The effects of chronic exposure to xylene on cognitive ability were studied in a group of 35 medical workers occupationally exposed to low-level concentrations of xylene for at least five years by using event-related potentials (ERPs), and compared with a control group of 21 subjects. The exposure to xylene was confirmed through determination of m-methylhippuric acid, a reliable biological indicator of xylene exposure, in pre- and post-shift urine. A dose-effect relationship between log m-methylhippuric acid and ERP log latency (p = 0.032), and the ERP amplitude (p = 0.047) was statistically significant. The group of medical workers showed significantly longer ERP log latency (p < 0.001) than did the control group with respect to factors of exposure to smoking, education and age as covariates. For the ERP amplitude the difference was found not to be significant (p = 0.263), probably due to high between subject variability. The cognitive impairment may occur in workers chronically exposed to xylene.