Mushroom Memristors

Mushroom memristors

Researchers from The Ohio State University found that common edible mushrooms can be grown and trained to act as organic memristors.

The team cultured samples of shiitake and button mushrooms, dehydrated them once mature to ensure long-term viability, connected them to special electronic circuits, and then electrocuted them at various voltages and frequencies.

“Mycelium as a computing substrate has been explored before in less intuitive setups, but our work tries to push one of these memristive systems to its limits,” said John LaRocco, a research scientist in psychiatry at Ohio State’s College of Medicine, in a press release. “We would connect electrical wires and probes at different points on the mushrooms because distinct parts of it have different electrical properties. Depending on the voltage and connectivity, we were seeing different performances.”

The volatile memory circuit was implemented using fungal memristors. (Credit: John LaRocco / The Ohio State University)

When used as RAM, the mushroom memristor could switch between electrical states at frequencies up to 5.85 kHz with about 90% accuracy. While performance dropped as the frequency of the electrical voltages increased, it could be fixed by connecting more mushrooms to the circuit.

Shiitake’s radiation resistance suggests possibilities for aerospace applications, the researchers said, but noted that viable fungal memristors would need to be far smaller than what they achieved, with future work aimed at miniaturizing the devices and improving cultivation techniques. [1]

Small hybrid LDO

Researchers from Ulsan National Institute of Science and Technology (UNIST) built an ultra-small hybrid low-dropout regulator (LDO) that uses a new digital-to-analog transfer (D2A-TF) method and a local ground generator to stabilize voltage and filter noise.

“Traditional hybrid LDOs often require large capacitors to smooth out digital-to-analog transitions, which can be a bottleneck. Our new design solves this problem with a seamless digital-analog transfer technique, making it both smaller and more efficient,” said Changmin An of UNIST in a statement.

When made with a 28nm process, the LDO measured 0.032 mm². In tests, it kept voltage ripple to 54 millivolts during rapid 99 mA current swings and restored the voltage to its proper level in 667 nanoseconds. It also achieved a power supply rejection ratio of –53.7 dB at 10 kHz with a 100 mA load, filtering out nearly all noise at that frequency. The LDO is designed to activate only during sudden power surges and consumes very little standby power. The team sees potential for use in next-generation AI chips and 6G communication modules. [2]

Photonic convolutions

Researchers from the University of Florida, University of California Los Angeles, and George Washington University developed a lens-based photonic chip capable of performing convolutions with high efficiency for image recognition and similar pattern-finding tasks.

The prototype chip uses two sets of miniature Fresnel lenses using standard manufacturing processes. Data is converted into laser light on-chip and passed through the lenses. The results are then converted back into a digital signal to complete the AI task. In tests, it achieved about 98% accuracy in classifying handwritten digits.

The chip can use different colored lasers to process multiple data streams in parallel, noted Hangbo Yang, a research associate professor at UF, in a press release. “We can have multiple wavelengths, or colors, of light passing through the lens at the same time. That’s a key advantage of photonics.” [3]

References

[1] J. LaRocco, Q. Tahmina, R. Petreaca, et al. Sustainable memristors from shiitake mycelium for high-frequency bioelectronics. (2025) PLoS One 20(10): e0328965. https://doi.org/10.1371/journal.pone.0328965

[2] C. An, H. An, H. Nam, et al. A −53.7-dB PSRR, Fast-Transient Output-Capacitor-Less Digital-Assisted Analog LDO Using Seamless Digital-to-Analog Transfer Technique. IEEE Journal of Solid-State Circuits. https://dx.doi.org/10.1109/jssc.2025.3602461

[3] H. Yang, N. Peserico, S. Li, et al. Near-energy-free photonic Fourier transformation for convolution operation acceleration. Advanced Photonics, 2025; 7 (05) http://dx.doi.org/10.1117/1.AP.7.5.056007