Musk's brain-computer interface can be made with a Raspberry Pi?

Click on "Uncle Wheat" above and select "Top/Star Public Account"

Welfare dry goods, delivered as soon as possible

Hello everyone, my name is Wheat.

Everyone must have heard of Musk, as well as Tesla, SpaceX, and brain-computer interfaces.

The first two are too difficult, but now you only need a Raspberry Pi board to make a low-end brain-computer interface?

No kidding, the Russian little brother Rakhmatulin really did it.

This down-to-earth device uses only a Raspberry Pi board as a processor and can process eight brain electrical signals in real time:

If you don't believe me, take a look at the live signal image:

These regular waveforms are the electrical signals that the brain produces as it performs activities.

At present, this project is open source, and the cost is not high. If you want to make a brain-computer interface yourself, you can't miss it, come and learn it~

Project address : https://github.com/Ildaron/EEGwithRaspberryPI

Raspberry Pi brain-computer interface

Before making the device, we need to have a rough framework.

The framework of the brain-computer interface is well understood: first, bioelectrical signals are read from the skull, then sent to the processor for processing, and finally the output signals are used to control other devices.

So the first step is to read the biological signals in the brain. For this purpose, Rakhmatulin made a small cap (the left side of the picture below) with 8 electrodes on it (the right side of the picture below).

The positions of these 8 electrodes are determined according to the electroencephalogram detection electrode positions of the International 10-20 system (as shown in the figure below), where the first letters represent different regions of the brain, such as F for the frontal lobe, P for the parietal lobe, and T for the Temporal lobe, etc.:

△ Legend: International 10-20 System

When the brain is performing different activities, it also produces distinct patterns of electrical signals at the same time. In this way, these electrodes can detect electrical signals for subsequent analysis and processing.

After having the brain signal, the next step is to process the signal. The third or fourth generation of Raspberry Pi can be used here. The following circuit diagram shows the circuit structure on the small yellow board:

8 of the 10 pins on the left in the above figure are connected to the 8 electrodes on the cap, 1 is connected to the reference level, and the other is connected to the offset signal.

The pins on the right are used to transmit data, with a sampling rate of 250 times per second to 16,000 times per second.

Rakhmatulin has open sourced the real-time signal detection and processing code on GitHub.

After connecting the power supply, the monitor, and putting on the small hat, the hardware part is ready:

△ Legend: Hardware part

Now all we have to do is do some simple movements to get the brain to generate electrical signals.

In the test, Rakhmatulin used the two movements of "chewing" and "blinking."

Every time I chew, there will be a peak in the electrical signal. In order to clearly see the changes in brain signals from the waveform, the little brother has done four groups of chewing movements, and each group chews 4 times and 3 times in turn. , 2 strokes, 1 stroke, forming a series of peaks.

Then the eight signals (from top to bottom) generated by the eight electrodes are recorded by the processor as shown in the figure below:

It can be clearly seen from the above figure that each action has a corresponding electrical signal peak, which is arranged according to the number of actions.

After the chewing action, the little brother did four sets of blinking actions, and the number of blinks in each set was 4, 3, 2, and 1 in turn.

Likewise, these electrical signal changes were recorded (pictured below). The first four groups of fluctuations are the electrical signals of chewing, and the last four groups of fluctuations are the electrical signals of blinking:

After passing the band-pass filter, the waveform is clearer, 4, 3, 2, 1, 4, 3, 2, 1... It's like the brain signal is doing radio gymnastics:

At this point, the mission of the brain-computer interface has been completed. The signal output from the processor can be connected to other devices to form different instructions, such as commanding robotic arms, toy cars, drones, and so on.

how about it? Is not it simple? There is no complicated hardware equipment in the whole process, you can do it at home, and you can start it if you have the conditions~

If you really don't want to do it, you can also pay attention to the announcement on the official website of Xiaoge. The crowdfunding event will be launched soon:

About the Author

The developer of this project, Ildar Rakhmatulin, is an electrical engineer and a graduate of the South Ural State University in Russia.

Before making the brain-computer interface, this guy also used the Raspberry Pi to make a processor to detect mosquitoes, and can also emit lasers to eliminate mosquitoes.

Most recently, he and his colleague Sebastian Völkl run a company that hopes to provide people with affordable brain-computer interface devices.

Reference link:

[1]https://github.com/Ildaron/EEGwithRaspberryPI
[2]https://www.hackerbci.com/
[3]https://arxiv.org/abs/2201.02228