AMPA-type glutamate receptors (AMPARs) are tetrameric ligand-gated ion channels formed as different combinations of GluA1-4 subunits encoded by the genes GRIA1-4. Various pathogenic variants of these genes have been described in patients with developmental delay, intellectual disability, autistic spectrum disorder and seizures. Here we report a heterozygous de novo pathogenic missense mutation in GRIA2 (c.1928 C>T, p.A643V) identified in a one-year-old male patient with seizures, developmental delay and failure to thrive. Electrophysiological investigation of heterologously expressed receptors showed GluA2 A643V to exhibit greatly slowed deactivation, and a markedly reduced extent of desensitization, compared with wild-type GluA2. When GluA2 A643V was coexpressed with the transmembrane AMPAR regulatory protein (TARP) γ2, or with both GluA1 and γ2 to generate TARPed heteromeric receptors, the slowed deactivation and decreased desensitization persisted. We found that the AMPAR negative allosteric modulator perampanel was able to fully block currents from GluA2 A643V/γ2 receptors, albeit with reduced potency compared with wild-type GluA2/γ2. The introduction of perampanel to the patient’s treatment regimen, alongside a modified Atkins diet, was associated with a marked reduction in seizure number, a resolution of failure to thrive, and clear developmental gains. Our study suggests that AMPAR gain-of-function (GoF) underlies the effect of the GRIA2 variant in our patient, and that perampanel may be beneficial in other patients with GRIA GoF variants.Author summaryRapid communication between brain cells is mediated by the excitatory neurotransmitter glutamate. Mutations affecting genes that encode glutamate-gated ion channels can result in neurodevelopmental disorders and seizures. In a young male patient, we identified a mutation in the gene responsible for production of GluA2, a key subunit of AMPA-type glutamate receptors. We showed that receptors containing GluA2 produced by this GRIA2 A643V variant were more sensitive to glutamate and gave rise to unusually prolonged or sustained responses, thus causing ‘gain-of-function’. We found that receptors affected by this mutation could, nonetheless, be inhibited by the AMPA receptor-targeting antiseizure medication, perampanel. Prompted by this finding we introduced perampanel alongside the patient’s ongoing antiseizure medication. This coincided with a decrease in the patient’s seizures and clear developmental gains. Taken together, our findings suggest that a GRIA2 gain-of-function mutation can cause neurological disease, that perampanel can be used to directly counteract the effect of the mutation, and that it may be beneficial in other GRIA gain-of-function patients.