Specialized Mechanisms Organizes Time Memories In The Brain. Researchers at the University of California, Irvine have discovered 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 in a scientific first. The discovery may be a critical first step toward comprehending memory failures associated with cognitive disorders such as Alzheimer’s disease and other forms of dementia.
The UCI researchers discovered evidence that the hippocampal network encodes and preserves progressions of experiences to aid in decision-making by combining electrophysiological recording techniques in rodents with statistical machine learning analysis of massive data sets. The team’s work was recently featured in a Nature Communications paper.
“Our brains keep a fairly accurate record of specific experiences or events. This ability enables us to function in daily life, but prior to this study, we lacked a firm grasp on the neuronal mechanisms underlying these processes “Norbert Fortin, an associate professor of neurobiology and behavior at UC Irvine, is the corresponding author. “However, because this type of memory is severely impaired in a variety of neurological disorders and simply with aging, we really need to understand how this brain function works.”
The project, which lasted more than three years, included phases of experimentation and data analysis. The researchers recorded the firing of neurons in the brains of rats while they were subjected to a series of odor identification tests. The scientists were able to measure the animals’ memory for the correct sequence and determine how their brains encoded these sequential relationships by presenting five distinct smells in various sequences.
“The analogy that comes to mind is computing,” Fortin explained. “If I were to implant electrodes in your brain – which we cannot; this is why we use rats – I would be able to see which cells are firing and which are not at any given time. This sheds light on the way the brain represents and computes information. When we record activity patterns in a structure, it’s as if we’re looking at a computer’s zeros and ones.”
Neuronal activity and inactivity measurements taken at millisecond intervals over several minutes provide a dynamic picture of the brain’s functioning. Fortin stated that he and his colleagues were able to “read the minds” of their subjects in some ways by observing the “coding” of the cells in rapid succession — which ones were firing and which ones were not.
“Thoughts move quickly when you’re thinking about something,” he explained. “You won’t be clinging to that memory for long. At the moment, it is represented, but we can see how quickly that can change.”
Fortin recognized early on that recording hippocampal activity would generate enormous amounts of raw data. He enlisted the assistance of statisticians at the Donald Bren School of Information & Computer Sciences from the project’s inception.
“My lab’s neuroscience questions at the time were far too advanced for our statistical knowledge. That is why we needed to collaborate with partners who possessed data science expertise “As Fortin stated.
“These emerging neuroscience studies make extensive use of data science techniques due to the complexity of their data,” said senior co-author Babak Shahbaba, a UCI Chancellor’s Fellow and professor of statistics. “Brain activity is recorded at the millisecond level, and these experiments last over an hour, so you can imagine how quickly the amount of data accumulates. It reaches a point where neuroscientists require more sophisticated techniques to accomplish what they had imagined but were unable to accomplish.”
He explained that when neurons encode information, such as memories, scientists can gain insight into the process by examining the pattern of spiking activity across all recorded neurons, collectively referred to as an ensemble.
“We discovered that we could treat these neural patterns as images, which enabled us to use deep machine learning techniques,” Shahbaba explained. “We used a convolutional neural network to analyze the data, which is a technique that is frequently used in image processing applications such as facial recognition.”
The researchers were able to decode the firing of neurons in order to retrieve information in this manner.
“We know what odor B’s signature looks like, just as we know what odors A, C, and D look like,” Fortin explained. “As a result, you can observe when those signatures reappear at a later point in time, such as when our subjects anticipate something that has not yet occurred. We’re seeing these signatures rapidly replayed as individuals consider the future.”
According to Shahbaba, the tools and methodologies developed for this project can be applied to a wide variety of problems, and Fortin’s research may expand to other brain regions.
According to Shahbaba, the study demonstrates the value of convergence research at institutions such as UCI: “I could immediately see the positive impact this is having on our students. Researchers in Norbert’s neuroscience group are taking data science classes and are now able to investigate some truly critical scientific questions that they were previously unable to investigate, and my own students are thinking fundamentally about the scientific method in a way that has never been seen before.”
“Through this collaboration,” he continued, “we are educating the next generation of scientists in the necessary skills for conducting interdisciplinary research.”
Fortin and Shahbaba collaborated on the project with Pierre Baldi, a UCI Distinguished Professor of computer science; Lingge Li, who earned a Ph.D. in statistics at UCI in 2020; Forest Agostinelli, who earned a Ph.D. in computer science at UCI in 2019 and is now an assistant professor at the University of South Carolina; Mansi Saraf and Keiland Cooper, both of whom are UCI Ph.D. students in neurobiology and behavior; The National Institutes of Health, the National Science Foundation, and the Whitehall Foundation provided funding.