Next generation Human Neocortical Neurosolver model with biophysically realistic somatic, dendritic and synaptic dynamics

The human brain is a vast, intricate landscape of electrical activity, but for scientists trying to understand how we think, feel, or perceive the world, much of this activity remains hidden. We can place sensors on the scalp to record the brain's electromagnetic whispers, known as electroencephalography or EEG, and magnetoencephalography or MEG. These tools capture the collective hum of billions of neurons firing in sync, yet the signals they record are like hearing a crowd cheer from a stadium without being able to see the individual fans. To understand the story behind the noise, researchers must bridge the gap between these large-scale signals and the tiny, microscopic circuits inside the brain. This requires a way to translate the broad waves of brain activity back into the specific behaviors of individual cells, a task that has long been difficult because the electrical properties of human neurons were poorly understood compared to those of animals.

A team of researchers has now built a significantly improved digital model to solve this problem, offering a clearer window into how the human brain generates its electrical signatures. They created a sophisticated computer simulation called the Human Neocortical Neurosolver, which acts as a translator between the microscopic world of cells and the macroscopic world of brain scans. The new version of this model is special because it is no longer based primarily on data from rats or monkeys. Instead, the researchers tuned the virtual neurons to match the actual electrical behavior of human brain tissue, using fresh data from human donors. By doing this, they have created a tool that can simulate not just the basic firing of neurons, but also complex, hidden events happening deep inside the long branches of brain cells, events that were previously impossible to track in a human context.

The core of this work involves rebuilding the digital architecture of the brain's outer layer, the neocortex, with a focus on two main types of cells: pyramidal neurons, which are the primary messengers, and inhibitory interneurons, which act as brakes to keep the network in balance. In previous versions of the model, these cells were simplified to make the computer simulations run faster, but they lacked the specific electrical quirks found in real human tissue. The researchers updated the model by adjusting the internal conductors and switches of these virtual cells to match recordings from human neurons. They found that human neurons behave differently than their animal counterparts, particularly in how they handle electrical signals traveling from the cell body out to their long, tree-like branches called dendrites.

One of the most significant discoveries in this new model is the ability to simulate calcium spikes, which are bursts of electrical activity that occur deep within the dendrites of certain neurons. In the deep layers of the human brain, specifically in the fifth layer, these neurons can generate powerful calcium spikes when they receive specific patterns of input. These spikes are not just random noise; they are complex events that integrate information from different parts of the brain and can trigger the cell to fire a signal to the rest of the body. The researchers demonstrated that their new model can reproduce these calcium spikes, showing that they happen when the neuron receives a rapid series of signals or when it gets input from two different directions at the same time. This is a crucial addition because these dendritic events are now known to contribute to the large electrical signals we can measure on the scalp, meaning that what we see in a brain scan might be partly driven by these hidden, deep-brain events.

The team also refined how these virtual cells connect to one another. In the brain, neurons do not just fire randomly; they are linked in precise patterns where the strength of the connection depends on how far apart the cells are. The new model incorporates these distance-dependent rules, ensuring that cells closer together influence each other more strongly than those further away. They adjusted the chemical signals, known as neurotransmitters, that pass between cells to match the speed and strength observed in human tissue. This included updating how inhibitory cells silence their neighbors, making sure the timing of these "brakes" matches the slow, deliberate pace found in human brain slices rather than the faster pace of rodents.

With this updated model, the researchers tested their ability to recreate real-world brain patterns. They simulated the brain's response to a sudden touch, a common experiment where a person feels a tap on their skin and their brain reacts with a specific electrical wave. The new model successfully reproduced the shape and timing of these waves, matching the data recorded from human participants. It also simulated the brain's natural rhythms, such as the alpha and beta waves that appear when the brain is at rest or moving. In these simulations, the model showed how the interplay between excitatory and inhibitory cells creates these rhythmic bursts of activity, and how the new, more realistic human cell properties changed the way these rhythms were generated compared to older models.

The researchers also explored how these different parts of the brain contribute to the final signal. They found that the deep layer neurons, with their long branches and ability to generate calcium spikes, play a dominant role in creating the large electrical currents measured by EEG and MEG. The model revealed that the calcium spikes in these deep cells create a sustained electrical pull that lasts longer than the quick flashes of activity in the upper layers. This suggests that the large waves seen on a brain scan are not just a simple sum of all neurons firing, but are heavily shaped by these specific, complex events happening in the deep layers. By visualizing the flow of electricity and the movement of calcium ions inside the virtual cells, the researchers could see exactly how the microscopic actions of individual neurons add up to the macroscopic signals we measure.

This work represents a major step forward in understanding the human brain because it moves beyond guessing based on animal data. The new model, which the researchers call the duecker_ET_model, is built on the actual physiology of human neurons, making it a more reliable tool for interpreting brain scans. It allows scientists to test hypotheses about what is happening inside the brain when we think, feel, or suffer from neurological conditions. For instance, if a patient has an unusual brain wave pattern, researchers can now use this model to see if the issue might be caused by a problem with calcium spikes in the deep layers or a mismatch in the timing of inhibitory signals. The model is open for other scientists to use, providing a shared platform to explore the mysteries of the human mind with a level of detail that was previously out of reach.

The researchers were careful to note that their model is a simplification, designed to be computationally efficient while still capturing the essential behaviors of the brain. They did not include every single type of neuron or every tiny detail of the cell's structure, as doing so would make the simulations too slow to be useful. Instead, they focused on the features that matter most for generating the electrical signals we can measure. They acknowledged that while their model reproduces many features of human neurons, there are still differences between the model and reality, particularly regarding how signals travel along the long branches of the cells. However, by grounding the model in human data, they have created a foundation that is far more accurate than previous attempts.

Ultimately, this paper provides a new lens through which to view the human brain. It transforms the abstract waves of EEG and MEG into a story of specific cells, specific connections, and specific electrical events. By showing that human neurons have unique properties, such as the ability to generate calcium spikes that shape our brain waves, the researchers have opened up new avenues for understanding how the brain works. The model serves as a bridge, connecting the invisible world of cellular biology with the visible world of brain imaging, allowing us to see the machinery behind the mind with greater clarity than ever before.