Abstract
This paper presents preliminary results on epileptic seizure detection. Combination of functional Near Infrared Spectroscopy (fNIRS) and Electroencephalogram (EEG) recordings shows enhanced performance compare to EEG recordings alone. Moreover, some results concerning the anticipation at which a seizure can be detected are also presented.

Introduction
Epilepsy is one of the top three most common neurological disorders; it is just below strokes and Alzheimer disease [1]. It is estimated that 65 million people around the world suffer from epilepsy, which represents approximately 1% of the global population [2]. On average, 60 new cases of epilepsy ۱۰۰,۰۰۰ people emerge every year [2]–[۴]. Epileptic seizures are the hallmark of epilepsy. They are brief episodes of perceptive or behavioral disturbances due to an abnormal excessive synchronization of a large group of neurons within the cerebral cortex. These episodes may last from few seconds to few minutes [1], [2].
Unpredictability of epileptic seizures represents one of the main problems for patients with this disorder [5]. Therefore, reliable and early detection of epileptic seizures is a problem of interest [6]. For years, several methods have been proposed to solve this problem; however, most of them have focused on the solely analysis of EEG recordings [6]. Nonetheless, recent publications suggest that information contained within fNIRS recordings (i.e., relative concentration of oxyhemoglobin, HbO, and deoxyhemoglobin, HbR) can be used to solve this problem [7]-[9].
Being able to detect epileptic seizures in advance would unfailingly derive in improvements to the quality of life of patients with this condition. Detecting seizures could allow patients with epilepsy and/or their caretakers to take precautions before the occurrence of an epileptic event. This would drastically minimize potential risks.
This work is organized as follows: in the second section a brief description of EEG and fNIRS recordings is presented. Third section describes the methodology followed in this work: ۱٫ The investigation of including multi-modal data and 2. Variations on the definition of pre-ictal window. Results from the applied methodology are presented and discussed in section number four. Finally, our conclusions as well as some guidelines on our future work are given in fifth section.