Seizure Detection Using Sequential and Coincident Power Spectra with Deterministic Finite Automata Rory A. Lewis 1,2 , Andrew M. White 1 1. Departments of Pediatrics & Neurology, University of Colorado Denver, Anschutz Medical Campus 2. Department of Computer Science, University of Colorado at Colorado Springs Rory.Lewis@ucdenver.edu, Andrew.White@ucdenver.edu Abstract In the continuing endeavor to create a robust method- ology for identifying epileptiform activity in the human brain the authors present a new tool to identify epileptic seizures in a noisy domain. This paper presents an addi- tional component for identifying epileptiform activity and seizures from Electroencephalograms (EEG). We present a system that combines 1) power spectra similarity mea- sures between channels and successive time intervals with 2) Deterministic Finite Automata (DFA) The power spectra similarity define a metric that separates the states between seizure and non-seizure. The use of DFA increases the rigor of this methodology even when noisy signals are encoun- tered. Herein we present a dual methodology that increases epileptoid identification in a noisy domain. 1 Introduction Epilepsy is a neurological disorder that makes people susceptible to recurrent unprovoked seizures due to electri- cal disturbances in the brain. Unfortunately, 30% of patients that suffer from epilepsy are not well controlled on medica- tion. Only a small fraction of these can be helped by seizure surgery [5]. Therefore, it would be life changing to a large number of individuals if a system could be developed that would predict a seizure hours, minutes, or even seconds be- fore its clinical onset. The challenge in this problem is that the dimensionality is huge; in the human brain there are approximately 100 billion neurons, each with about 1000 connections (synapses)[20]. Even in the rat brain it is es- timated that there are approximately 200 million neurons [4], [1]. The connections are wired such that the problem is highly chaotic. In a certain class of seizures it would be helpful if they could be detected even a few seconds prior to the start of a seizure. The dimensionality of the problem A B C D E F G H I J Figure 1. Implantable Tethered System Devices: (A-D) Stereotaxic placement of cortical elec- trodes. (E) Dental cement polymer applied to hold the electrodes in place. Note dental ce- ment on q-tip. (F) The tethered pre-amplifier connects to the implanted electrodes and sends the signals to Epilepsy Monitoring Unit) can be significantly reduced, with only a small loss of informa- tion by recording electrical potentials at multiple points on the surface of the skull or, using depth electrodes, in the hippocampus (EEG). EEGs are accepted as one of the best means of evaluating neurocognitive functions [12]. EEG spike/seizure detection and prediction is made more com- plicated by the following: (1) For a single individual, no two seizures or even their EEG correlates are exactly alike, (2) seizures from different individuals vary significantly, (3) there is no single metric that consistently changes during all