The analysis, drawn from the Home Air Filtration for Traffic-Related Air Pollution (HAFTRAP) trial, showed that participants aged 40 or above who used HEPA filtration completed the Trail Making Test – Part B 12% faster than those who used sham units during the preceding month (54.0 seconds versus 61.4 seconds, p=0.02). No significant cognitive benefit was observed among participants under 40.
The randomized crossover trial, conducted in the United States, enrolled 119 healthy adults living full-time within 200 metres of highways in Somerville, Massachusetts. Each participant received both HEPA and sham filtration for separate one-month periods, with a one-month washout period in between.
While the study found no significant overall difference in test completion times across all age groups, age was a significant moderator of the effect (test for interaction p=0.02). The improvement was specific to Part B of the test, which assesses executive function and mental flexibility—cognitive domains linked to the frontal and temporal lobes of the brain.
In a subset of 19 homes with air monitoring, HEPA filtration reduced indoor PM2.5 concentrations by 52% and ultrafine particle counts by 32% compared with sham filtration. The researchers noted that their findings align with neurobiological evidence showing that particulate matter exposure may reduce white matter in brain regions responsible for these cognitive skills.
The study authors wrote that "among older healthy adults, there was an improvement in cognition following one month of in-home HEPA filtration."
The trial was conducted between September and April across the 2020–2024 period, with enrolment limited to cooler months when highway-related ultrafine particle concentrations tend to be higher. The research received approval from the UConn Health Institutional Review Board and was registered on ClinicalTrials.gov (NCT04279249).
The researchers acknowledged several limitations, including a predominantly White, higher-income, and well-educated participant sample, which may limit generalisability. They also noted that the Trail Making Test has a known learning effect, as participants often improve simply by retaking the test. Additionally, because participants did not have pre-existing cognitive impairments, the findings are not directly applicable to cognitively impaired individuals.
