All project evidence

PFA research program · PFI framework

Plant Physiological Intelligence

Estimate plant state from multiple observations and express uncertainty for a specific agricultural decision.

Jump to the test plan

Published research

Independent studies · Free-to-read full text

  1. A machine-learning–powered spectral-dominant multimodal soft wearable system for long-term and early-stage diagnosis of plant stresses

    Jiang et al. · Science Advances · 2025 · 10.1126/sciadv.adw7279

    The research team combined local leaf temperature, humidity and transmission spectra to characterize stress and nutrient deficiencies in tomatoes. This supports the measurement approach in a defined crop and leaves the PFI framework to be evaluated.

  2. PlantRing: A high-throughput wearable sensor system for decoding plant growth, water relations, and innovating irrigation

    Sun et al. · Plant Communications · 2025 · 10.1016/j.xplc.2025.101322

    The team used a wearable ring to measure stem dynamics in relation to growth and water status. This provides a comparator for a crop-specific irrigation study.

Technical feasibility

Plant-mounted sensors can collect complementary signals. A useful PFI model still requires a defined physiological outcome, independent reference measurements and performance on plants held out from model fitting.

Proposed first test

In one crop and a controlled irrigation experiment, compare predictions against an independent water-status reference. Split the evaluation by plant and run. Compare multimodal inference with the strongest single-sensor baseline.

Proposed pass criterion

Advance if the model improves prespecified prediction error on held-out plants and identifies unreliable predictions when inputs are missing. Validate transfer to new crops and sites separately.

These are research targets. This brief reports independent findings and a proposed test plan for SpectraCortex.