Brainless decisions

30 November 2019 - By: Alex Evans

Brainless decisions

Human society is built on the sharing and processing of information, something that has long been the focus of advancement in the computer sciences - but mechanical computers aren’t the only computing-capable resource at our disposal, as even some of Earth’s simplest of organisms can be harnessed for their decision-making abilities. Andrew Adamatzky, head of the Unconventional Computing Lab at the University of the West of England, Bristol, UK, recounts his experiences working with computers of the soft and gooey kind.


As with many experimental scientists, Andrew’s interest in non-traditional computing first started when he became curious about new applications for his current research. “Years ago, my primary research interest was in computer codefinite state machines called cellular automata [1],” says Andrew. “It was while using cellular automata to solve geometry tasks that I realised that we can use similar algorithms to prototype chemical-based computers.” These chemical computers, also known as diffusion-reaction or gooware computers, rely on naturally occurring chemical reactions to emulate different types of mathematical calculations [2]. The applications of this technology are still actively being researched, with recent successes including improve route calculations for satellite navigation [3] and pattern recognition [4]. Andrew himself has even created a chemical pocket calculator! However, while his research with chemical computers proved them be capable of solving a wide variety of logical and mathematical tasks, Andrew noted that they were not able to physically grow in new directions or solve graph optimisation problems – issues that were soon dealt with by a new direction of research.


“I was sent a sample of the slime mould, Physarum polycephalum, by my peer Soichiro Tsuda, who suggested that I culture and observe it growing,” explains Andrew. “Over time, I realised that the configurations of the slime mould’s protoplasmic tubes could be used for a whole suite of computing applications.” One of Andrew’s first forays with his newly acquired slime moulds was to explore their decision-making processes when it comes to growth and how these processes could be applied to computing. “The plasmodium of Physarum is a single cell that relies on a protoplasmic network to sense its environment and optimise the transfer of nutrients to different parts of its body,” he explains. When Andrew presented his slime mould with oat flakes carved out into the shapes of Greater London and other major urban areas in the UK, he found that the growth of the mould did a surprisingly great job at replicating England’s major motorway system [5] and even demonstrated how it could be used to reconfigure transport networks in the event of a natural or man-made disaster. “I soon realised that the slime mould, in computational terms, can be seen as a reaction-diffusion system encapsulated in an elastic membrane,” adds Andrew. “After that, I got really curious in solving all manner of mathematical problems with slime mould and set about producing living prototypes of slime mould computing devices.”


And this research certainly isn’t just theoretical, as the PhyChip project [6] led by Andrew and the more recent PhySense project [7] were specifically focused on developing computing machines with living components. “In principle, it is possible to prepare a database of slime mould’s electrical responses to various stimuli and then use it as a biosensor,” says Andrew. “There’s no reason that this should differ for plants and fungi.” If you find yourself with slime mould and no idea of what to use it for, Andrew has published a list of “Thirty Eight Things to do with Live Slime Mould” [8] that includes many computational and problem-solving tasks, but also a few more creative outlets such as the creation of art and music!


“After my work with slime mould, I then started to look for other living substrates that could be used to implement computation,” says Andrew. After dabbling in the calculating applications of plant roots, Andrew and his colleagues concluded that there were probably better options to be working with. “Unfortunately, my experiments with plants weren’t as productive as the ones with slime mould,” says Andrew, “however, we were still able to produce a conceptual paper outlying possible developments for these green machines [9].” Instead of plants, it turned out that marvellous mushrooms proved to be the key to Andrew’s next line of enquiries. “First, I discovered that mushrooms are capable of generating interesting electrical potential spikes in response to different stimuli [10],” he explains. “It didn’t take long to realise that we can attempt to implement computing devices with mushrooms where information is encoded in these spikes of electrical activity [11]”. Although he has not yet produced any experimental prototypes of fungal computers, this is something that he is eagerly working towards as part of his new FUNGAR project [12]. “We are now considering the potential for fungi in the context of architecture, where buildings will be made from fungal materials, with some living parts to allow for sensing and computing,” adds Andrew. “This project has only just started so we expect to make plenty of exciting discoveries in the field of fungal computing.”


As well as being relatively easy to culture and experiment on in a controlled lab environment, these simple yet sophisticated organisms possess other qualities that make them ideal candidates for biological computation. “Firstly, these creatures lack a nervous system but are capable of producing interesting adaptive behaviours,” explains Andrew. “For example, when we compute with fungi, we can represent our input data as chemical, optical or electrical signals and then interpret the subsequent electrical activity as the results of the computation.” Throughout his career, Andrew has worked with all manner of different non-traditional computers, but he says that his discoveries about the puzzle-solving abilities of slime moulds and fungi are amongst his favourites: “I was especially pleased when I found that slime moulds can simulate transport networks, but discovering that we can record a spiking electrical activity of fungi by inserting electrodes in their fruiting bodies was also very exciting.” Looking forward to the next stages in this area of research, Andrew has his eyes set firmly on making these biological computers more intuitive and adaptable for incorporation into everyday use. “Although the exact experimental setups have differed between projects, a key goal of each has been to demonstrate that we can implement computation with living substrates,” concludes Andrew. “It turns out that we can create all kinds of fascinating computing, sensing and electronic devices with slime mould and fungi.”



[1] Adamatzky, Andrew I. Identification of cellular automata. CRC Press, 2014Taylor and Francis, 1994[..1] .

[2] Adamatzky, Andrew, Benjamin De Lacy Costello, and Tetsuya Asai. Reaction-diffusion computers. Elsevier, 2005.

[3] Suzuno, Kohta, et al. "Maze solving using fatty acid chemistry." Langmuir 30.31 (2014): 9251-9255.

[4] Cronin, Leroy, et al. "A programmable chemical computer with memory and pattern recognition." (2019).

[5] Adamatzky, Andrew, and Jeff Jones. "Road planning with slime mould: if Physarum built motorways it would route M6/M74 through Newcastle." International Journal of Bifurcation and Chaos 20.10 (2010): 3065-3084.



[8] Adamatzky, Andrew. "Thirty eight things to do with live slime mould." arXiv preprint:1512.08230 (2015).

[9] Adamatzky, Andrew, et al. "Computers from plants we never made: Speculations." Inspired by nature. Springer, Cham, 2018. 357-387.

[10] Adamatzky, Andrew. "On spiking behaviour of oyster fungi Pleurotus djamor." Scientific reports 8.1 (2018): 1-7.

[11] Adamatzky, Andrew. "Towards fungal computer." Interface focus 8.6 (2018): 20180029.


 [..1]First edition was published in 1994, evidence: