Syndrome-specific experimental models are increasingly central to antiseizure drug development for developmental and epileptic encephalopathies (DEEs). Given genetic heterogeneity, early onset, complex seizure phenotypes, developmental impairment, and elevated SUDEP risk that characterize DEEs, the face and predictive validity of these models are critical determinants of translational success. As we emphasized in a recent review (Potschka and Pérez-Pérez, Epilepsia 2025;66:4597–4623; doi: 10.1111/epi.18581), careful appraisal of these validity domains is urgently needed.
Face validity in syndrome-specific DEE models refers to the extent to which they recapitulate key clinical features, including seizure types, age at onset, disease progression, neurodevelopmental non-seizure symptoms, and mortality. Genetic mouse and zebrafish models of Dravet syndrome represent the most extensively characterized examples, reproducing hyperthermia sensitivity, spontaneous seizures, interneuron dysfunction, and SUDEP risks. While this presentation focuses on anti-seizure effects, it emphasizes that these models enable assessment beyond seizure suppression by encompassing behavioral alterations, cognitive deficits, and motor dysfunction.
Predictive validity—defined by concordance between preclinical and clinical drug responses—remains more challenging to establish. In Dravet models, several clinically effective therapies, including fenfluramine and stiripentol (particularly in rational combinations), demonstrate antiseizure efficacy, supporting translational alignment. Moreover, contraindicated sodium channel blockers fail to improve or aggravate phenotypes in relevant models, further reinforcing predictive relevance. However, inconsistencies exist: some approved therapies show limited or variable effects in specific testing paradigms (e.g., hyperthermia-induced seizures), and age, genetic variant, dosing paradigm, and model-specific features influence responsiveness. For most DEEs beyond Dravet syndrome, systematic pharmacological validation remains sparse, limiting robust conclusions about predictive validity.
Additional challenges affect both validity domains. Incomplete penetrance of seizure phenotypes, reliance on provoked seizures in some models, species differences in neurodevelopment, limited heterogeneity relative to patient populations, and the labor-intensive nature of continuous EEG monitoring constrain the generalizability of findings. High mortality and ethical considerations further complicate the conduct of preclinical studies.
Overall, syndrome-specific DEE models can offer superior face validity and the potential for enhanced predictive validity compared with non-specific seizure models, particularly when evaluating both seizure and non-seizure outcomes. Nevertheless, comprehensive pharmacological characterization and validation are essential to strengthen confidence in their predictive power and to support successful mechanism-informed and disease-targeting antiseizure drug development for DEEs.