How an Amazing Implant in the Brain Creates Texts from Thoughts at 90% Accuracy

Imagine sitting down to type, and you open a document. You begin thinking your words and watch them appear on the screen!

This is not as ridiculous as it sounds, according to a new study.

Mind-blowing technology

This breakthrough has been made possible because of a new brain-computer interface (BCI). As a result, individuals suffering from paralysis can now achieve this.

Simply by imagining the writing motions of making letters, a man with a spinal injury translated his thoughts into text. And he did this just as quickly as any average person would type text on their smartphone.

The technology can produce about 90 characters per minute with fantastic accuracy of more than 90% after autocorrect. This fascinating system surpasses all other neural implants by leaps and bounds in its performance.

Powerful algorithm

The keystone of this technology is its algorithm. Its routine is based on an extremely powerful neural network called the recurrent neural network (RNN) and some machine learning technology.

It works because its neural implant employs artificial intelligence to transform a person’s electrical brain signals – which are generated whenever they imagine handwriting text. Then this information is displayed on a computer screen in real-time.

Researchers are delighted that this technology will help empower severely paralyzed people to communicate with the world around them. They see it as a powerful way of letting ‘locked-in’ people express and share their thoughts.

A bonus discovery

As it turns out, ‘mind texting’ could be just the beginning. The study also revealed that artificial intelligence seems better suited for decoding the more complex human behaviors – rather than the simpler ones, as was previously believed.

The implications of this discovery are hard even to imagine. Especially whenever the potential of a brain-computer symbiosis is reimagined.

While scientists admit that there’s much research needed, they consider the results of this study as a significant milestone. It opens the door wide for future brain-computer interface applications. One reason for such optimism is that it uses machine learning methods that are constantly evolving and improving. Thus, upgraded models will always be readily available in the future.

machine learning artificial intelligence

Handless typing

This research is actually is part of the renowned BrainGate project, which has been on the cutting edge of developing neural interfaces for over a decade. This mission has helped those who are paralyzed to restore communications with others. They’ve seen the kind of magic that happens after inserting a chip that’s surgically embedded in the top layer of the brain.

BrainGate has several incredible accomplishments under its belt. They created one implant that allows people to use robotic arms with their thoughts. Another project lets paralyzed people move computer cursors with their minds on a digital tablet.

The inner workings

This has been made possible because a central processor and a motor cortex are still intact after someone suffers a paralysis – at least for simple movements like grasping and reaching. You could compare it to cutting the cable of your wireless router. While you’ll lose Internet access, your network is still in place.

These neural implants connect straight to the source, the electrical brain signals that drive all our movements. Then these signals are decoded and translated into a language that computers can understand. From this point, the information control yet another output – such as a screen cursor or a robotic arm.

Mindtexting precision

The notion of using machine learning algorithms was a brilliant idea because this field is evolving very rapidly. Moreover, it as provides a solid connection between artificial intelligence and neuroscience.

However, every algorithm is dependent on the quality of data it receives. The ironic thing with this project is since they dealt with complex behaviors, the data was very distinguishable. Therefore, the algorithm worked wonderfully. This was when they realized that complex data is better suited for AI applications.

Final thoughts

This new technology is not quite ready for the clinics yet. It needs to be tested on more people and modified by adding a few standard typing functions, like editing and deleting. And the research team intends to give mindtexters the ability to use symbols and capital letters.

Of course, the most important thing is their mission. They want to work with patients who cannot speak or communicate for whatever reason and restore their desire to engage with the outside world. This study has dramatically moved the needle towards that beautiful conclusion.