Neurons are fickle. Electric fields are more reliable for information

Electric fields can represent information stored in working memory, allowing the brain to overcome “representational drift” or the inconsistent involvement of individual neurons.

Brain - artistic effect.

Brain – artistic effect. Image credits: Pixel, CC0 Public Domain

As the brain attempts to take into account information, such as a list of groceries that we need to buy on the way home, a new study suggests that the most consistent and reliable representation of that information includes personal Neurons do not have electrical activity, but an overall electric field that they collectively generate.

In fact, whenever neuroscientists have looked at how brains represent information in working memory, they have found that from trial to trial, even when repeating the same task, the involvement and activity of individual cells. varies (a phenomenon called “representational drift”)

in a new study neuroimageScientists from The Pickover Institute for Learning and Memory at MIT and the University of London found that regardless of the specific neurons involved, the overall electric field generated provides a stable and consistent indication of the information animals are tasked with remembering. went.

Caption: Based on electrode measurements of electrical activity in individual brain cells, researchers mathematically estimated the amplitude of a neural electric field at each electrode in an 800 millisecond time frame. They found that the regions representing information held in working memory more reliably and stably than direct measures of neural activity. Image credit: Dimitris Pinotis

In a sense, once established, the field imposes itself on neurons like conductors of an orchestra in which each neuron is a solo musician, says Dimitris Pinotis, lead and corresponding author of the study. Even if the musicians change, the conductor coordinates to produce the same result as in the chairs.

“This ensures that the brain can still function even if some neurons die,” says Pinotsis, an associate professor at the University of London and a research associate at the Picower Institute. “The field ensures that a group of neurons receives the same output even after different parts change. The brain doesn’t need individual neurons, just the conductors, the electric field, to be the same.”

Co-author Earl Miller, the Pickover Professor of Neuroscience in MIT’s Department of Brain and Cognitive Sciences, says that electric fields can provide the brain with a level of information representation and integration that exceeds the level of individual details encoded by single neurons or circuits. is summed up. ,

“The brain may employ this mechanism to operate on a more holistic level, even if the details drift,” he says.

Measurement and Mathematical Modeling

In the study, Pinotis and Miller tested whether the electric field was constant and whether it contained task-related information. To do this they used a combination of direct measurements of neural activity made in animals as they performed a working memory game, and subsequent mathematical analysis to isolate and estimate the electric fields associated with the task. It was not possible to directly measure the electric fields alone, says Miller, because the implanted electrode arrays they use measure nerve activity individually and the EEG electrodes that sit on the outside of the skull are too wide and normal to reflect the pattern. Huh. Specific information presented by a small group of neurons.

“You have to record every detail and then do the mathematically necessary half-step,” Miller says.

During the game the animals were shown a dot in one of six positions on the edge of a screen which would then be blank. After a brief pause, or delay period, they must direct their gaze from the center of the screen to the position they had just marked. During the pause, while the animals were to keep the direction directed, Pinotsis and Miller were recording the electrical activity of neurons on the surface of the brain.

As expected, there was a lot of noise in the signal even when comparing the rounds of the game where the recall condition was the same. At first, in line with representational drift, the involvement of individual neurons varied, but the electrodes picked up activity not only from neurons involved in the task but also from cells working on unrelated things.

Therefore, to isolate game-related patterns among this inconsistent jumble of neural activity, Pinotis developed a mathematical strategy of tracking correlated activity between neurons during the latency period. By determining which neurons were working together and in unison to perform a task, he could determine their connectivity and therefore the information flowing between them. From there, using standard principles of biophysics, they calculated the electric field their activity was producing around the brain surface patches they occupied.

As Miller quipped, “the regions were ‘above’ the brain, but still ‘of the brain.

Importantly, the projected electric fields exhibited properties demonstrating that they represented information retained in the animal’s memory. For example, they were more consistent than the underlying neural activity when the direction of recall was the same. They also differed in different but consistent ways depending on which cited situation was to be remembered – far more so than neural activity. And when the scientist trained software called a “decoder” to predict which direction the animals were in attention, the decoder was relatively better at doing it based on electric fields based on neural activity.

That’s not to say that variation between individual neurons is meaningless noise, Miller says. The thoughts and sensations people and animals experience, even when they repeat the same task, can change minute by minute, causing different neurons to behave differently than they just did. It is important for the memory task to be completed that the overall region remains consistent in its representation.

“This stuff we call representational drift or noise, may be the actual computation that the brain is doing, but the point is, at the next level of the electric field, you can get rid of that drift and just get the signal. are,” Miller says.

The researchers speculate that this region also appears to be a tool the brain may employ to fabricate information flows to ensure desired outcomes. By implying that a particular region emerges, it directs the activity of the participating neurons.

In fact, it’s one of the next questions scientists are investigating: Could an electric field be a means of controlling neurons?

“We are now using this work to ask whether information flows from the macroscale level of an electric field to the microscopic level of individual neurons,” Pinotis says. “To draw an analogy with orchestra, we are now using this work to ask whether the style of a conductor alters the way an individual member of an orchestra plays his instrument.”

Written by David Orenstein

Source: Massachusetts Institute of Technology

Leave a Reply

Your email address will not be published.