| Epileptic seizures correspond to episodes of increased rhythmicity of the normally chaotic activity in the biological neural networks (BNNs). It is believed that through suppression of harmful rhythmic modes, seizures may be eliminated.; In this thesis, we propose a hybrid neural network (HNN) that uses artificial neural networks (ANNs) to manipulate biological neural networks. The ANNs consist of two major networks: a detection network and a controller. The Detection network is responsible for the detection of state transitions that occur in BNNs. The controller will provide sufficiently small non-invasive electric field perturbations to the BNN. Utilizing the forces of the system vector field, it is able to restore the system state to its normal high complexity dynamic.; Using the mapped clock oscillators (MCO) model, we also demonstrated that detection and control of a frequency encoded dynamical system can be performed successfully using only its “raw” information. |
