Bold claim, but true: the way a baby’s brain learns to track its own thoughts and movements could completely change how we understand development—and most people have never even heard about how this actually works.
Title, authors and affiliations
Neuronal Activity: Keeping track of moving targets.
Authors:
- Renata Batista-Brito
- Geoffrey Terral
Affiliations:
- Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, United States
- Department of Psychiatry and Behavioral Sciences, Albert Einstein College of Medicine, United States
- Department of Genetics, Albert Einstein College of Medicine, United States
Here is the fascinating part: Jure Majnik and colleagues have developed a clever way to follow the exact same neurons in young mice, day after day, while their brains are rapidly changing during early life.
A rapidly changing newborn brain
In the first days after birth, the brain is in constant motion—cells grow, circuits reorganize, and activity patterns shift as animals move from simple sensory responses to active exploration of the world.
Traditionally, scientists have tried to understand this process by taking snapshots: they record neuronal activity at a few separate ages and then compare those snapshots.
But this “before and after” style can miss what is really happening in between, because it does not show how individual neurons themselves evolve over time.
Here’s where it gets controversial: can we truly claim to understand brain development if we never watch the same neurons grow and change, but instead only compare different cells at different ages?
Why tracking the same neurons is so hard
Manually, researchers can sometimes track single cells in small, sparse populations over several days just by visually inspecting images and annotating them by hand.
However, when the goal is to follow dense populations of hundreds of neurons, especially in very young animals, things quickly become extremely difficult.
The cortex of a developing animal does not stay still: it expands, stretches, and remodels rapidly as the brain grows.
On top of that, individual neurons change shape, some cells disappear, and their relative positions shift as the surrounding tissue moves and matures.
Methods that work quite well in adult animals—where the structure is more stable—often fail when applied to the squirmy, changing brains of young pups.
And this is the part most people miss: the technical difficulty of tracking cells over time can silently bias what we think we know about development.
The two-photon imaging strategy
To overcome these challenges, Majnik and colleagues at INMED, INSERM, and Aix-Marseille University (with Jean-Claude Platel and Rosa Cossart as joint corresponding authors) turned to a powerful imaging approach known as two-photon calcium imaging.
Using this method, they repeatedly imaged the same cortical region in mouse pups every single day from postnatal day 8 (P8) to postnatal day 14 (P14).
This developmental window is especially important: during this period, the brain undergoes dramatic anatomical and functional changes that lay the groundwork for later perception and behavior.
Rather than simply taking images and trying to line them up against a fixed reference map, they combined this imaging strategy with a novel computational tool called Track2p.
This tool attempts to align images across days in a way that respects natural growth, instead of fighting against it.
By aligning each day’s image to the previous day—rather than forcing all days to match a rigid template—the method smooths out distortions caused by tissue expansion and growth.
How Track2p matches neurons across days
Once the images are aligned in this growth-aware way, the next step is to figure out which shapes correspond to which neurons over time.
Track2p does this by comparing both the positions and the shapes of neuronal signals in the images.
It then estimates how much each neuronal “footprint” overlaps from one day to the next, with stronger spatial overlap suggesting a higher probability that two shapes belong to the same cell.
Majnik and colleagues did not simply trust the algorithm blindly.
They carefully validated Track2p’s performance by comparing its assignments to meticulous human labeling.
The result: the automated method performed comparably to trained experts, giving researchers confidence that they can scale up from a handful of cells to hundreds.
From a few neurons to hundreds
Tracking a few isolated neurons by hand, day after day, is possible—but it is slow, exhausting, and does not scale to large populations.
Extending that manual process to hundreds of neurons in each animal would demand enormous time and effort, making many experiments practically impossible.
The method developed by Majnik and colleagues breaks through this barrier.
With Track2p, scientists can now follow hundreds of neurons in the same brain as it develops, opening the door to detailed studies of how activity patterns change over time.
This is not just a convenience; it fundamentally changes the kinds of questions researchers can ask.
Instead of asking “How do brains at age X differ from brains at age Y?” they can now ask “How does this particular neuron’s activity evolve as the animal matures?”
An abrupt shift at P11
Using this new approach, the team uncovered a striking and somewhat surprising result: around postnatal day 11 (P11), neuronal activity in the cortex undergoes an abrupt transition rather than a slow, gradual drift.
Before this time point, neurons tend to fire together in large, highly synchronized bursts—like a crowd chanting in unison.
After P11, this picture changes dramatically: activity becomes more diverse, more decorrelated, and less dominated by big, synchronized events.
Alongside this shift in synchrony, the patterns of neuronal activity become more complex.
This increased complexity suggests that the circuit is gaining the ability to encode richer and more detailed information.
Here is a subtle but provocative point: this kind of transition hints that the brain might pass through discrete developmental “states,” rather than just getting incrementally better at processing information.
Do you think development is more like a series of switches being flipped, or a smooth, continuous dial turning up?
Linking brain activity to movement
The study also sheds light on how the developing cortex begins to respond to the animal’s own behavior.
By combining calcium imaging with video recordings of the pups, the researchers could examine how neuronal activity related to movement.
During the early days of this developmental window, the pups’ movements had little measurable effect on cortical activity patterns.
After P11, however, a clear relationship between movement and neuronal firing emerged.
Some neurons became more active when the pups moved, while others reduced their firing during movement.
Importantly, these movement-related patterns were not random: they were stable over time and could be used to predict the animal’s behavior from neural signals alone.
Because the same neurons were tracked across the entire developmental period, Majnik and colleagues could pinpoint when these movement-linked patterns first appeared and show that they remained consistent as the brain continued to mature.
Multiple developmental processes are likely to influence the timing of this shift, including the maturation of inhibitory circuits, changes in neuromodulatory systems, and the onset of more active, self-generated exploration of the environment.
Why longitudinal tracking really matters
One of the key messages of this work is that important developmental transitions can easily be overlooked if neuronal activity is only examined at separate, isolated ages.
Looking at different sets of neurons at P8 and P14, for example, might suggest a smooth change, even if individual neurons actually undergo an abrupt reorganization around P11.
By following the exact same cells over multiple days, the study reveals dynamic changes that snapshot-based experiments simply cannot capture.
Beyond the biological discoveries, the methodological advance is itself a major contribution.
Track2p is open-source and designed to be user-friendly, which means other research groups can adopt and adapt it for their own developmental studies.
This opens exciting possibilities for monitoring neuronal development at unprecedented resolution, not only in mice but potentially in other models and brain regions.
Here’s a thought-provoking angle: if more labs start using longitudinal tools like Track2p, could some established “textbook” ideas about brain development end up being revised or even overturned?
Publication details and licensing
Article and author information:
- Authors: Renata Batista-Brito and Geoffrey Terral
- Year of publication: 2025
- Title: Neuronal Activity: Keeping track of moving targets
- Journal: eLife, volume 14, article e109627
- DOI: https://doi.org/10.7554/eLife.109627
This article is distributed under the terms of the Creative Commons Attribution License (CC BY 4.0).
This license allows unrestricted use, distribution, and redistribution of the work, as long as the original authors and source are properly credited.
Metrics, downloads and citation tools
At the time described, the article has:
- Citations: 0
- Views, downloads, and citations aggregated across all published versions on the eLife platform
Readers can access:
- Download options to obtain the full article (for example, as a PDF)
- Links to open the article’s references in various online citation-management services
- Tools to download citation information in formats compatible with different reference managers
Author names for citation:
- Renata Batista-Brito
- Geoffrey Terral
Citation format example:
Batista-Brito R, Terral G (2025) Neuronal Activity: Keeping track of moving targets. eLife 14:e109627. https://doi.org/10.7554/eLife.109627
A question for you
Now for the part that might spark debate: if key developmental changes are abrupt and tightly timed, should more neuroscience studies move away from simple age-group comparisons and instead focus on tracking individual neurons or circuits over time?
Do you agree that many classic developmental studies may have missed critical transitions because they did not follow the same cells longitudinally, or do you think age-based snapshots are still enough for most questions?
Share where you stand—does this paper change how you think brain development should be studied?
