Project Summary According to the WHO, an estimated 5 million people worldwide receive a diagnosis of epilepsy annually. Unfortunately, misdiagnosis rates range between 20% to 42%. A false positive leads to inappropriate treatment with unnecessary antiseizure medication with potential adverse reactions, failure to receive suitable therapy for the correct diagnosis, and unnecessary restrictions that arise with the stigma of epilepsy. A false negative comes with increased risks of seizure recurrence, status epilepticus, and premature death. The diagnosis of epilepsy depends on a comprehensive clinical history, neurological examination, and ancillary studies including scalp electroencephalography (EEG). Scalp EEG can confirm an epilepsy diagnosis if abnormalities indicating epilepsy, such as random interictal (between seizure) epileptiform discharges (IEDs) or focal slow wave activity, are present and detected by visual inspection. However, the sensitivity of abnormalities being present in the EEG varies from 29-55%, and the ability for clinicians or EEG technicians to detect them by visual inspection varies. EEG artifacts can both mask IEDs and be mistaken for IEDs. Consequently, it takes multiple visits, months, or even years to be accurately diagnosed. We propose to further develop and validate EpiScalp, a revolutionary EEG analytics algorithm to enhance diagnostic accuracy. EpiScalp produces a risk score between 0-1 from 10-20 minutes of EEG data and does not rely on the presence of EEG abnormalities. Instead, our novel algorithm predicts epilepsy in resting-state (no seizure) brain networks using a dynamic network model. In Phase 1, EpiScalp underwent evaluation on 198 patients with EEGs void of abnormalities during their first visits. An alarming 54% (107) of these patients were misdiagnosed as having epilepsy when they did not. EpiScalp achieved definitive diagnoses for 168 of 198 patients with low (epilepsy unlikely) and very (epilepsy likely) risk scores, demonstrating remarkable accuracy, sensitivity, and specificity at 93%, 92%, and 95% respectively. EpiScalp could have reduced misdiagnoses from 54% to 17% (a 69% reduction). EpiScalp's predictive capabilities extend to patients undergoing long-term EEG monitoring within the epilepsy monitoring unit (EMU). Notably, 30-50% of EMU beds are occupied by non-epileptic patients, resulting in care delays for those requiring admission due to seizure exacerbation or medically refractory epilepsy needing surgical evaluation. This issue significantly impacts center efficiency and care quality. Through risk-based patient triage, EpiScalp improves care quality for epilepsy patients in need and empowers centers to efficiently assess more potential surgical candidates. This optimization of EMU resources enables timelier treatment for patients. In Phase 2, we aim to validate EpiScalp through prospective observational studies involving first visit and EMU patients at 3 renowned epilepsy centers in the US...