For decades, microtubules were treated primarily as the structural beams of a neuron. These hollow polymers help maintain axons and dendrites, transport molecular cargo, and organize cell division in other tissues. But a growing body of research is testing a more ambitious possibility: that microtubules also participate directly in intraneuronal signal processing.
The idea is not that microtubules replace synapses or membrane ion channels. Rather, their rapidly changing structures and dense molecular interactions could provide an additional layer of computation inside the neuron—one that operates alongside conventional electrical signaling.
A dynamic substrate, not a static scaffold
Microtubules are built from tubulin proteins that continuously assemble and disassemble. This “dynamic instability” allows the cytoskeleton to remodel in response to activity, development, and injury. Microtubules also bind a large assortment of regulatory proteins, including motor proteins and microtubule-associated proteins, while supporting long-range transport of vesicles, RNA, and organelles.
These properties make them plausible information-processing elements. A local change in microtubule stability could alter cargo delivery, protein localization, or the geometry of a dendrite. Conversely, electrical activity and calcium-dependent signaling can modify microtubule-associated proteins, creating feedback between a neuron’s membrane state and its internal architecture.
Electrical effects remain the controversial frontier
Some experimental and theoretical studies have reported that microtubules can conduct or influence electrical signals through ionic flows along their surfaces or through collective behavior in the tubulin lattice. Other work has proposed that mechanical vibrations, conformational changes, or resonant interactions might affect how signals propagate within the polymer.
These claims remain actively debated. Measurements made in purified proteins or isolated microtubules do not automatically demonstrate a computational role in intact neurons. The intracellular environment is crowded, noisy, and strongly regulated; water, ions, binding proteins, and neighboring cytoskeletal structures can alter the behavior observed in simplified systems.
The most defensible current picture is therefore pluralistic. Microtubules clearly process information indirectly by controlling transport and structural plasticity. Whether they also carry biologically meaningful electrical or vibrational signals over relevant distances is still unresolved.
Why the question is resurfacing now
Modern microscopy, genetically encoded sensors, and single-molecule methods are making it possible to observe cytoskeletal changes with far greater spatial and temporal precision. Researchers can now track microtubule growth, motor traffic, calcium signals, and synaptic activity in the same neuronal compartments. Computational models are also beginning to connect molecular-scale dynamics with changes in dendritic integration and axonal reliability.
This convergence is shifting the debate away from a simple question—“Are microtubules conscious computers?”—toward experimentally tractable ones. Do microtubule modifications alter the timing of cargo delivery after synaptic stimulation? Can activity-dependent lattice changes measurably influence membrane excitability? Are particular microtubule-associated proteins acting as molecular filters, amplifiers, or memory elements?
A cautious expansion of neuronal computation
If microtubules contribute directly to signal processing, the discovery would broaden the scale at which neural computation must be understood. Information would not reside only in spikes, synapses, and network connectivity, but also in the regulated molecular state of each neuron’s interior.
Even without a direct electrical role, the cytoskeleton is already a powerful computational intermediary: it integrates biochemical signals, controls transport routes, and determines how neurons change over time. The emerging research does not overturn established neuroscience. It adds a deeper layer—one in which cellular architecture is not merely the hardware for computation, but part of the computation itself.
