Objective: Our knowledge of disease mechanisms in epilepsy is biased by findings originating from cross-sectional studies and advanced stages of epilepsy. We provide a new perspective by collecting systematically longitudinal data from patients who present in early stages (ES). Methods: The Halifax First Seizure Clinic, founded in 2008, uses a comprehensive multimodal data basis addressing clinical presentation, neuroimaging, EEG findings, genetics, cognition, comorbidities, social parameters, and life style. Follow-up visits are 6, 12 and 24 months. Results: Out of 575 patients we identified 3 subgroups 1) Strictly first seizure, n=187, 2) New-onset epilepsy (> 1 seizure < 12 months), n=149, and 3) Newly-diagnosed epilepsy (seizures > > 12 months), n=50. 189 patients were excluded (not proven seizure or other conditions e.g syncope etc.). Our interim analyses suggest: A) pharmacoresistance presents in highly diverse patterns and is rarer than expected in ES, B) Congenital malformations seem to have an excellent treatment prognosis in ES, C) Marijuana consumption is significantly more prevalent at initial assessment (comparison general population), D) Preceding psychiatric comorbidities associated with reduced amygdalar volume may be a predictor for seizure recurrence. Conclusions: The preliminary and illustrative findings of our pilot study challenge current concepts of disease evolution in epilepsy.