For over a century, neuroscientists have searched for the physical trace of memory, the elusive “engram,” or neural pattern believed to encode experiences in the brain. Now, researchers may have uncovered something even more intriguing: a mathematical reason why human perception and memory seem tuned to a particular limit.
In a new study published in Scientific Reports, researchers from King’s College London, Loughborough University, and the Skolkovo Institute of Science and Technology have developed a theoretical model showing that the brain’s ability to form and retain memories might depend on the dimensionality of the sensory world it perceives.
The results suggest that there is an optimal number of sensory dimensions—around seven—where memory storage reaches its peak efficiency before declining in higher dimensions.
“One of the intriguing consequences of the model is the apparent existence of the optimal number of senses in evolving neural and neuromorphic systems,” the researchers write. “The largest capacity of the conceptual space, that is, the most rich perception of the external world, would be attained when the number of senses is equal to 7 — the critical dimension where the number of survived/retained different concepts is maximal.”
At its core, the study poses the question: How many senses does an intelligent system need to remember the most about its environment?
To explore this idea, the researchers built what they call a “kinetic model” of memory engrams. These are mathematical representations showing how memories form, change, and fade within a conceptual space—something like a multidimensional map of experiences. In this model, each engram acts almost like a living object that grows, shrinks, or merges depending on how often it is “hit” by stimuli such as sights, sounds, or sensations from the world around it.
“If each feature is associated with a different sense, the critical dimension corresponds to an optimal number of senses for a system aiming at keeping the maximal number of different concepts in its memory,” researchers write.
In essence, if every sense corresponds to a new dimension in perception, then there appears to be a natural limit—seven—beyond which the brain’s ability to store distinct concepts begins to drop off.
The research builds on the century-old notion of the engram introduced by German biologist Richard Semon in 1904. For decades, neuroscientists have tried to locate these neural “memory traces” through brain-imaging and optogenetic experiments, identifying specific neuron clusters that reactivate during recall. However, while these studies revealed where memories may reside, they could not fully explain how memories evolve and compete over time.
This new study bridges that gap through mathematics. Using Monte Carlo simulations and analytical solutions, the researchers modeled how memory engrams behave when exposed to a constant barrage of stimuli.
In their simulations, memories formed when multiple sensory impressions clustered together, becoming stronger with repeated exposure. Left unstimulated, they slowly expanded and lost focus, a metaphorical “forgetting.”
This push and pull between remembering and forgetting turned out to create a steady balance. However, when the team simulated engrams existing in spaces with different numbers of sensory dimensions, they discovered something surprising.
As the number of dimensions increased, so did the number of unique memories—up to a point. Beyond the seventh dimension, memory capacity began to decline, with overlapping and interference between engrams reducing the system’s efficiency.
“As we consider the ultimate capacity of a conceptual space of a given number of dimensions, we somewhat surprisingly find that the number of distinct engrams stored in memory in the steady state is the greatest for a concept space of seven dimensions,” co-author and professor at Skoltech AI, Dr. Nikolai Brilliantov, said in a press release. “Hence the seven senses claim.”
This “critical dimension” may help explain more than just memory mechanics. It could also reflect why biological sensory systems evolved the way they did.
Humans are traditionally thought to have five senses—sight, hearing, smell, taste, and touch—but neuroscience now recognizes a few more, including proprioception, our awareness of body position, and equilibrioception, the sense of balance.
The idea that cognition might peak around seven inputs isn’t entirely new. Psychologist Dr. George A. Miller famously proposed in 1956 that the average human can hold about “seven, plus or minus two” pieces of information in working memory at once. This new model offers a potential physical and mathematical rationale behind that long-observed cognitive limit.
According to researchers, the balance between sensitivity and precision, as well as the trade-off between being open to new experiences and preserving sharp, distinct memories, may reflect a universal principle. Systems that are highly receptive to new information tend to form blurred, overlapping memories. In contrast, systems that are overly selective risk missing new experiences altogether.
“The higher the receptivity, the less sharp the learned concept becomes,” the researchers note, comparing this tension to the bias-variance trade-off in machine learning, where systems must balance generalization against overfitting.
In biological terms, the finding suggests that evolution may have “tuned” human sensory capacity to the sweet spot where perception, learning, and memory remain maximally efficient. Adding more senses, or sensory dimensions, might not necessarily improve cognition but could instead overload the brain’s conceptual space, causing interference between memories.
The study’s kinetic framework for engrams could inform artificial intelligence and neuromorphic computing, where memory systems modeled on biological brains must balance learning flexibility with information stability. An AI system with too many input channels might experience the same problem as a biological brain exceeding its optimal sensory dimensions, with saturation leading to confusion rather than insight.
“In addition to revealing and justifying the existence of critical dimension in the space of concepts, the proposed kinetic model of the dynamics of memory engrams may offer novel interpretation of empirical data, both existing and new,” the researchers write. “For example, one can empirically examine and explore the mechanisms of engrams’ merging and fragmentation.”
Researchers suggest experimental ways to test their theory. For example, by presenting subjects with sequences of sensory stimuli and measuring how distinct their memory traces remain, scientists could estimate real-world equivalents of the model’s variables—such as the rate of learning and forgetting.
Ultimately, the study offers a tantalizing perspective: that the architecture of human perception—the way we see, hear, feel, and balance—may not just be a product of biology, but a reflection of deep mathematical laws governing how systems remember.
“Our conclusion is, of course, highly speculative in application to human senses, although you never know,” Dr. Brilliantov added. “It could be that humans of the future would evolve a sense of radiation or magnetic field.”
Tim McMillan is a retired law enforcement executive, investigative reporter and co-founder of The Debrief. His writing typically focuses on defense, national security, the Intelligence Community and topics related to psychology. You can follow Tim on Twitter: @LtTimMcMillan. Tim can be reached by email: tim@thedebrief.org or through encrypted email: LtTimMcMillan@protonmail.com