“…Specifically, during substrate-guided mapping, bipolar electrograms (EGMs) recorded from damaged myocardial areas are carefully inspected by electrophysiologists, in order to identify those late and fractionated pathological deflections associated with arrhythmogenic areas. These pathological potentials, generally named abnormal ventricular potentials (AVPs), have been often associated with high-frequency components in the scientific literature [3]- [7], and more recently, their spectral contents have been deeply investigated [8], highlighting the presence of peculiar spectral signatures for AVPs. However, there is no evidence that such spectral signatures may be effective for the automatic recognition of AVPs and physiological EGMs, despite the usefulness of artificial intelligence tools for the recognition of arrhythmogenic zones in patients suffering from postischaemic VT has been already investigated [9], [10].…”