BACKGROUND: For the majority of congenital brain malformations, the underlying cause remains unknown. Recent studies have implicated rare copy number variations (CNVs) in their etiology. METHODS: Here, we used array-based molecular karyotyping to search for causative CNVs in 33 fetuses of terminated pregnancies with prenatally detected brain malformations and additional extracerebral anomalies. RESULTS: In 11 fetuses, we identified 15 CNVs (0.08 Mb to 29.59 Mb), comprising four duplications and eleven deletions. All larger CNVs (> 5 Mb) had also been detected by prenatal conventional karyotyping. None of these CNVs was present in our 1307 healthy in-house controls (frequency < 0.0008). Among these CNVs, we prioritized six chromosomal regions (1q25.1, 5q35.1, 6q25.3-qter, 11p14.3, 15q11.2-q13.1, 18q21.1) due to their previous association with human brain malformations or owing to the presence of a single gene expressed in human brain. Prioritized genes within these regions were UBTD2, SKA1, SVIP, and, most convincingly, GPR52. However, re-sequencing of GPR52 in 100 samples from fetuses with brain malformations or patients with intellectual disability and brain malformations revealed no disease-causing mutation. CONCLUSION: Our study suggests chromosomal regions 1q25.1, 5q35.1, 6q25.3-qter, 11p14.3, 15q11.2-q13.1, and 18q21.1 to be involved in human brain development. Within three of these regions, we suggest UBTD2, GPR52, and SKA1 as possible candidate genes. Because the overall detection rate of array-based molecular karyotyping was slightly higher (23%) than that of conventional prenatal karyotyping (20%), we suggest it's use for prenatal diagnostic testing in fetuses with nonisolated brain malformations.