Human Brain Project: Neurons in the Brain Follow a Mathematical Pattern
Scientists involved in the Human Brain Project have discovered a mathematical rule that determines how neurons are distributed in our brains. This rule predicts how neurons are spread across different brain regions and could help researchers create accurate models to better understand brain function and develop new treatments for neurological disorders.
In the fascinating world of statistics, the logarithm of a continuous random variable often follows what is known as a log-normal distribution. Defined by its mean and standard deviation, this distribution is visualized as a bell-shaped curve, but it is wider than the normal distribution.
A team of researchers from the Jülich Research Center and the University of Cologne in Germany found that the number of neurons in the outer layer of nervous tissue in various mammals fits a log-normal distribution. The key difference is that the normal distribution’s bell curve is symmetrical, while the log-normal distribution is asymmetrical with a heavy right tail, due to many small values and a few significantly larger ones.
The structure and function of the brain depend on the number and arrangement of neurons. The density of neurons in different regions and layers of the brain’s outer tissue—the cerebral cortex—varies greatly.
“The distribution of neuron density affects network connectivity,” says neuroscientist Sacha van Albada from the Jülich Research Center. “For example, if synapse density is constant, regions with lower neuron density will receive more synapses per neuron.”
Until now, the statistical distribution of neuron density was largely unknown, even though research has provided fascinating insights into the cellular makeup of our brains.
To conduct their study, the team used nine open datasets covering seven different species: mice, marmosets, macaques, galagos, owl monkeys, baboons, and humans. When comparing neuron density in different regions of the cortex, a common pattern of log-normal distribution emerged.
“Our results are consistent with the observation that a surprisingly large number of brain characteristics follow log-normal distributions,” the authors write in their paper.
The log-normal distribution is a natural result of multiplicative processes, just as the normal distribution is a natural result of adding many independent variables.
“The prevalence of the log-normal distribution in nature can be explained by the fact that it arises from multiplying many independent variables,” notes Alexander van Megen, the lead researcher on this topic as part of his doctoral dissertation in computational neuroscience at the Jülich Research Center.
The researchers suggest that the way the cortex is organized may be a byproduct of development or evolution, not directly related to computation. However, previous studies suggest that variations in the brain’s neural network are not just byproducts and may actively help animals learn in changing environments. The presence of the same brain structure across different species and in many brain regions points to the special significance of the log-normal distribution.
“We can’t be sure how the log-normal distribution of neuron density affects brain function, but it’s likely related to high network heterogeneity, which could be computationally beneficial,” explains study co-author Aitor Morales-Gregorio, a computational neuroscientist at the Jülich Research Center.
Scientists hope this discovery will shed light on how the brain stores and retrieves information, as well as how it acquires new knowledge. In the ongoing search for effective treatments for brain diseases, this could be a starting point for developing new drugs targeting specific brain regions.
The decade-long Human Brain Project, aimed at creating a shared research infrastructure to advance neuroscience, computing, and brain-related medicine, is coming to an end, having brought many interesting discoveries along the way.
The research was published in the journal Cerebral Cortex.