In a scientific first, researchers at the University of California, Irvine have uncovered the fundamental mechanisms by which the hippocampus region of the brain organizes memories into sequences and how this can be used to plan future behavior. The discovery may be a critical first step towards understanding memory lapses in cognitive disorders such as Alzheimer’s disease and other forms of dementia.
By combining electrophysiological recording techniques in rodents with statistical machine learning analysis of huge troves of data, UCI researchers uncovered evidence suggesting that the hippocampal network encodes and preserves the progressions of experiences to facilitate decision making. The team’s work is the subject of an article recently published in Nature Communication.
“Our brain keeps a pretty good record of when specific experiences or events occur. This ability helps us function in our daily lives, but before this study we did not have a clear idea of the neural mechanisms behind these processes,” said corresponding author Norbert Fortin, associate professor of neurobiology and behavior at the UCI. with everyone is that this type of memory is highly impaired in a variety of neurological disorders or just with aging, so we really need to know how this brain function works.”
The project, which lasted more than three years, included phases of experimentation and data analysis. The researchers monitored the firing of neurons in the rats’ brains as they underwent a series of odor identification tests. By presenting five different smells in various sequences, the scientists were able to measure the animals’ memory of the correct sequence and detect how their brains captured these sequential relationships.
“The analogy I would think of is computing,” Fortin said. “If I were to stick electrodes in your brain — we can’t; that’s why we use rats — I could see which cells are firing and which aren’t firing at any given time. That gives us insight into how the brain represents and computes information When we record patterns of activity in a structure, it is like seeing zeros and ones in a computer.
Obtained at millisecond intervals over several minutes, measurements of neural activity and inactivity present a dynamic picture of brain functioning. Fortin said he and his colleagues were, in some ways, able to “read the minds” of their subjects by visualizing the “coding” of cells – which fired and which did not – in rapid succession.
“When you think about something, it moves quickly,” he said. “You’re not stuck on that memory for long. Right now it’s pictured, but we can see how it’s changing very quickly.”
Fortin knew early on that readings of hippocampal activity would result in huge amounts of raw data. From the early stages of the project, he called on statisticians from the Donald Bren School of Information & Computer Sciences.
“The neuroscience questions we had at the time in my lab were far too advanced for the statistical knowledge we had. That’s why we needed to involve partners with expertise in data science,” Fortin said.
“These emerging studies in neuroscience rely on data science methods due to the complexity of their data,” said co-lead author Babak Shahbaba, UCI Chancellor’s Fellow and professor of statistics. “Brain activities are recorded on the scale of milliseconds, and these experiments last over an hour, so you can imagine how quickly the amount of data increases. It’s getting to a point where neuroscientists need more techniques. advanced to accomplish what they had imagined but were not able to implement.”
He noted that when neurons encode information such as memories, scientists can gain insight into this process by examining the pattern of peak activity across all recorded neurons, collectively known as the ensemble.
“We discovered that we could process these neural patterns like images, which unlocked our ability to apply deep machine learning methods,” Shahbaba said. “We analyzed the data with a convolutional neural network, which is a methodology frequently used in image processing applications such as facial recognition.”
In this way, the researchers were able to decode the firing of neurons to retrieve information.
“We know what the B scent signature looks like, just like we know A, C, and D,” Fortin said. “Because of this, you can see when these signatures reappear at a different time, such as when our subjects are anticipating something that hasn’t happened yet. We see these signatures replay rapidly as they think about the to come up .”
Shahbaba said the tools and methodologies developed during this project can be applied to a wide range of problems, and Fortin could expand his research area to other brain regions.
The study is an example of the power of convergence research at institutions such as UCI, Shahbaba said: “I could directly see the difference it makes for our students. Researchers from Norbert’s Neuroscience Group are taking data science classes and can now ask really important scientific questions that they couldn’t study in the past, and my own students are fundamentally thinking about the scientific method in an unprecedented way.”
He added: “Through this collaboration, we are training the next generation of scientists, who have the skills required to conduct interdisciplinary research.
Fortin and Shahbaba were joined on the project by Pierre Baldi, professor emeritus of computer science at UCI; Lingge Li, who earned a Ph.D. in statistics at the UCI in 2020; Forest Agostinelli, who earned a doctorate. in computer science from UCI in 2019 and is now an assistant professor at the University of South Carolina; Mansi Saraf and Keiland Cooper, Ph.D. UCI. neurobiology and behavior students; Derenik Haghverdian, holder of a UCI doctorate. student in statistics; and Gabriel Elias, postdoctoral researcher at the UCI. Funding was provided by the National Institutes of Health, the National Science Foundation and the Whitehall Foundation.