Every year millions of people experiencing the pain of memory loss. There are many reasons: such as a large number of veterans and football player of traumatic brain injury, stroke and alzheimer's disease, such as the elderly Even we all go through the normal brain aging. A loss of memory seems inevitable, but a maverick neuroscientists are trying to electronic therapy. DARPA funded biomedical engineer at the university of southern California Theodore Berger, Ph.D., is testing a memory-boosting implants, the device can imitate to form the new long-term memory signal processing.

This revolutionary implantable devices has help to the memory of mice and monkeys coding (memory encoding), now it is to test human epilepsy patients, it is very exciting beginning, will accelerate the rapid development in the field of memory retrieval. In order to achieve this goal, however, the team's first job is cracked memory encoding.

Decoding the memory

In the beginning, Berger will know that he is a big problem facing. They don't pursue matching memory every detail in the process, but at least a proper model is put forward. "Of course people will ask: can you use a device to simulate it? Can you make this equipment in any brain operation? It is these things make people think I'm crazy. They think it is too difficult." Berger said, but the research team is the solid first step.

Hippocampus is buried deep in the brain circuitry and fold in an area that is the key part of short-term memory into long-term memory. In the center of the hippocampus, memory is produced by a specific number of neurons in a certain period of time, a series of electrical impulses. This is very important, which means that the process can be reduced to mathematical equations, and form a computing framework. For the exploration of the problem, Berger is not alone.

By tracking animal study of neurons activations, neuroscientists began to decipher support memory encoding of information flow in the hippocampus. The key point of this process from (input) of the hippocampus CA3 area to CA1 (the output of the hippocampus) strong electronic signals. Impaired memory in the brain, this signal will be blocked, so if can use silicon chips to create it, we can restore (or even enhance memory.

Close the gap

The brain's memory encoding are hard to crack, the reason lies in the nonlinear characteristic of neural network, signal noise and at the same time are often overlap, this leads to some of the input signal is inhibited or enhanced. Has hundreds of thousands of neurons in a network, any tiny change will be amplified resulting in the great changes in the output. It's like a chaos of the black box. However, with the help of modern computing technology, Berger think he mastered the approximate solutions: with its mathematical theorems for chip programming, and then see if the brain can accept the chip as a substitute or memory module.

Team first in mice with simple tasks. They trained rats to push two lever for delicious food, and then record when the right control lever in mice, the hippocampus CA3 area to the CA1 area a series of electrical impulses. They will details the short-term memory into long-term memory mode of signal conversion, and then use these information (memory is essentially electronic signals) programming on external memory chips.

Then they gave the mice injected a can temporarily block formation and long-term memory storage capacity of the drug, which control lever to push the mice forgot to find food in this matter. Then they will microelectrodes implanted the hippocampus, and later to CA1 area in memory encoding (output) emits pulses. The result is very effective, through the external memory module, mice regained the ability to choose the right control lever.

Encouraged by good result, Berger in monkeys next implant test the memory, the focus on areas of the brain called the prefrontal cortex (prefrontal cortex), the region's function is to receive and adjust the memory of encoded by the hippocampus.

Electrodes implanted in the monkey brain, the team to give the monkeys watching a series of repeated image, and then catch monkeys recognized before seen images of the activity of the prefrontal cortex. Then, they use a lot of cocaine specific brain regions by blocking the monkey's memory. Then, the researchers used guided by "memory encoding" programming of electrode brain signal processing process back on track, the monkey's memory improved significantly. A year later, the team further confirmed their implants can save memory by the hippocampus damaged monkeys and memory loss.

Human memory implants

Last year, the team began to carefully tested its memory implants in human volunteers. Because the brain surgery is risky, team recruited root has been used to track epilepsy electrodes implanted in the brain of 12 patients with epilepsy. Recurrent epilepsy will gradually formed in the damage to the hippocampus, the key part of long-term memory, so if the implant can play a role, also has benefits for these patients.

Team let volunteers have a series of images, 90 seconds after reminding them before can see pictures. With the training, volunteers team recorded the firing pattern area CA1 and CA3 area. Use of these data, the team developed an algorithm (a special human memory encoding), can be predicted according to the input of CA3 area, CA1 area cell activity pattern. Compared with the actual firing pattern in the brain, the algorithm about 80% accuracy.

Although it is not perfect, but it's a good start. Using this algorithm, researchers began to use the input signal analogue to stimulate output cell. Although song dong, assistant professor admit that they have already use this method to cure a women with epilepsy, but he is still very cautious, only said that although the hope, but success is still too early to say. The hippocampus, unlike motor cortex, the corresponding relationship between different parts of the body have a clear its activities cannot be clearly seen, so the song caution is necessary.

"It's hard to explain why the stimulus input leads to predictable results, it's hard to say that the implants can save those hurt by the hippocampus output and memory loss." Dr RIKEN institute for brain science McHugh said, "but the data is convincing." Berger was very happy, and said: "I didn't think it can be used in the human brain."

But the work is still far from complete. The next few years, Berger, want to see if the chip can in different situations to help rebuild long-term memory. Now, after all, the algorithm is based on the team recorded a special case, it is effective for different types of input are? Berger believed and still full of hope, could also say that they will find a perfect adaptation model of most conditions, after all, the brain is limited by its own physical structure, so the electrical signal processing pattern is so much in the hippocampus. "Our goal is to improve the memory badly damaged people's quality of life," Berger says, "if I can restore its ability to form long-term memory half, I'll be happy as hell, the majority of patients will be very happy."