Scientists Construct Functional Memristor Using Shiitake Mushroom Mycelium

Too long to read? Go for the highlights below.

• Researchers at Ohio State University built working memristors, circuit elements that remember past electrical states—using shiitake mushroom mycelium instead of conventional silicon or titanium dioxide

• The fungal memristor achieved 5,850 Hertz performance with 90% accuracy, switching signals approximately every 170 microseconds, approaching commercially available memristor speeds

• The technology offers potential for low-cost, scalable, and biodegradable computing components, requiring resources as minimal as a compost heap and homemade electronics

---

The quest to develop computers that mimic brain function has led researchers down an unexpected path: using the underground networks of shiitake mushrooms as biological memory components. Scientists at Ohio State University have successfully constructed memristors from Lentinula edodes mycelium, demonstrating that fungal tissue can store and process information comparably to silicon-based hardware.

Mycelial Networks as Neural Analogues

Memristors represent circuit elements that retain memory of previous electrical states, functioning analogously to synapses—the junctions between neurons managing information flow. Developing brain-like computing requires components exhibiting similar behavioural characteristics. Mycelial networks have attracted scientific interest precisely because they behave similarly to neural networks: both possess comparable structural organisation and transmit information via electrical and chemical signals.

Microscopic mycelial network showing structural similarities to neural pathways used in neuromorphic computing models. Credits: John LaRocco et al.

However, transforming fungal networks into functional computing components requires deliberate engineering. The research team selected shiitake mushrooms based on the species' robustness, resilience, and resistance to environmental stressors including radiation. Nine samples were seeded in substrate-filled petri dishes with shiitake spores, cultivated under controlled temperature and humidity conditions.

Once mycelium sufficiently covered each petri dish, researchers dried samples in well-ventilated areas under direct sunlight to ensure long-term viability. This preparation rendered each sample ready for computational testing, connected to purpose-built circuits flooded with electrical currents.

Performance Metrics and Electrical Properties

According to psychiatrist John LaRocco, the team connected electrical wires and probes at different mycelial points, as distinct regions possess varying electrical properties. Depending on voltage and connectivity, the fungal tissue exhibited different performance characteristics.

The resulting mushristor achieved 5,850 Hertz performance with 90% accuracy, switching signals approximately 5,850 times per second, or one switch every 170 microseconds. The slowest commercially available memristors operate at slightly under twice that speed, rendering these initial results promising for early-stage research.

Schematic of the “mushristor” test circuit, showing electrical probes attached to shiitake mycelium to measure memory-like resistance changes. Credits: John LaRocco et al.

Testing revealed that increasing electrical voltage decreased mushroom performance. Researchers compensated for this limitation by incorporating additional mushrooms into circuits. This scalability suggests potential pathways for enhancing computational capacity through fungal network expansion.

Energy Efficiency and Material Advantages

The technology's appeal extends beyond novel biological substrates. LaRocco notes that developing microchips mimicking neural activity eliminates substantial power requirements for standby modes or when machines remain unused. This characteristic represents a significant potential computational and economic advantage over conventional silicon-based systems.

The approach offers material benefits unavailable with current hardware: potentially low-cost production, environmental sustainability, and biodegradability—properties conspicuously absent from many contemporary computer components.

Neuromorphic and bio-based computing approaches consume far less energy than conventional silicon architectures while offering biodegradable materials. Credits: John LaRocco et al.

LaRocco suggests that everything required to explore fungi and computing could range from compost heaps and homemade electronics to culturing factories with pre-made templates. All represent viable options using presently available resources.

Practical Applications and Development Timeline

Despite these promising results, mycelial computers will not power consumer devices imminently. The findings do, however, indicate an auspicious avenue for future research toward accessible, low-cost, and biodegradable components. Potential applications span personal devices to aerospace systems, though substantial development remains necessary before commercial viability.

The research addresses fundamental challenges in neuromorphic computing, systems designed to replicate brain structure and function. Traditional computing architectures consume substantial energy maintaining continuous electrical states. Brain-inspired systems promise dramatically reduced power consumption by processing information only when necessary, mimicking biological neural efficiency.

Fungal memristors contribute to this paradigm by offering biological substrates capable of storing electrical state information without continuous power input. The mycelium's natural electrical conductivity and ability to maintain distinct electrical properties across different regions provide inherent advantages for information processing and storage.

The work demonstrates that unconventional biological materials merit serious consideration for next-generation computing hardware. Whether fungal computing transitions from laboratory curiosity to practical technology depends on substantial improvements in switching speed, reliability, and integration with conventional electronic systems. Nonetheless, the successful construction of functional memristors from shiitake mycelium establishes proof of concept for biological computing components derived from readily cultivated fungal tissue.