PubMed · 2026-06-28
Researchers developed a smarter algorithm that can untangle the overlapping electrical signals produced by different cell types inside a plant leaf, something older methods couldn't do reliably. This opens a cleaner window into how leaves sense and respond to light at the cellular level.
The second-order polynomial-mapped FastICA (Poly2) achieved Spearman correlation coefficients of 0.82 and 0.87 in simulations, outperforming all linear and higher-order comparison methods.
Real leaf surface recordings successfully separated signals attributed to guard cells and mesophyll cells, with correlation coefficients of 0.80 and 0.82 respectively against known reference curves.
Statistical testing at 95% confidence confirmed that plant leaf electrical signals have inherent nonlinear dynamics, explaining why standard linear signal-separation methods fail on them.