What we are learning

The plant was always signalling. We finally had the instrumentation to hear it.

For most of agricultural history, plants have been managed from the outside: weather, soil, schedules, visual symptoms, and crop outcomes after the fact. Syntheflora is built on a different premise. If the plant's internal physiology can be measured continuously, then agriculture can respond to the biological system itself.

This page documents the research foundation behind that premise. The science matters because Syntheflora's AI agronomy agents are only useful if the plant signals they work from are real, measurable, and interpretable. Some findings are peer-reviewed. Some are technical application notes. Some are active observations still being written up. They are presented separately, because credibility depends on knowing which evidence belongs to which category.

How this evidence is presented.

Each finding on this page is labelled by evidence type. Credibility depends on knowing which category each result belongs to.

Evidence type What belongs here How it should be read
Peer-reviewed foundation Journal-published plant electrophysiology, biofeedback, and stimulus classification work Scientific proof base
Technical foundation CYBRES application notes and system documentation Methodology and instrument capability
Preprint / technical research Ozone detection and related biological sensing work Strong signal; verify publication status before calling peer-reviewed
Active research Tomato agency, root water redistribution, ongoing observations Promising observations, not final claims
Deployment outcomes Water, quality, energy, compliance ranges Commercial proof, with caveats by crop and region
Peer-reviewed foundation Biomimetics, 2024

When plants were given biofeedback control, they did not follow fixed schedules.

What was tested

CYBRES tested biofeedback-based phytoactuation in vertical farming and controlled-environment agriculture. The system measured plant physiological signals and used them to control light, irrigation, fertilization, and environmental parameters.

What was found

Biofeedback control produced measurable operational outcomes. Microgreen production cycles shortened from 7 days to 4–5 days. Wheatgrass production shortened from 10 days to 7–8 days. Pea biomass increased by 30% in combination with biofeedback-based irrigation. Energy optimization reached 25–30% compared with a non-optimized 16:8 lighting schedule.

Why it matters

This paper establishes the principle behind Syntheflora Cortex: plant physiology can become a control signal. The grower no longer has to assume that a fixed environmental schedule is optimal. The plant's response can guide AI agents for water, energy, climate, and micro-trial decisions.

Kernbach, S. "Biofeedback-Based Closed-Loop Phytoactuation in Vertical Farming and Controlled-Environment Agriculture." Biomimetics 2024, 9, 640. doi:10.3390/biomimetics9100640

Peer-reviewed foundation Bioinspiration & Biomimetics, 2023

Plant electrical signals are not noise. They are classifiable information.

What was tested

Researchers exposed Zamioculcas zamiifolia and tomato plants to controlled stimuli including wind, heat, red light, and blue light, then measured electrical potential and tissue impedance signals.

What was found

The study produced 1,864 electrical potential time series. Discriminant analysis classifiers achieved 100% accuracy for binary classification, 100% for three-class classification, and 99.1% accuracy for five-class classification. Statistical approaches outperformed deep-learning approaches in this dataset.

Why it matters

This is foundational for Syntheflora. If a plant's electrical activity reliably encodes different stimuli, then internal plant physiology can serve as the signal layer for Stress Detection, Pathogen Early Warning, Decision Support, and plant-aware automation.

Buss, E. et al. "Stimulus Classification with Electrical Potential and Impedance of Living Plants." Bioinspiration & Biomimetics 18 (2023) 025003. doi:10.1088/1748-3190/acbad2

Technical / preprint foundation arXiv, 2024

Plants can function as biological detectors of environmental stress.

What was tested

CYBRES measured electrochemical impedance in tobacco and tomato plants exposed to low ozone concentrations indoors and under outdoor conditions.

What was found

51 days of measurements across 948 sensor-plant attempts. Biological response delay of approximately 10–20 minutes after ozone exposure. 92% confidence in detecting elevated ozone when pooling data from at least three plants. Outdoor differentiation between high- and low-ozone days was demonstrated.

Why it matters

This extends the Syntheflora thesis beyond farm inputs. A plant's internal physiology can register environmental stress events that electronic instruments may not flag in the same way. This supports future government, ecological, and sentinel monitoring use cases.

Kernbach, S. "In-situ biological ozone detection by measuring electrochemical impedances of plant tissues." arXiv:2411.16321. Publication status to confirm before labelling peer-reviewed.

In progress

Research continues. These findings are not yet published.

The instrumentation does not stop generating data between publications. Two findings currently in active documentation are described below — presented not as claims, but as observations in the process of being written up.

In progress

When configured to let a tomato plant govern aspects of its environment, the system observed cycles that did not match 24-hour patterns.

When the Syntheflora biofeedback system was configured to allow a tomato plant to govern its own light cycles, nutrient timing, and irrigation through real-time physiological signals, it did not run the 24-hour cycles that controlled-environment agriculture assumes. Growth accelerated measurably.

This is a separate observation from the published microgreen and pea findings, with a productive commercial crop species producing a different pattern of self-directed behaviour.

Paper in preparation. Do not publish specific growth numbers until approved.
In progress

Monitoring of root-zone dynamics has documented conditions in which plant roots appear to emit water outward into the surrounding rhizosphere.

Standard irrigation models treat roots as absorbers — water moves from soil into root. Monitoring of root-zone dynamics documented plants actively emitting water from root tissue into the surrounding rhizosphere under specific conditions, appearing to hydrate dry zones around the root system.

If confirmed through further controlled experimentation, this has implications for deficit irrigation strategy, rhizosphere management, and the relationship between plants and the soil microbiome.

Preliminary observation. Active investigation.

The published and technical foundation.

All research listed below used the CYBRES phytosensing hardware — the instrument that Syntheflora brings to commercial deployment. The science is verifiable. Every paper is available in full.

  1. 01

    “Biofeedback-Based Closed-Loop Phytoactuation in Vertical Farming and Controlled-Environment Agriculture”

    Kernbach, S. — CYBRES GmbH, Stuttgart, Germany. Biomimetics 2024, 9, 640.

    Demonstrates closed-loop biofeedback control of grow light, irrigation, and nutrition. Documents adaptive photoperiodic rhythms, production cycle reduction, pea biomass increase, and energy optimisation results.

    ↗ doi:10.3390/biomimetics9100640
  2. 02

    “Stimulus Classification with Electrical Potential and Impedance of Living Plants”

    Buss, E., Weidner, E., Mohan, R., Burghardt, T., Kernbach, S. et al. Bioinspiration & Biomimetics 18 (2023) 025003.

    Classifies plant electrical responses to wind, heat, red light, and blue light using discriminant analysis and deep learning. Achieves 99.1% classification accuracy with discriminant analysis across five stimulus classes.

    ↗ doi:10.1088/1748-3190/acb3b0
  3. 03

    “In-Situ Biological Ozone Detection by Measuring Electrochemical Impedances of Plant Tissues”

    Kernbach, S. — CYBRES GmbH, Stuttgart, Germany. arXiv:2411.16321. Publication status to confirm.

    Demonstrates biological detection of low atmospheric ozone concentrations using electrochemical impedance measurements. 51 days of continuous automated experiments. 92% detection confidence from pooled plant data.

    ↗ arXiv:2411.16321
  4. 04

    “Application Note 28 — Using Phytosensors in Precision Agriculture, Vertical Farms, Hydroponics and Agricultural AI Applications”

    Kernbach, S. — CYBRES GmbH. CYBRES GmbH Application Note, v.0.6, July 2024.

    Technical reference document describing sensor deployment methodology, data channel definitions, embedded controller operation, feedback protocol design, and AI integration in agricultural applications. Use this as the technical citation for sensor capabilities and deployment methodology.

    ↗ Download from CYBRES
  5. 05

    “Human-AI Collaborative Evaluation of Plant Physiology with Multisensor Data”

    Kernbach, S. & Gemini AI Platform — CYBRES GmbH, Stuttgart, Germany. CYBRES GmbH, 2024.

    Documents the integration of the CYBRES phytosensing hardware with Google Gemini AI for real-time interpretation of multisensor plant physiological data. Establishes the analytical framework for the Syntheflora commercial platform.

For researchers

Build on the published foundation.

Syntheflora provides university and institutional research teams with access to the full phytosensing suite, raw data export, Python/API access, AI interpretation, and CYBRES technical support at near-cost pricing.

The purpose is not only instrument access. It is to expand the evidence base for plant-state intelligence across more crops, environments, and research questions.

Papers produced using the Syntheflora system are attributed to CYBRES GmbH as the instrument source, following standard scientific citation practice. There is no requirement to produce specific outcomes, publish within a specific timeframe, or share data before you choose to publish.

Apply for the Research Programme Ask a question first