More than 90% of scalp Electroencephalogram (EEG) fails to identify the Hippocampal Epileptiform Activity (HEA). The objective was to develop and validate a machine-learning algorithm that accurately detects the HEA from a standard scalp EEG, without the need for intracranial electrodes. This was done by analyzing the intracranial electrodes in the brain. A diagnostic study was conducted from 2008 to 2021 with 141 eligible participants (97 with temporal lobe epilepsy [TLE] and 44 healthy controls [HCs] without epilepsy), and electroencephalogram (EEG) data were used from patients with TLE and HCs to train and validate a deep neural network. It was concluded that there are three ways to diagnose the same – novel, quantitative, and non-invasive. HEAnet computational algorithm may improve the diagnosis and treatment of epilepsy by detecting HEA from scalp EEG.