We present the secondary-analysis of neurocognitive tests in the "Bumetanide in Autism Medication and Biomarker" (BAMBI;EUDRA-CT-2014-001560-35) study, a randomized double-blind placebo-controlled (1:1) trial testing 3-months bumetanide treatment ([less than or equal to] 1 mg twice-daily) in unmedicated children 7-15 years with ASD. Children with IQ [greater than or equal to] 70 were analyzed for baseline deficits and treatment-effects on the intention-to-treat-population with generalized-linear-models, principal component analysis and network analysis. Ninety-two children were allocated to treatment and 83 eligible for analyses. Heterogeneous neurocognitive impairments were found that were unaffected by bumetanide treatment. Network analysis showed higher modularity after treatment (mean difference:-0.165, 95%CI:-0.317 to - 0.013,p = 0.034) and changes in the relative importance of response inhibition in the neurocognitive network (mean difference:-0.037, 95%CI:-0.073 to - 0.001,p = 0.042). This study offers perspectives to include neurocognitive tests in ASD trials.