A living input for a moving machine
Imagine a robot whose movements change with the activity of a living fungal culture. The intriguing part is the connection between two systems: a growing organism and a machine built to act on measurements. What would it take to make that connection useful?
In a peer-reviewed study published in Science Robotics on 28 August 2024, Anand Kumar Mishra and colleagues linked fungal electrical activity to robotic movement. This was a laboratory demonstration of biohybrid control. The living material was king oyster mycelium, Pleurotus eryngii, grown on agar. [1]
What mycelium actually is
A mushroom is one conspicuous structure that some fungi produce. Much of the organism’s vegetative body consists of fine filaments called hyphae. A connected mass of those filaments is mycelium. It can spread across a surface or through material, placing living cells in contact with their surroundings. These filaments are fungal structures, rather than plant roots. [3]
Fungi obtain carbon and energy from organic compounds. They release enzymes that break down material outside their cells, then absorb smaller molecules. Their cells have membranes, nuclei and mitochondria; chitin contributes to their cell walls. That biological complexity matters when thinking about a living sensor: a culture has nutritional and environmental requirements, rather than simply a factory specification. [3]
The familiar mushroom silhouette is therefore a poor guide to the organism’s full structure. A pale network spread through a culture dish may be less conspicuous, but it is living fungal tissue, with many cells exposed to changing local conditions.
Electrical activity does not require a brain
Fungal electrical recordings predate this robot experiment. In a peer-reviewed 2018 study, Andrew Adamatzky recorded changing extracellular electrical potentials from fruiting bodies of pink oyster fungi, Pleurotus djamor. The work described spontaneous spike trains and responses to thermal stimulation. This was a different species and a different fungal structure from the 2024 robot’s king oyster mycelium. [4]
“Extracellular” means that electrodes measure electrical changes outside cells. A recorded voltage pattern is an observation; identifying its physiological origin and function is a further scientific task. A spike-shaped trace alone does not establish a nervous system, a language or an intention. The useful question for engineering is narrower: can a biological measurement provide a repeatable input for a defined operation?
How the culture entered the control loop
Electrodes recorded the culture, an interface reduced vibration and electromagnetic interference, and a controller converted patterns into actuator commands. The researchers built a walking soft robot and a wheeled robot. Ultraviolet exposure altered movement; engineered control could override the fungal input. The tissue provided signals, while artificial actuators produced movement. [1] [2]
Think of this as a translation chain. A measurement becomes a command only because engineers define how to interpret it. Changing that translation could change the machine’s behaviour without changing the organism. This distinction helps explain why a moving robot is not, by itself, evidence that a fungus has learned to drive.
Why confirming the signal matters
Independent fungal electrophysiology experiments illustrate the importance of controls. In a 2024 BioSystems paper, Davin Browner and Andrew Adamatzky used microelectrode arrays to record biofilms of Hericium erinaceus, the lion’s mane fungus. They examined low-frequency extracellular potentials, including activity at or below 1 hertz. One hertz means one cycle per second. [5]
The researchers compared cultures with medium controls and applied nystatin, an antifungal that disrupts membrane function. Electrical activity ceased after treatment, supporting a biological contribution to the recordings. That result does not establish the mechanism of every fungal spike, and it does not independently validate the robot’s complete controller. It demonstrates the kind of experiment needed to distinguish living activity from a signal merely recorded near living material. [5]
Later work sharpened that measurement problem. In a peer-reviewed iScience study published online on 2 September 2025, Matteo Buffi and colleagues developed circuit boards with paired electrodes for fungal recordings. They used Fusarium oxysporum, shielding and frequency analysis, with comparisons between colonized and uncolonized electrodes. [6]
Their recordings changed with fungal growth and biocide treatment, supporting a biological origin. Analysing how frequencies change over time helped separate patterns from background interference. This was another experimental system, rather than a replication of the mobile robot. Its relevance is methodological: reliable biohybrid control begins with knowing what the electrodes are actually detecting. [6]
What the robot study leaves unresolved
The robot paper reports signal weakening, difficulty resolving small signals below 50 microvolts, and lag associated with a 30-second sampling interval. Its authors also call for further validation to separate ultraviolet responses from potential heat effects. These limitations qualify the demonstration. [1]
A useful follow-up, in my assessment, would compare several independently grown cultures under the same conditions. Does the same stimulus repeatedly produce an identifiable response? Does the relationship persist as a culture ages? Can temperature, moisture and mechanical movement be changed separately? These are proposed tests, rather than results claimed for the prototype.
Another practical question is specificity. A sensor that changes when something happens may still be difficult to interpret. A robot would need a dependable way to distinguish a target condition from unrelated changes before acting on the measurement. Reliability should be judged against the intended job, including the consequences of a mistaken response.
Where the idea could become useful
The researchers proposed future chemical sensing, including agricultural applications. Their demonstration did not establish a field-ready system that diagnoses soil conditions or decides fertilizer doses. [2]
For a future development programme, I would compare a fungal module with conventional sensors doing the same task. Useful criteria would include response time, accuracy, maintenance, replacement frequency and total operating demands. A living component would earn its place by offering a measurable advantage, rather than by being biologically unusual.
The memorable lesson is an engineering possibility: biological activity can become an input to machinery. The next step is to show when that input remains interpretable, dependable and worth maintaining. Watching a prototype move makes the idea tangible; carefully designed comparisons will determine whether it becomes a practical tool.
Sources / References
- Mishra, A. K., Kim, J., Baghdadi, H., Johnson, B. R., Hodge, K. T., and Shepherd, R. F. Sensorimotor control of robots mediated by electrophysiological measurements of fungal mycelia. Science Robotics 9(93), eadk8019, 2024 (28 August 2024).
- Nutt, D. Biohybrid robots controlled by electrical impulses — in mushrooms. Cornell Chronicle, 2024 (28 August 2024).
- Clark, M. A., Douglas, M., and Choi, J. Biology 2e, Section 24.1: Characteristics of Fungi. OpenStax, 2018.
- Adamatzky, A. On spiking behaviour of oyster fungi Pleurotus djamor. Scientific Reports 8, 7873, 2018 (18 May 2018).
- Browner, D., and Adamatzky, A. Micro-electrode array recording of extracellular electrical potentials of liquid static surface fermented Hericium erinaceus. BioSystems 245, 105298, 2024 (Online 17 August 2024).
- Buffi, M., et al. Detection of electrical signals in fungal mycelia in response to external stimuli. iScience 28(10), 113484, 2025 (Published online 2 September 2025; issue 17 October 2025).
0 Comments