AI Just Learned How To Boost The Brain’s Memory

Silicon Updates

Has it ever crossed your mind that, how exactly the human brain works remains to be the biggest puzzle -yet to be fully understood? Sure it is, even with all the stories about hacking of the brain and so on. In fact, scientists can explain almost anything found in space or deep seas, but when it comes to the gray matter in between our ears, they say it’s so complex that even the brain doesn’t understand itself. But, things are changing now -and the hope is that artificial intelligence (AI) will help us to finally grok our brains.

Having the full details of how the brain functions can have unique benefits. When the possibility of hacking into the brain came into the picture, it ignited lots of hope in relation to treating mental complications. Well, we can’t say that the tech hasn’t achieved much because it has –and one thing that brain hacking revealed to us is that stress is the most destructive thing to human health. However, now, AI is taking a different approach by helping the brain to remember and work faster.

Al Can Decode and Enhance Memory

Source: Netdna-ssl

In a research paper published at Nature Communications –which also happens to be the latest issue, a team of scientists from the University of Pennsylvania, led by Michael Kahana -reveals how machine-learning algorithms can be used to first decode, and then enhance human memory.  Well, you may want to know immediately how that is even possible. The system does that by delivering precisely timed pulses of electric charge to the brain.

The basic idea here is manipulating the black boxes found in the brain by powering one black box to unlock the potential of the next one in line and subsequently trigger the rest. Well, so far, this sounds like a good idea because at least it is possible to influence the brain or help it become active –based on the data pattern in an AI algorithm system. The problem is that it is not clear yet whether there are side effects to it.

How AI Learned To Boost The Brain’s Memory

Source: Kahana ET A

Kahana who is a psychologist, together with his team, collaborated with 25 epilepsy patients to undertake are research with aims of seeing if it’s possible to boost brain memory, by studying its patterns. They then collected the data and fed it into machine learning to create an AI brain Algorithm. Each patient had 100 to 200 electrodes inserted into their brains. The goal was to monitor electrical activities related to the seizure. The set up was placed in a way that it records high-resolution brain signals during memory tasks.

Building the Algorithm and Working the System

After setting up implants, Kahana states that his team got a sense of what it exactly looks like when the mind is memorizing stuff. At the same time, the team was busy gathering thousands of voltage variations and measurement per second from the electrodes that come from the patient’s brain. In other words, the patients were given books to read and memorize the words in a bit to internalize them. This is the point where the higher voltage was collected. Later, they tested the percentage of how each patient recalled his potion -to build up data that directly connects to the patterns associated with the brain activity -that resulted in either remembering or forgetting words.

After repeating the above procedure over, and over -the team managed to collect enough data, which they used to create patient-specific algorithms that could foretell the words a patient can most likely remember -in connection with their electrode activity.

The Kicker

Source: Cell

However, the most interesting part of the research is that the electrodes did beyond just reading the neural activity –they also played a role in stimulating the activity. As in, after a few seconds, when the brain sees a new word, the trained algorithm was able to decide whether the brain was in the position to remember it or not. If not, the loop system was able to trigger a stimulation that boosted the brain’s capability to act.

The bottom line is, while this sounds more mechanical than biological, a deeper look into this will artificial intelligence will help improve brain science as well as empower AI research. In other words, after understanding the brain, AI engineers who specialize in creating brains for robots will have a better picture of how exactly a true human brain transacts. In the event, that will help them create the long-awaited REAL AI, –one that could do everything without human necessary at any particular point or stage.

Comments

comments