The human brain is genetically mosaic. Pathogenic somatic mutations occurring during brain development can cause brain malformations in children, with symptoms including cerebral palsy and drug-resistant epilepsy. Microscopically, brain malformations are characterised by disrupted cortical layering and the presence of abnormal cell types including “dysmorphic neurons”. Patients often require surgical removal of the affected cortex for seizure control. The genetics underlying brain malformations and the biology of dysmorphic neurons remain incompletely understood. Here we investigate brain malformations at single-cell resolution using patient-derived surgical brain tissue. Genetic analysis using targeted panel deep sequencing (>500X depth) identified low allele frequency (1.5%~5.7%), brain-specific somatic variants in 10 of 30 cases. All variants were identified in genes involved in the mTOR signalling pathway, including MTOR, RHEB, TSC1, TSC2, and DEPDC5. These variants were predicted to result in hyperactivation of the pathway, which was subsequently confirmed by immunohistochemical analyses of brain sections. Further characterisation of brain tissues revealed a “mutation gradient”, in which the highest mutation load was observed in brain region with the strongest epileptic discharge and the most severe histopathology. We used laser capture microdissection to isolate dysmorphic neurons and showed that pathogenic somatic variant can be identified in dysmorphic neurons but not in normal neurons. We performed single-nucleus RNA-sequencing to analyse the transcriptomic signature of dysplastic (n=8) and normal brain specimens (n=4) at single-cell resolution. We captured 47,706 nuclei, corresponding to 25 cell clusters representing different brain cell types. We identified a unique cluster enriched in the dysplastic brain specimens, which likely represents the dysmorphic neurons. Detailed transcriptomic profiling of this cluster may provide novel genes to delineate the biology of dysmorphic neurons. Our results offer novel insights into the genetics and transcriptomics of brain malformations, which will be applicable to the broader research in epilepsy and brain development.