Automatic glissade determination through a mathematical model in electrooculographic records

TL;DR

This paper presents a mathematical and machine learning-based method to automatically detect glissades in electrooculographic records, improving the analysis of ocular movements related to neural programming failures.

Contribution

It introduces a novel procedure combining a Gauss series model and machine learning to identify glissades in saccadic eye movements.

Findings

Effective detection of glissades using the proposed model

High accuracy in classifying presence or absence of glissades

Potential for improved ocular movement analysis in clinical settings

Abstract

The glissadic overshoot is characterized by an unwanted type of movement known as glissades. The glissades are a short ocular movement that describe the failure of the neural programming of saccades to move the eyes in order to reach a specific target. In this paper we develop a procedure to determine if a specific saccade have a glissade appended to the end of it. The use of the third partial sum of the Gauss series as mathematical model, a comparison between some specific parameters and the RMSE error are the steps made to reach this goal. Finally a machine learning algorithm is trained, returning expected responses of the presence or not of this kind of ocular movement.