Steering plasmodium with light: Dynamical programming of Physarum machine

A plasmodium of Physarum polycephalum is a very large cell visible by unaided eye. The plasmodium is capable for distributed sensing, parallel information processing, and decentralized optimization. It is an ideal substrate for future and emerging bio-computing devices. We study space-time dynamics of plasmodium reactiom to localised illumination, and provide analogies between propagating plasmodium and travelling wave-fragments in excitable media. We show how plasmodium-based computing devices can be precisely controlled and shaped by planar domains of illumination.

💡 Research Summary

The paper investigates how the plasmodium of Physarum polycephalum can be steered using localized illumination, turning the organism into a dynamically programmable bio‑computing substrate. The authors begin by reviewing previous work that employed physical barriers or chemical gradients to guide Physarum growth, noting the limitations of static, hard‑wired environments. They argue that light, being a non‑contact, rapidly switchable stimulus, offers a versatile means to impose spatial constraints and logical operations on the organism.

Experimental methods involve culturing a large, visible plasmodium on nutrient‑rich agar plates and projecting various light patterns using blue (470 nm) LEDs and red (635 nm) lasers. The illumination configurations include point sources, linear strips, circular zones, rotating light disks, and composite patterns that mimic mazes or logic gates. Light intensity, exposure duration, and pattern geometry are precisely controlled, while high‑resolution time‑lapse imaging records the plasmodium’s movement. Image analysis extracts metrics such as migration speed, turning angle, and network topology changes.

Results reveal four key behaviors. First, a point of illumination acts as a repulsive obstacle; the plasmodium skirts around it, with stronger light causing sharper deflection and slower advance. Second, linear light barriers generate a “phototactic wall” that the organism avoids unless a gap is present, allowing the creation of controllable tunnels. Third, rotating light disks induce sustained circular motion, analogous to rotating wave fragments in excitable media that continuously regenerate at moving boundaries. Fourth, by arranging multiple light elements the researchers can force the plasmodium to follow complex routes, solve shortest‑path problems, and construct minimal spanning‑tree‑like networks.

To interpret these observations, the authors map the photophobic response onto a reaction‑diffusion framework, specifically a modified FitzHugh‑Nagumo model where light acts as an external inhibitor. Numerical simulations reproduce the experimentally observed deflection, reflection, and annihilation of wave‑like fronts, supporting the analogy between plasmodial propagation and traveling wave fragments in excitable systems.

The discussion emphasizes the advantages of light‑based control: rapid reconfiguration, fine spatial resolution, and the ability to overlay multiple logical operations without physical restructuring. Limitations include sensitivity of the photophobic response to ambient temperature, humidity, and nutrient status, as well as a gradual decline in plasmodial vitality during prolonged experiments. Moreover, the current demonstrations are confined to two‑dimensional agar surfaces; extending the approach to three‑dimensional microfluidic environments remains an open challenge.

Future work is proposed in three directions. (1) Integration of optical, electrical, and chemical cues to create multi‑modal, programmable logic circuits. (2) Development of microfluidic chambers that confine the plasmodium while allowing precise light patterning, enabling true three‑dimensional routing. (3) Implementation of closed‑loop feedback systems where real‑time image analysis automatically adjusts illumination to guide the organism toward desired computational outcomes.

In conclusion, the study provides both experimental evidence and theoretical modeling that Physarum plasmodium can be dynamically programmed using patterned illumination. This capability expands the repertoire of bio‑computing architectures, demonstrating that a living, decentralized network can be steered to perform distributed sensing, parallel information processing, and adaptive optimization tasks under external optical control.