New maps of brain activity challenge century-old anatomical boundaries

New research challenges the century-old practice of mapping the brain based on how tissue looks under a microscope. By analyzing electrical signals from thousands of neurons in mice, scientists discovered that the brain’s command center organizes itself by information flow rather than physical structure. These findings appear in the journal Nature Neuroscience.

The prefrontal cortex acts as the brain’s executive hub. It manages complex processes such as planning, decision-making, and reasoning. Historically, neuroscientists defined the boundaries of this region by studying cytoarchitecture. This method involves staining brain tissue and observing the arrangement of cells. The assumption has been that physical differences in cell layout correspond to distinct functional jobs.

However, the connection between these static maps and the dynamic electrical firing of neurons remains unproven. A research team led by Marie Carlén at the Karolinska Institutet in Sweden sought to test this long-standing assumption. Pierre Le Merre and Katharina Heining served as the lead authors on the paper. They aimed to create a functional map based on what neurons actually do rather than just where they sit.

To achieve this, the team performed an extensive analysis of single-neuron activity. They focused on the mouse brain, which serves as a model for mammalian neural structure. The researchers implanted high-density probes known as Neuropixels into the brains of awake mice. These advanced sensors allowed them to record the electrical output of more than 24,000 individual neurons.

The study included recordings from the prefrontal cortex as well as sensory and motor areas. The investigators first analyzed spontaneous activity. This refers to the electrical firing that occurs when the animal is resting and not performing a specific task. Spontaneous activity offers a window into the intrinsic properties of a neuron and its local network.

Free daily newsletter

The team needed precise ways to describe this activity. Simply counting the number of electrical spikes per second was insufficient. They introduced three specific mathematical metrics to characterize the firing patterns. The first metric was the firing rate, or how often a neuron sends a signal.

The second metric was “burstiness.” This describes the irregularity of the intervals between spikes. A neuron with high burstiness fires in rapid clusters followed by silence. A neuron with low burstiness fires with a steady, metronomic rhythm.

The third metric was “memory.” This measures the sequential structure of the firing. It asks whether the length of one interval between spikes predicts the length of the next one. Taken together, these three variables provided a unique “fingerprint” for every recorded neuron.

The researchers used a machine learning technique called a Self-Organizing Map to sort these fingerprints. This algorithm grouped neurons with similar firing properties together. It allowed the scientists to visualize the landscape of neuronal activity without imposing human biases.

The analysis revealed a distinct signature for the prefrontal cortex. Neurons in this area predominantly displayed low firing rates and highly regular rhythms. They did not fire in erratic bursts. This created a “low-rate, regular-firing” profile that distinguished the prefrontal cortex from other brain regions.

The team then projected these activity profiles back onto the physical map of the brain. They compared the boundaries of their activity-based clusters with the traditional cytoarchitectural borders. The two maps did not align.

Regions that looked different under a microscope often contained neurons with identical firing patterns. Conversely, regions that looked the same structurally often hosted different types of activity. The distinct functional modules of the prefrontal cortex ignored the classical boundaries drawn by anatomists.

Instead of anatomy, the activity patterns aligned with hierarchy. In neuroscience, hierarchy refers to the order of information processing. Sensory areas that receive raw data from the eyes or ears are at the bottom of the hierarchy. The prefrontal cortex, which integrates this data to make decisions, sits at the top.

The researchers correlated their activity maps with existing maps of brain connectivity. They found that regions higher up in the hierarchy consistently displayed the low-rate, regular-firing signature. This suggests that the way neurons fire is determined by their place in the network, not by the local architecture of the cells.

This finding aligns with theories about how the brain processes information. Sensory areas need to respond quickly to changing environments, requiring fast or bursty firing. High-level areas need to integrate information over time to maintain stable plans. A slow, regular rhythm is ideal for holding information in working memory without being easily distracted by noise.

The study then moved beyond resting activity to examine goal-directed behavior. The mice performed a task where they heard a tone or saw a visual stimulus. They had to turn a wheel to receive a water reward. This allowed the researchers to see how the functional map changed during active decision-making.

The team identified neurons that were “tuned” to specific aspects of the task. Some neurons responded only to the sound. Others fired specifically when the mouse made a choice to turn the wheel.

When they mapped these task-related neurons, they again found no relation to the traditional anatomical borders. The functional activity formed its own unique territories. One specific finding presented a paradox.

The researchers had established that the hallmark of the prefrontal cortex was slow, regular firing. However, the specific neurons that coded for “choice”—the act of making a decision—tended to have high firing rates. These “decider” neurons were chemically and spatially mixed in with the “integrator” neurons but behaved differently.

This implies a separation of duties within the same brain space. The general population of neurons maintains a slow, steady rhythm to provide a stable platform for cognition. Embedded within this stable network are specific, highly excitable neurons that trigger actions.

The overlap of these two populations suggests that connectivity shapes the landscape. The high-hierarchy network supports the regular firing. Within that network, specific inputs drive the high-rate choice neurons.

These results suggest that intrinsic connectivity is the primary organizing principle of the prefrontal cortex. The physical appearance of the tissue is a poor predictor of function. “Our findings challenge the traditional way of defining brain regions and have major implications for understanding brain organisation overall,” says Marie Carlén.

The study does have limitations. It relied on data from mice. While mouse and human brains share many features, the human prefrontal cortex is far more complex. Additionally, the recordings focused primarily on the deep layers of the cortex. These layers are responsible for sending output signals to other parts of the brain.

The activity in the surface layers, which receive input, might show different patterns. The study also looked at a limited set of behaviors. Future research will need to explore whether these maps hold true across different types of cognitive tasks.

Scientists must also validate these metrics in other species. If the pattern holds, it could provide a new roadmap for understanding brain disorders. Many psychiatric conditions involve dysfunction in the prefrontal cortex. Understanding the “normal” activity signature—slow and regular—could help identify what goes wrong in disease.

This data-driven approach offers a scalable framework. It moves neuroscience away from subjective visual descriptions toward objective mathematical categorization. It suggests that to understand the brain, we must look at the invisible traffic of electricity rather than just the visible roads of tissue.

The study, “A prefrontal cortex map based on single-neuron activity,” was authored by Pierre Le Merre, Katharina Heining, Marina Slashcheva, Felix Jung, Eleni Moysiadou, Nicolas Guyon, Ram Yahya, Hyunsoo Park, Fredrik Wernstal & Marie Carlén.