ABSTRACT Epilepsy is one of the most common neurological disorders affecting ~65 million people worldwide. Of those, 25-35% are not responsive to pharmacological treatment, despite the development of new antiepileptic drugs in the previous decades. One of the main barriers hindering the development of better therapies for epileptic seizures, such as closed-loop neuromodulation for seizure prevention and abatement, is the lack of understanding of how seizures initiate, spread and terminate over cortical and subcortical regions. Progress thus far has been hampered by the challenge of measuring in humans neural activity at the multiple scales of ensembles of single neurons and large-scale brain networks. In addition, most previous studies have focused on biophysical mechanisms for seizure initiation at seizure onset zones. An overlooked aspect of focal seizures is the formation/maintenance of local and large-scale pathological neuronal networks and the time-varying susceptibility of brain dynamics to seizure initiation and spread (generalization). We hypothesize that these pathological multiscale networks are maintained via the recurring activation of epileptiform spatiotemporal patterns not only during seizures but also during interictal and preictal periods. We will address these problems in patients with pharmacologically intractable focal epilepsy by recording ensemble of single-neurons via intracortical 96-microelectrode arrays (96-MEA, 4 mm X 4 mm) and large-scale brain networks via intracranial EEGs (Truccolo et al., 2011, 2014; Wagner et al., 2015). Neural activity at these multiple levels will be recorded continuously 24hr/day, over a period of ~1-2 weeks. Furthermore, we will determine the association between recurrent pathological patterns and changes in the brain's susceptibility to spread of excitation and seizures by actively probing neural dynamics with a recently developed real-time closed-loop intracranial electrical stimulation platform (Sarma et al., 2016). Three specific AIMs will: (1) Test the hypothesis that multiscale ictal patterns recur not only during seizures but also during interictal periods, becoming part of the resting state networks' repertoire; (2) Test the hypothesis that precursor biomarkers of seizure initiation include the reactivation of multiscale ictal network patterns; (3) Test the hypothesis that ictal pattern reactivation during interictal periods is accompanied by increases in the brain's susceptibility to both local and large-scale spread of excitation: probing neural dynamics with closed-loop electrical stimulation.