⚡ Your Crops Are Already Reporting Stress — Days Before You Can See It
In a commercial-style greenhouse in Switzerland, tomato plants signaled iron deficiency about six days before their young leaves began to pale. No tissue test, no drone flight, no scout walking the rows — just an electrode on the stem and a machine learning model reading the plant's own electrical activity.
That result is one of a growing body of studies pulled together in a new research synthesis from the Electric Plant Company, a startup building technology to decode plant electrical signals. The review, Plant Perception and Electrical Communication by Lee von Kraus, PhD, makes a case that agronomists should take seriously: plants are continuous, high-resolution environmental sensors, and we are only beginning to learn how to listen to them.
von Kraus, L. (2026). Plant Perception and Electrical Communication. Electric Plant Company. https://www.electricplant.co/research/plant-perception
Here is what the science actually shows, what it could mean for producers, and where the evidence still has a long way to go.
🌿 Plants Have a Wiring System — It Just Runs Through the Phloem
Animals move information through nerves. Plants do it through the phloem — specifically, the chain of sieve element cells that runs the length of the plant from root tip to leaf.
The mechanism is strikingly similar. In both kingdoms, a stimulus triggers an action potential: a self-propagating wave of ion channels opening and closing in sequence down the length of the cell chain. Animal neurons rely primarily on sodium, potassium, and calcium ions. Plant sieve elements use potassium, chloride, and calcium. The main practical difference is speed — plant action potentials last considerably longer than their animal counterparts.
Plants also produce a broader vocabulary of electrical signals than animals do. Alongside action potentials, they generate variation potentials (slow waves driven by hydraulic pressure changes in the xylem), system potentials, and localized wound potentials. These signals can overlap and interact when a plant faces more than one stress at once — a detail that matters enormously in a real field, where stresses rarely arrive one at a time.
The common thread connecting nearly every stimulus to an electrical response is calcium. Whether the trigger is wounding, drought, cold, or a pathogen, the signal is typically translated into a rapid shift in cellular calcium that then drives membrane voltage changes — and those changes can travel long distances through the plant within minutes.
Choi, W.-G., Hilleary, R., Swanson, S.J., Kim, S.-H., & Gilroy, S. (2016). Rapid, long-distance electrical and calcium signaling in plants. Annual Review of Plant Biology, 67, 287–307. https://doi.org/10.1146/annurev-arplant-043015-112130
🔬 What Plant Electrical Signals Are Known to Encode
Decades of plant electrophysiology research have linked distinct electrical responses to a long list of stimuli: wounding, touch, temperature, light wavelength, soil moisture, herbivorous insects, pathogens, pollination, and chemical exposure in air, soil, and water. That information moves between roots and shoots to regulate photosynthesis, respiration, vascular transport, and gene expression.
For most of that history, though, the research had a serious practical limitation. Recordings were made inside electromagnetically shielded laboratory enclosures (Faraday cages) because ordinary environmental electrical noise swamped the plant's faint signals. That made the work scientifically interesting but agronomically irrelevant.
Machine learning has changed that.
🤖 The Breakthrough: Reading Plant Signals Outside the Lab
The turning point came when researchers began recording plants in realistic growing environments and letting machine learning models separate meaningful signal from background noise.
Water stress. A Swiss team recorded hydroponic tomatoes in a commercial-style greenhouse — no Faraday cage — and trained a gradient boosted tree model that distinguished water-deficit plants from well-watered plants with 98.5% accuracy. The electrical shift appeared early in the irrigation reduction, before visible wilting.
Tran, D., Dutoit, F., Najdenovska, E., Wallbridge, N., Plummer, C., Mazza, M., Raileanu, L.E., & Camps, C. (2019). Electrophysiological assessment of plant status outside a Faraday cage using supervised machine learning. Scientific Reports, 9, 17073. https://doi.org/10.1038/s41598-019-53675-4
Iron deficiency. In the study opening this article, the model flagged iron deprivation at approximately day 2, while the first visible symptom — pale young leaves — did not appear until about day 8. Accuracy was more modest (roughly 76–79%), and predictions fluctuated during the early window, but the lead time is the headline.
Tran, D. & Camps, C. (2021). Early diagnosis of iron deficiency in commercial tomato crop using electrical signals. Frontiers in Sustainable Food Systems, 5, 631529. https://doi.org/10.3389/fsufs.2021.631529
Nitrogen deficit. A deep learning model applied to stem electrical recordings identified nitrogen-deficit stress with about 88% accuracy on individual samples, rising to roughly 96% when ten consecutive predictions were combined. In some plants, stress was detected one to two days before visual symptoms appeared.
González i Juclà, D., Najdenovska, E., Dutoit, F., & Raileanu, L.E. (2023). Detecting stress caused by nitrogen deficit using deep learning techniques applied on plant electrophysiological data. Scientific Reports, 13, 9633. https://doi.org/10.1038/s41598-023-36683-3
Disease and pests. Barley plants have shown systemic electrical changes that reveal pathogenic infection days before the first disease symptoms. For spider mites in tomato, a model reached about 80% accuracy in identifying infested plants — though notably, that study classified plants that were already heavily infested, and the authors list pre-symptomatic detection as future work.
Zaiosc Simmi, F., Dallagnol, L.J., Almeida, R.O., Dorneles, K.R., & Souza, G.M. (2023). Barley systemic bioelectrical changes detect pathogenic infection days before the first disease symptoms. Computers and Electronics in Agriculture, 209, 107832. https://doi.org/10.1016/j.compag.2023.107832
Najdenovska, E., Dutoit, F., Tran, D., Plummer, C., Wallbridge, N., Camps, C., & Raileanu, L.E. (2021). Classification of plant electrophysiology signals for detection of spider mites infestation in tomatoes. Applied Sciences, 11(4), 1414. https://doi.org/10.3390/app11041414
Across the broader literature, the Electric Plant review reports that newer multi-class models — those asked to tell apart several different stresses, not just stressed versus healthy — have achieved accuracies ranging from 73% to above 99%.
🌐 The "Internet of Plants" — Promising, but Read the Fine Print
The review also explores communication between plants, and here a scientist's caution is warranted.
Some of the evidence is solid. Researchers demonstrated that an injured plant can pass electrical signals directly to a neighbor through touching leaves under humid conditions, triggering stress-acclimation responses in the receiving plant. Connecting two plants with a copper wire reproduced most of the same effects — strong evidence that the signal is electrical.
Szechyńska-Hebda, M., Lewandowska, M., Witoń, D., Fichman, Y., Mittler, R., & Karpiński, S.M. (2022). Aboveground plant-to-plant electrical signaling mediates network acquired acclimation. The Plant Cell, 34, 3047–3065. https://doi.org/10.1093/plcell/koac150
Underground signaling through fungal (mycorrhizal) networks is a different story. It has become a popular narrative, but a 2023 analysis of the field literature concluded that peer-reviewed field evidence for trees sending warning signals through these networks is lacking, and that the narrative has been amplified by citation bias and over-interpretation of limited greenhouse results.
Karst, J., Jones, M.D., & Hoeksema, J.D. (2023). Positive citation bias and overinterpreted results lead to misinformation on common mycorrhizal networks in forests. Nature Ecology & Evolution. https://doi.org/10.1038/s41559-023-01986-1
Mycorrhizal networks unquestionably matter for nutrient and water acquisition — that is foundational soil biology. Whether they function as a reliable plant-to-plant communication channel in the field remains an open question.
🚜 What This Means for Your Operation
1. The plant becomes the sensor. Soil moisture probes, weather stations, and tissue tests measure inputs and conditions. Plant electrophysiology measures the plant's response — how the crop itself is experiencing its environment. These are complementary data streams, not competing ones. A soil probe tells you what water is available; a plant signal could tell you whether the crop is actually stressed by its absence.
2. The early-detection window is where the money is. Most nutrient and disease problems are diagnosed only after visible symptoms appear — by which point yield potential is often already compromised. A two-to-six-day head start on iron deficiency, nitrogen stress, or fungal infection could shift management from reactive rescue to preventive correction: a targeted fertility adjustment, an earlier scouting trip, or a better-timed fungicide application.
3. Plant pathology is a natural first application. My own research career has centered on how plants defend themselves at the leaf surface — antimicrobial proteins like phylloplanins that form a first line of defense against pathogens. Electrical signaling is how the alarm spreads from that first point of contact to the rest of the plant. Being able to "hear" that alarm in real time would be a genuine advance in disease management.
4. A reality check before anyone buys sensors. Producers should understand where this evidence currently stands:
Most of the machine learning studies were conducted on hydroponic greenhouse tomatoes, often with 12 to 16 plants per experiment.
Much of the foundational applied work comes from a single research consortium in Switzerland. Independent replication across groups is still limited.
Open-field row crops — corn, soybeans, wheat, tobacco, hemp — face wind, rain, temperature swings, and electrode attachment challenges that greenhouse studies have not yet solved at scale.
Several authors explicitly describe their results as preliminary and call for larger datasets and testing across species.
The likely first commercial adopters are controlled-environment and high-value specialty crop operations — greenhouse vegetables, nursery production, and potentially hemp — where per-plant economics can justify the sensor investment.
🧭 The Bottom Line
Plants have spent hundreds of millions of years evolving sophisticated systems to sense their environment and communicate stress internally. The Electric Plant review is a valuable synthesis of how far the science of reading those signals has come — particularly in the last five years, as machine learning has moved recordings out of the shielded lab and into realistic growing conditions.
The core biology is well established. The early-detection results are genuinely exciting. The path from greenhouse tomato to Kentucky row crop is real but not yet paved. For producers and agtech investors alike, the right posture is informed optimism: watch this space closely, ask hard questions about validation data, and be ready to move when field-scale evidence arrives.
💼 Put the Science to Work on Your Operation
Higher Ground Plant Consulting LLC translates cutting-edge plant science into practical, field-ready decisions. Whether you are a producer evaluating new monitoring technology, an agtech company validating a sensing platform, or an investor conducting technical due diligence on an emerging agricultural technology — our team brings the scientific rigor to separate genuine breakthroughs from hype.
Led by Dr. Brian C. King (PhD Crop Science, MBA — University of Kentucky), Higher Ground has supported producers, agribusinesses, and research institutions with evidence-based agronomic and plant science consulting for over 20 years, with expertise spanning plant pathology, plant–microbe interactions, and agricultural biotechnology.
Our services include:
✅ Emerging agtech evaluation — sensor, precision ag, and AI-based monitoring platforms ✅ Independent technical due diligence for agtech investors and acquirers ✅ Plant disease diagnostics and integrated disease management strategy ✅ Crop nutrition and stress-monitoring program design ✅ Research trial design and validation studies for agtech companies ✅ Federal grant identification and writing (USDA SBIR, NIFA, SARE, NRCS)
If you are weighing whether a new plant-sensing technology is ready for your operation or your portfolio, the answer lives in the validation data — and we know how to read it.
📩 Schedule a consultation today.
🌐 highergroundplantconsulting.com 📧 brian@highergroundplantconsulting.com
Science-based. Field-proven. Kentucky-rooted.
Verified References
von Kraus, L. (2026). Plant Perception and Electrical Communication. Electric Plant Company. https://www.electricplant.co/research/plant-perception ✅ Primary source — read in full. Note: company research synthesis, not peer-reviewed.
Choi, W.-G., Hilleary, R., Swanson, S.J., Kim, S.-H., & Gilroy, S. (2016). Rapid, long-distance electrical and calcium signaling in plants. Annual Review of Plant Biology, 67, 287–307. DOI: 10.1146/annurev-arplant-043015-112130 ✅ Cited in source (ref. 39) — DOI confirmed at annualreviews.org
Tran, D., Dutoit, F., Najdenovska, E., Wallbridge, N., Plummer, C., Mazza, M., Raileanu, L.E., & Camps, C. (2019). Electrophysiological assessment of plant status outside a Faraday cage using supervised machine learning. Scientific Reports, 9, 17073. DOI: 10.1038/s41598-019-53675-4 ✅ Cited in source (ref. 42) — confirmed at Nature.com; 98.5% water-deficit accuracy confirmed in paper
Tran, D. & Camps, C. (2021). Early diagnosis of iron deficiency in commercial tomato crop using electrical signals. Frontiers in Sustainable Food Systems, 5, 631529. DOI: 10.3389/fsufs.2021.631529 ✅ Cited in source (ref. 46) — confirmed at Frontiers; D+2 detection vs. D+8 symptoms confirmed in paper
González i Juclà, D., Najdenovska, E., Dutoit, F., & Raileanu, L.E. (2023). Detecting stress caused by nitrogen deficit using deep learning techniques applied on plant electrophysiological data. Scientific Reports, 13, 9633. DOI: 10.1038/s41598-023-36683-3 ✅ Cited in source (ref. 47) — confirmed at Nature.com
Zaiosc Simmi, F., Dallagnol, L.J., Almeida, R.O., Dorneles, K.R., & Souza, G.M. (2023). Barley systemic bioelectrical changes detect pathogenic infection days before the first disease symptoms. Computers and Electronics in Agriculture, 209, 107832. DOI: 10.1016/j.compag.2023.107832 ✅ Cited in source (ref. 22) — title, journal, and volume confirmed on ScienceDirect (PII S016816992300220X) ⚠️ Full text was paywalled; DOI follows the article number and should be click-tested before publishing. Article claims limited to what the title states.
Najdenovska, E., Dutoit, F., Tran, D., Plummer, C., Wallbridge, N., Camps, C., & Raileanu, L.E. (2021). Classification of plant electrophysiology signals for detection of spider mites infestation in tomatoes. Applied Sciences, 11(4), 1414. DOI: 10.3390/app11041414 ✅ Cited in source (ref. 44) — confirmed at MDPI
Szechyńska-Hebda, M., Lewandowska, M., Witoń, D., Fichman, Y., Mittler, R., & Karpiński, S.M. (2022). Aboveground plant-to-plant electrical signaling mediates network acquired acclimation. The Plant Cell, 34, 3047–3065. DOI: 10.1093/plcell/koac150 ✅ Cited in source (ref. 57) — confirmed via Plantae / Polish Academy of Sciences listing
Karst, J., Jones, M.D., & Hoeksema, J.D. (2023). Positive citation bias and overinterpreted results lead to misinformation on common mycorrhizal networks in forests. Nature Ecology & Evolution. DOI: 10.1038/s41559-023-01986-1 ✅ DOI and authors confirmed from author-posted PDF (Univ. of Mississippi) ➕ Added independently — NOT in source reference list; included to provide scientific balance on the mycorrhizal-network communication claim.
Citation notes:
References 2–8 are drawn directly from the Electric Plant review's own reference list and each was verified individually against the publisher or an indexed record.
Reference 9 was added for balance and is disclosed as an outside citation.
Accuracy figures and detection lead times stated in the article were confirmed against the original papers, not taken from the review's summary.
⚠️ Correction vs. source framing: The review lists the spider mite study (ref. 7) as evidence of detection; the paper itself classified heavily infested plants and does not demonstrate pre-symptomatic detection. The article states this explicitly.
Article drafted for LinkedIn and website blog by Higher Ground Plant Consulting LLC | Dr. Brian C. King, PhD, MBA | CEO & Principal Investigator
#PlantElectrophysiology #PlantSignaling #EarlyDiseaseDetection #PlantPathology #NutrientDeficiency #CropMonitoring #PlantSensors #AIinAgriculture #MachineLearning #PrecisionAgriculture #AgTech #PlantScience #CropScience #Agronomy #Agriculture
