System for bidirectional, technically mediated communication with substrate-bound, non-vascular biological networks with electrical signaling activity using domain-specific artificial intelligence

Abstract

System for bidirectional, technologically mediated communication with substrate-bound, non-vascular biological networks, in particular fungal mycelial networks, mycorrhizal networks and plant tissue networks, characterized by an arrangement for in-situ communication within the natural or artificial substrate of the network and under real environmental conditions, comprising the combination of: a) at least one biosensor positioned in or on the substrate for detecting electrical, electrochemical and acoustic signals of the biological network, b) at least one frequency generator for targeted acoustic or electromagnetic stimulation of the biological network, c) a control unit comprising a domain-specific artificial intelligence (AI) configured as a closed, adaptively learning control loop.

Description

1. Field of the invention

[0001] The invention relates to a system for bidirectional, technologically mediated communication with substrate-bound, non-vascular biological networks exhibiting electrical signaling activity, in particular fungal mycelial networks, mycorrhizal networks, lichens, and plant tissue networks, using domain-specific artificial intelligence. The system operates in situ within the natural or artificial substrate of the biological network and under real environmental conditions. 2. Biological Foundations

[0002] Fungal mycelial networks are among the oldest and most widespread biological communication systems on Earth. They have existed for at least 400 million years and connect over 90 percent of all land plants in symbiotic information networks (mycorrhizae). The following properties relevant to the invention have been scientifically proven: Mycelial networks transmit electrical signals whose frequency and spiking patterns exhibit structural similarities to human languages ​​(Adamatzky et al., Royal Society Open Science, 2022).

[0003] Mycelial networks can reliably transmit electrical signals in the range of 100 Hz to 10,000 Hz and discriminate between frequencies (Propagation of electrical signals by fungi, ScienceDirect, 2023; Przyczyna et al., Biosystems, 2022).

[0004] Fungal cell walls possess mechanoreceptors that respond to acoustic vibration and frequency stimulation, altering growth and signaling activity (Robinson et al., Flinders University, 2024).

[0005] Mycorrhizal networks actively transmit warning, nutrient, and stress signals between connected plants and respond to environmental changes with measurable changes in their electrical signaling patterns.

[0006] Fungal tissues exhibit properties of memristors, capacitors and sensors (Fungal Electronics, Adamatzky, University of the West of England).

[0007] Crucially for the present invention, the electrical signal activity of these networks occurs without specialized nerve tissue and differs fundamentally from neuronal action potentials in signal characteristics, frequency spectrum, and generation mechanism. This necessitates a domain-specific artificial intelligence trained exclusively on signal patterns of fungal and plant networks. 3. State of the art and delimitation

[0008] The prior art known to us includes the following relevant systems, from which the present invention clearly distinguishes itself: Bidirectional AI-supported closed-loop systems for neural networks (brain-machine interfaces, neural prostheses): These are exclusively designed for neural tissues with specialized nerve tissue and are neither designed nor transferable to substrate-bound, non-vascular biological networks.

[0009] Adaptive bioelectronic wound therapy systems with AI-supported closed-loop (a-Heal, 2025): These are limited to human and animal tissue in a medical context.

[0010] Biosensors for detecting plant communication via mycorrhizal networks (Bletsas et al., Technical University of Crete): These detect signals unidirectionally without AI interpretation and without feedback stimulation.

[0011] AI-based image analysis for the classification of mycorrhizal colonization (AMFinder, Evangelisti et al., 2021): This analyzes static image data and does not communicate bidirectionally with the living network.

[0012] PEMF therapy devices: These stimulate biological tissue with predefined frequencies without recording and interpreting biological feedback signals and without an adaptive control loop.

[0013] None of the systems known to us form a closed, adaptively learning control loop between a technical system and a living, substrate-bound, non-vascular biological network—especially not a fungal or plant network—including domain-specific AI-supported real-time signal interpretation and adaptive frequency stimulation in situ and under real environmental conditions. This combination is, to the best of our knowledge, unprecedented. 4. Technical Impact and Advantages

[0014] The invention enables, for the first time, technically mediated bidirectional communication with living, substrate-bound, non-vascular biological networks. The key technical effects include: First-ever closure of an adaptive control loop between a technical system and a non-neuronal biological network in real time and under real environmental conditions - not in the laboratory, but in situ in the natural or artificial substrate.

[0015] Continuous ecosystem status monitoring without invasive interventions in the biological system.

[0016] Targeted manipulation of biological network activity to promote plant growth, soil regeneration and pest control.

[0017] Scalability from single-site research facilities to comprehensive ecosystem monitoring networks.

[0018] In a preferred embodiment, the system can be operated energy-autonomously by decentralized biological energy sources – in particular biogas plants, fuel cells based on organic substrates, or microbial fuel cells (MFCs) that extract energy directly from the biological substrate. This embodiment does not constitute a separate subject matter of protection, but significantly increases the practical applicability of the system, especially in remote agricultural and forestry areas without grid connection. 5. Application areas

[0019] The invention is particularly applicable in the following areas: Precision agriculture: Real-time monitoring of soil condition and plant stress - especially nutrient deficiency, drought stress and pollution - as well as targeted stimulation of nutrient transport without chemical fertilizers.

[0020] Ecosystem protection and forestry: Early warning systems for pest infestation, drought and fire hazard via extensive mycelial networks.

[0021] Soil regeneration: Accelerating the regeneration of damaged soils through targeted frequency stimulation of mycorrhizal networks.

[0022] Basic research: Deciphering the signaling language of substrate-bound, non-vascular biological networks and further developing domain-specific AI models.

[0023] Developing regions: Deployment in regions without health and agrochemical infrastructure through self-sufficient, decentralized system architecture in preferred embodiment.

[0024] Marine and aquatic ecology: Monitoring of aquatic fungal networks, lichen ecosystems and similar substrate-bound biological networks in aqueous substrate including fresh and salt water with organic support structures. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited non-patent literature

[0000] Adamatzky et al., Royal Society Open Science, 2022

[0002] Propagation of electrical signals by fungi, ScienceDirect, 2023; Przyczyna et al., Biosystems, 2022

[0003] Robinson et al., Flinders University, 2024

[0004] AMFinder, Evangelisti et al., 2021

[0011]

Claims

[1] System for bidirectional, technologically mediated communication with substrate-bound, non-vascular biological networks, especially fungal mycelial networks, mycorrhizal networks and plant tissue networks, characterized by an arrangement for in-situ communication within the natural or artificial substrate of the network and under real environmental conditions, comprising the combination of: a) at least one biosensor positioned in or on the substrate for detecting electrical, electrochemical and acoustic signals of the biological network, b) at least one frequency generator for targeted acoustic or electromagnetic stimulation of the biological network, c) a control unit comprising a domain-specific artificial intelligence (AI) configured as a closed, adaptively learning control loop. [2] System according to claim 1, characterized bythat the biological network is a mycorrhizal network that is present in the soil or a comparable terrestrial substrate. [3] System according to claim 1, characterized by that the biological network is a marine or aquatic fungal network, a lichen, or a fungus-like network such as slime molds (Myxomycota). [4] System according to any one of the preceding claims, characterized by , that the domain-specific artificial intelligence (c) is exclusively trained and optimized on signal patterns of fungal and plant networks to perform pattern recognition and interpretation of biological signals independent of neurobiological models. [5] System according to any one of the preceding claims, characterized by, that the domain-specific artificial intelligence exhibits continuous learning behavior according to feature c), through which it recognizes and classifies the signal patterns of the respective connected specific biological network over time and independently optimizes its stimulation parameters. [6] System according to any one of the preceding claims, characterized by that the artificial intelligence is designed to derive information about the state of the connected ecosystem from the interpreted signals, in particular nutrient deficiency, drought stress, pest infestation, fire hazard, environmental changes or pollution, and to initiate targeted countermeasures by frequency stimulation of the network. [7] System according to any one of the preceding claims, characterized by, that the biosensor is set up to detect geometrically regular growth and propagation patterns, and the artificial intelligence is configured to analyze these and feed them back into the network as frequency sequences. [8] System according to any one of the preceding claims, characterized by , that the system is configured in software to maintain a closed adaptive learning control loop in such a way that it continuously analyzes the signals detected by the biosensor (a) using the artificial intelligence (c) trained exclusively on fungal and plant patterns, generates stimulation signals for the frequency generator (b) based on this analysis, and processes the subsequent network response for feedback adjustment of the stimulation parameters. [9] System according to any one of the preceding claims, characterized by, that the domain-specific artificial intelligence according to feature c) is trained and validated exclusively with signal patterns of fungal, mycorrhizal and plant biological networks, and that the system is designed in terms of construction and software exclusively for use in substrate-bound, non-vascular biological networks, thereby structurally excluding a functional transfer to systems with specialized nerve tissue through the system architecture.