Summary: New research finds that the brain’s internal map of space changes each time we traverse a familiar environment. Even when mice experienced identical virtual mazes with tightly controlled sensory input, hippocampal neurons activated in different patterns on successive runs.
These results indicate that spatial memories are dynamic and continually updated rather than permanently fixed. The pattern of change — often called representational drift — may help the brain encode when events happened as well as where, and the findings have implications for understanding memory, learning and age-related decline.
Key Facts:
- Dynamic maps: The hippocampus updates spatial representations even when the external environment remains the same.
- Representational drift: Different neurons can encode the same path on different repetitions, so the neuronal signature of a given experience evolves across repetitions.
- Age link: Highly excitable neurons stay more stable over time, and because excitability declines with age, this mechanism may relate to age-related memory changes.
Source: Northwestern University
New study from Northwestern neurobiologists shows the brain’s internal GPS changes with each navigation of a stable environment.
Northwestern researchers report that even when animals navigate a perfectly reproducible environment, the specific hippocampal neurons that represent that space shift over time. In other words, walking the same route repeatedly does not always activate the exact same population of “map-making” cells.
The study is scheduled for publication on Wednesday (July 23) in the journal Nature.
“Our results confirm that spatial memories in the hippocampus are not static or permanently fixed,” said Daniel Dombeck, senior author and professor of neurobiology at Northwestern’s Weinberg College of Arts and Sciences. “Rather than finding one immutable group of cells that stores a memory, we see that memory representations move across neurons. The same experience can recruit different cells on different occasions, and that recruitment slowly evolves over time.”
Dombeck led the work with collaborators from his laboratory — Jason Climer, Heydar Davoudi and Jun Young Oh. Climer is a co-first author and is now an assistant professor at the University of Illinois, Urbana-Champaign.
Background: a longstanding mystery
The hippocampus, situated in the temporal lobe, is central to spatial memory and navigation. For many years, researchers assumed that the same hippocampal place cells would encode the same location whenever it was visited. Yet prior imaging studies in mice observed that different neurons often fired when animals ran the same maze on different days, a phenomenon that raised questions about whether those differences reflected true neural change or merely variations in behavior or sensory cues.
“People wondered whether the animals truly experienced the maze the same way each time,” Dombeck explained. “Subtle changes in speed, smell or other environmental factors could have altered neural activity.”
Extreme experimental control: identical sensory input
To address those concerns, the team built on a multisensory virtual reality system previously developed in Dombeck’s lab. That setup provided identical visual cues across runs. Mice ran on treadmills so speed and locomotion were precisely measured, and the researchers standardized olfactory input by placing small cones near the animals’ noses to deliver the same scent conditions each session.
Even with these rigorous controls — identical visuals, measured speed and matched smells — the hippocampal representations still drifted: slightly different groups of neurons represented the same virtual maze during different runs. This confirmed that representational drift is an intrinsic feature of how the hippocampus encodes space, not merely an artifact of changing external conditions or behavior.
“We controlled for everything we possibly could, and yet the neuronal ensemble still changed,” Dombeck said. “I expected to find the opposite, that identical experiences would produce identical neuronal patterns. Instead, representations slowly shift.”
Why it matters for aging and memory stability
Although the drift was widespread, there was a consistent predictor of stability: the intrinsic excitability of individual neurons. Place cells that were more excitable — easier to activate — tended to maintain their representation of the maze across runs, while less excitable cells were more likely to change. Because neuronal excitability diminishes with age, this correlation suggests a possible mechanism linking aging to decreased stability of spatial representations and memory retention.
“A small subset of highly excitable neurons appears better able to hold onto the original memory,” Dombeck noted. “The weaker-firing cells are the ones that change, so some core of the original representation persists in a limited population.”
Dombeck and colleagues are exploring why the brain allows similar experiences to be encoded differently over time. One plausible idea is temporal tagging: by encoding identical events with slightly different neuronal patterns, the brain can separate occurrences that happen at different times, preserving unique memories of each repetition.
“Even if two experiences are identical, they occur at different moments,” Dombeck said. “The brain may deliberately encode them slightly differently so we can recall those separate instances.”
Funding: The study, “Hippocampal representations drift in stable multisensory environments,” received support from the National Institutes of Health (grants R01MH101297, T32AG020506 and 1F32NS116023).
About this memory and neuroscience research news
Author: Amanda Morris
Source: Northwestern University
Contact: Amanda Morris – Northwestern University
Image credit: Neuroscience News
Original research: Closed access. Title: “Hippocampal representations drift in stable multisensory environments” by Daniel Dombeck et al., published in Nature.
Abstract
Hippocampal representations drift in stable multisensory environments
Previous experiments tracking hippocampal place cells in mice navigating the same real-world environment have documented substantial changes in neural representations over days. Whether that representational drift reflects an intrinsic function — for example, to distinguish similar experiences occurring at different times — or is driven by subtle sensory or behavioral differences has remained unclear.
Using a multisensory virtual reality system to exert tight experimental control, the researchers show that differences in sensory environment or behavior do not substantially alter drift rate. They also identify neuronal excitability as the strongest predictor of long-term representational stability: more excitable place cells exhibit less drift. These results demonstrate that representational drift occurs even under highly reproducible conditions and highlight excitability as a key factor in long-term stability of hippocampal memory representations.