In 2015 Min-sun Lee had 24 young men repot a Peperomia and work on a word processor, 15 minutes each, in a randomised crossover. Diastolic blood pressure was 65.26 after the plant and 71.75 after the computer, a difference of 6.5 mmHg at p = 0.001. Lee’s own caveat belongs with the figure: the control was mental work, which raises sympathetic activity by itself, and all 24 subjects were healthy male students.
That is the shape of this field. The firmest numbers come from doing something with plants. Having them in the room shows up more in direction than in size.
Ke-Tsung Han gathered 42 records in 2022 and could pool only 16 of them. Across those, indoor plants moved diastolic blood pressure by −2.526 (95 per cent CI −4.142 to −0.909) and academic achievement by 0.534. Effects on EEG alpha and beta waves, on attention and on response time ran in the expected direction without reaching significance. Han names the limit plainly: two or three records per function.
Hilary Serra put 41 people through a working memory task in 2025, once with a removable living wall in the room and once with a shelf of objects standing in the same place. Heart rate variability showed more relaxation with the wall, and it showed it in a real room and in a virtual copy of that room alike. Positive emotions rose only in the real one.
Qinghua Lei measured the dose in 2021. Fifteen participants worked at green coverage ratios of 0, 0.2, 5, 12 and 20 per cent of the visible surface. Output improved at every ratio above zero, but 0.2 and 5 per cent changed nothing psychologically, and only 12 and 20 per cent moved brain activity. Lei settled on 12 per cent as the working optimum and called 20 per cent overwhelming.
Marlon Nieuwenhuis ran three field experiments in Dutch and British offices in 2014, two of them over weeks and months, and reported better outcomes with plants in all three. Andrew Thatcher reproduced the laboratory result in South Africa in 2020, then took it into two call centres for six weeks and for fourteen, and could not reproduce it on any measure there. Both stand on this sheet because neither has been answered.
Neither paper is open to us in full text, so the entries below carry no figure. Lee measured 24 people, Serra 41 and Lei 15.