Objective: Prior studies have shown improved neurocognition with initiation of antiretroviral treatment (ART) in HIV. We hypothesized that stopping ART would be associated with poorer neurocognitive function. Methods: Neurocognitive function was assessed as part of ACTG 5170, a multicenter, prospective observational study of HIV-infected subjects who elected to discontinue ART. Eligible subjects had CD4 count >350 cells/mm3, had HIV RNA viral load <55,000 cp/mL, and were on ART (≥2 drugs) for ≥6 months. Subjects stopped ART at study entry and were followed for 96 weeks with a neurocognitive examination. Results: A total of 167 subjects enrolled with a median nadir CD4 of 436 cells/mm3 and 4.5 median years on ART. Significant improvements in mean neuropsychological scores of 0.22, 0.39, 0.53, and 0.74 were found at weeks 24, 48, 72, and 96 (all p < 0.001). In the 46 subjects who restarted ART prior to week 96, no significant changes in neurocognitive function were observed. Conclusion: Subjects with preserved immune function found that neurocognition improved significantly following antiretroviral treatment (ART) discontinuation. The balance between the neurocognitive cost of untreated HIV viremia and the possible toxicities of ART require consideration. Classification of evidence: This study provides Class III evidence that discontinuing ART is associated with an improvement in 2 neuropsychological tests (Trail-Making Test A & B and the Wechsler Adult Intelligence Scale–Revised Digit Symbol subtest) for up to 96 weeks. Resuming ART was not associated with a decline in these scores for up to 45 weeks.