and naturally it will not cause the cynomolgus monkey to become comatose.

So if there is some truth to this wild hypothesis, how can we explain that the cynomolgus monkey's consciousness cannot be transmitted back after more than two hours? The problem probably does not lie with the cynomolgus monkey's body, but may be due to the fact that the consciousness has changed in the brain-like body, a structure different from the brain, and can no longer adapt to the theater's brain. Perhaps it's like the plasticine that originally filled a certain pipe was pulled out of the pipe, spread out and stuck in an unconstrained environment, and became difficult to stuff back into the pipe.

This conjecture is quite ambitious, but the question is how to prove it?

Asking for recommendation votes every day...