We find that the human PLA2G6 enzyme functions as an A2 phospholipase, hydrolyzing the sn-2 acyl chain of phosphatidylcholine, and as a lysophospholipase, hydrolyzing the sn-1 acyl chain of lysophosphatidylcholine, the product of its A2 phospholipase reaction. We find that mutations associated with INAD and NBIA profoundly impair enzyme function in both phospholipase and lysophospholipase assays. In contrast, mutations associated with dystonia-parkinsonism mutations do not impair catalytic function. Our results provide insight into pathogenic mechanisms underlying the Y-27632 spectrum of neurodegenerative diseases caused by PLA2G6 mutations. We find that mutations associated with NBIA/ INAD impair the catalytic activity of the PLA2G6 protein. In contrast, mutations associated with dystonia-parkinsonism do not impair catalytic activity. These results clarify the mechanisms underlying the phenotypic expression; selective effects on protein function, rather than other genetic or environmental factors, produce the two different disease spectrums, NBIA/INAD and dystonia-parkinsonism. Comparison of the effects of a glutamine versus a tryptophan substitution at the 741 position further illustrates the correlation between enzymatic activity and disease phenotype. A tryptophan substitution for arginine at position 741 produces an 80% reduction in activity, associated with INAD in one patient and NBIA in another patient, while glutamine substitution results in no apparent reduction in catalytic activity and is associated with a dystonia-parkinsonism phenotype. Although our studies do not detect a loss of catalytic function resulting from dystonia-parkinsonism mutations, the recessive pattern of inheritance suggests that they are more likely to cause a loss of function rather than a dominant gain of function. The R632W mutation has been observed in one patient with INAD in addition to three siblings with dystonia-parkinsonism. In the patient with INAD, a heterozygous R632W mutation was found in combination with a heterozygous V691del mutation, which is distinct from the situation in dystonia-parkinsonism patients who have all been homozygous for the R632W mutation. The complete loss of enzyme function caused by the V691del mutation suggests that a genotype-phenotype correlation may exist based on the degree of impairment in PLA2G6 function, and that impairment below a certain level may cause the early onset INAD/NBIA disease phenotype and more widespread effects in the nervous system. Although our assays did not detect impairment of catalytic activity, they do not exclude the possibility of impaired enzyme function in vivo. An alternative explanation is that an additional PLA2G6 mutation was present in this patient but was not detected by sequencing of PLA2G6 exons. Our results suggest that the R632W and R747W mutations might alter PLA2G6 function by increasing the catalytic rate for PC. We did not observe an increase in the catalytic rate for LPC. Further experiments are needed to determine whether these mutations alter the relative catalytic rates for PC and LPC.