Personalised healthcare services are striving for improving the quality of individuals’ life by proposing the customization of medical decisions and products. Interpreting the biomedical signals, such as Electroencephalogram (EEG) and Electrocardiogram (ECG), falls into these types of services, and covers numerous applications from real-‐time monitoring to diagnosis of different health conditions. The conventional time series analyses lack the required robustness to address the need for interpretation of biomedical signals. In order to address this need, our ambition is focused in increasing the classification accuracy along with the minimum feedback from experts within a patient-‐specific framework. The main goal is to study and develop the state-‐of-‐the-‐art nonlinear methods in order to interpret the complex behaviour of the brain and heart using bioelectric signals and provide complementary information to characterize specific underlying patterns such as providing early warning for brain seizures or heart arrhythmias. More specifically, we aim to explore and develop robust and novel method in a sense of reconstruction and characterization of the signals’ dynamics. In order to achieve crucial improvements over the state-‐of-‐the-‐art techniques, the focus will be particularly drawn onto the Deep vs. “Divide and Conquer” Learning using Operational Neural Networks (ONNs).