Yiming Shao ran 43 people through desk work in three rooms in 2024, one with a living wall, one with a fake living wall and one with a bare white wall, recording EEG and heart rate variability throughout. Negative psychological states built up about three times more slowly with the real wall than in the plain office. After resting in the living-wall room people came back less alert and less focused, and Shao’s recommendation follows from it: put the wall in the rest area, not beside the desks.
The rest of this literature mostly measures the building. Ibtihaj Alsadun ran two identical office rooms in Dhahran for twelve months in 2025, one of them retrofitted with a modular hydroponic green wall, with sensors on carbon dioxide, volatile organic compounds and PM2.5 and a sub-meter on the cooling. The green wall room used 13.5 per cent less cooling energy across the year, 574.5 kilowatt hours, rising to 17.1 per cent at the summer peak. Carbon dioxide fell 14.1 per cent during occupied hours, volatile organic compounds 28.1 per cent and PM2.5 20.9 per cent. The economic case in that paper is modelled from published relationships rather than measured, which belongs in a different column from the sensor readings.
Facade designers are asking plants to do the work of a blind. Luciana Kristanto shaded a test office in a tropical climate with Vernonia elliptica at two foliage densities in 2026. At a leaf area index of 1.5 the glass luminance dropped by up to 56.1 per cent while the daylight factor held at 6.2 per cent, and glare was a Daylight Glare Index of 21.67, which rates as perceptible rather than disturbing.
Fatemeh Delkhosh replaced the leaves with algae. Panels of spirulina at one part algae to 60 parts water left 85 per cent of the office space glare-free across the year, and holding the density reached after a single day of growth raised that to 95.5 per cent.
Victor Pozzobon modelled the same idea across four American climate zones using five years of weather data. In hot dry conditions a microalgae biofacade raised the share of time with comfortable indoor humidity from 19.7 to 62.5 per cent. In cold humid and hot humid zones it made no difference worth having, and the thermal gain over ordinary double glazing was 0.42 gigajoules per module per year.
Four of these five measure light, air, energy or humidity. Only Shao measured a person, and there were 43 of them.