Automatic identification of oculomotor behavior using pattern recognition techniques Alexandra I. Korda a , Pantelis A. Asvestas b,n , George K. Matsopoulos a , Errikos M. Ventouras b , Nikolaos P. Smyrnis c,d a School of Electrical and Computer Engineering, National Technical University of Athens, Athens, Greece b Department of Biomedical Engineering, Faculty of Technological Applications, Technological Educational Institute of Athens, Agiou Spyridonos Str., Egaleo 12243, Greece c Laboratory of Sensorimotor Control, University Mental Health Research Institute, Athens, Greece d Psychiatry Department, National and Kapodistrian University of Athens, Eginition Hospital, Athens, Greece article info Article history: Received 23 October 2014 Accepted 3 March 2015 Keywords: Saccades Microsaccades Blinks Fixation Classification Neural network Velocity threshold algorithm abstract In this paper, a methodological scheme for identifying distinct patterns of oculomotor behavior such as saccades, microsaccades, blinks and fixations from time series of eye's angular displacement is presented. The first step of the proposed methodology involves signal detrending for artifacts removal and estimation of eye's angular velocity. Then, feature vectors from fourteen first-order statistical features are formed from each angular displacement and velocity signal using sliding, fixed-length time windows. The obtained feature vectors are used for training and testing three artificial neural network classifiers, connected in cascade. The three classifiers discriminate between blinks and non-blinks, fixations and non-fixations and saccades and microsaccades, respectively. The proposed methodology was tested on a dataset from 1392 subjects, each performing three oculomotor fixation conditions. The average overall accuracy of the three classifiers, with respect to the manual identification of eye movements by experts, was 95.9%. The proposed methodological scheme provided better results than the well-known Velocity Threshold algorithm, which was used for comparison. The findings of the present study indicate that the utilization of pattern recognition techniques in the task of identifying the various eye movements may provide accurate and robust results. & 2015 Elsevier Ltd. All rights reserved. 1. Introduction Gesture, speech and eye movements are frequently analyzed by experts in order to explain human behavior. As a result, eye move- ments have enjoyed burgeoning attention in recent years as a tool for studying human behavior [1]. Perhaps the most important reason for their usefulness is that eye movements indicate the focus of visual attention although covert attention can be focused away from the point of visual fixation [2]. The eyes do not remain still when viewing a visual scene; they have to move constantly to build up a mental “map” from interesting parts of the scene [3]. The main reason for this is that only a small central region of the retina, the fovea, is able to perceive with high acuity. According to Leigh and Zee [4], four basic types of eye movements can be identified: (1) Saccades, that are fast voluntary movements, which bring the fovea in the region of interest within the visual field, (2) smooth eye pursuit, that involves a slow continuous movement of both eyes in order to follow a moving visual stimulus in the visual field, (3) vergence, that is the disconjugate slow movements of the eyes that converge or diverge in order to foveate an object in three- dimensional space, and (4) fixation, that is the inhibition of all eye movements which keeps the eyes locked on a particular location in the orbit. When fixing the gaze, there exist microscopic and unnoticed motions of the eye, called fixational eye movements. Furthermore, saccades can be divided into two distinct groups: major saccades, that are easily observed even with naked eye and minor saccades that are virtually unobservable without special instrumen- tation [5]. The smallest saccades, called ‘microsaccades’, are involun- tary eye movements produced during attempted visual fixation. They are the largest and fastest of the fixation eye movements [6]. Saccades have been extensively examined in normal vision towards the understanding of human behavior [7]. Apart from the study of normal vision, saccadic deviations have been also mea- sured in special groups, such as patients with psychiatric disorders (such as attention deficit hyperactivity disorder (ADHD) and schi- zophrenia), young children and elderly people in an attempt to differentiate saccadic characteristics between those conditions and normal controls [8]. Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/cbm Computers in Biology and Medicine http://dx.doi.org/10.1016/j.compbiomed.2015.03.002 0010-4825/& 2015 Elsevier Ltd. All rights reserved. n Correspondening author. Tel.: þ30 210 5385375. E-mail address: pasv@teiath.gr (P.A. Asvestas). Computers in Biology and Medicine 60 (2015) 151–162