The current prevalence of cognitive impairment in HIV-infected individuals is surprisingly high, even in those with undetectable plasma HIV RNA. The aetiology is unknown but one possibility is inadequate control of persistent central nervous system (CNS) HIV infection. The CNS Penetration Effectiveness (CPE) rank has been proposed to predict how well an antiretroviral regimen treats CNS infection. Fabbiani et al. report that 'correcting' the CPE rank of each drug in an individual's regimen for the results of genotypic susceptibility (the CPE-GSS score) results in better ability to predict whether the regimen will improve cognition. The CPE-GSS score may help us better understand the aetiology of HIV-associated cognitive impairment. Whether it will be useful in the management of individual patients requires further study.