An improved algorithm for automatic detection of saccades in eye movement data and for calculating saccade parameters

This analysis of time series of eye movements is a saccade-detection algorithm that is based on an earlier algorithm. It achieves substantial improvements by using an adaptive-threshold model instead of fixed thresholds and using the eye-movement acceleration signal. This has four advantages: (1) Adaptive thresholds are calculated automatically from the preceding acceleration data for detecting the beginning of a saccade, and thresholds are modified during the saccade. (2) The monotonicity of the position signal during the saccade, together with the acceleration with respect to the thresholds, is used to reliably determine the end of the saccade. (3) This allows differentiation between saccades following the main-sequence and non-main-sequence saccades. (4) Artifacts of various kinds can be detected and eliminated. The algorithm is demonstrated by applying it to human eye movement data (obtained by EOG) recorded during driving a car. A second demonstration of the algorithm detects microsleep episodes in eye movement data.     

Application of digital filters in the processing of eye movement data

Digital filter techniques have been applied to the analysis of eye movement data. Methods were developed to calculate eye velocity and eye acceleration in real-time from an electronystag-mogram (ENG) signal that was recorded using a one-pole RC high-pass filter in the preamplifier. Nonrecursive, finite impulse-response digital filters were designed to remove the effects of the RC high-pass filter and calculate the first and second time derivatives of the ENG signal, as well as remove high-frequency noise. Applying these new techniques to the analysis of vestibular nystagmus enables estimation of the transfer characteristics of the vestibuloocular system.     

Analysis of methods to determine the latency of online movement adjustments

When studying online movement adjustments, one of the interesting parameters is their latency. We set out to compare three different methods of determining the latency: the threshold, confidence interval, and extrapolation methods. We simulated sets of movements with different movement times and amplitudes of movement adjustments, all with the same known latency. We applied the three different methods in order to determine when the position, velocity, and acceleration of the adjusted movements started to deviate from the values for unperturbed movements. We did so both for averaged data and for the data of individual trials. We evaluated the methods on the basis of their accuracy and precision, and according to whether the latency was influenced by the intensity of the movement adjustment. The extrapolation method applied to average acceleration data gave the most reliable estimates of latency, according to these criteria.     

Updating target location at the end of an orienting saccade affects the characteristics of simple point-to-point movements

Six results are reported. (a) Reaching accuracy increases when visual capture of the target is allowed (e.g., target on vs. target off at saccade onset). (b) Whatever the visual condition, trajectories diverge only after peak acceleration, suggesting that accuracy is improved through feedback mechanisms. (c) Feedback corrections are smoothly implemented, causing the corrected and uncorrected velocity profiles to exhibit similar shapes. (d) Initial kinematics poorly predict final accuracy whatever the condition, indicating that target capture is not the only critical input for feedback control. (e) Hand and eye final variability are unrelated, suggesting that gaze direction is not a target signal for arm control. (f) Extent errors are corrected without modification of movement straightness; direction errors cause path curvature to increase. Together these data show that movements with straight paths and bell-shaped velocity profiles are not necessarily ballistic.     

Detecting eye movements in dynamic environments

To take advantage of the increasing number of in-vehicle devices, automobile drivers must divide their attention between primary (driving) and secondary (operating in-vehicle device) tasks. In dynamic environments such as driving, however, it is not easy to identify and quantify how a driver focuses on the various tasks he/she is simultaneously engaged in, including the distracting tasks. Measures derived from the driver’s scan path have been used as correlates of driver attention. This article presents a methodology for analyzing eye positions, which are discrete samples of a subject’s scan path, in order to categorize driver eye movements. Previous methods of analyzing eye positions recorded in a dynamic environment have relied completely on the manual identification of the focus of visual attention from a point of regard superimposed on a video of a recorded scene, failing to utilize information regarding movement structure in the raw recorded eye positions. Although effective, these methods are too time consuming to be easily used when the large data sets that would be required to identify subtle differences between drivers, under different road conditions, and with different levels of distraction are processed. The aim of the methods presented in this article are to extend the degree of automation in the processing of eye movement data by proposing a methodology for eye movement analysis that extends automated fixation identification to include smooth and saccadic movements. By identifying eye movements in the recorded eye positions, a method of reducing the analysis of scene video to a finite search space is presented. The implementation of a software tool for the eye movement analysis is described, including an example from an on-road test-driving sample.     

Evaluating and reporting data quality in eye movement research

In order to judge the degree of confidence one should have in the results of an experiment using eye movement records as data, it is necessary to have information about the quality of the eye movement data themselves. Suggestions are made for ways of assessing and reporting this information. The paper deals with three areas: characteristics of the eye movement signal, algorithms used in reducing the data, and accuracy of the eye position data. It is suggested that all studies involving eye movement data should report such information. Appendices include linear interpolation algorithms for mapping from the eye movement signal to stimulus space and a way of obtaining an index of accuracy for each data point.     

The extraretinal signal from the pursuit-eye-movement system: its role in the perceptual and the egocentric localization systems

The accuracy of perceptual judgment of the distance of a moving target tracked at various velocities by pursuit eye movements was examined in relation to the amount of two types of eye movement (smooth pursuit eye movement and compensatory saccade) involved in eye tracking. The perceptually judged distance became shorter as the amount of pursuit-eye-movement component in eye tracking increased. A detailed analysis of the eye-movement data and the size of perceptual underestimation indicated that the underestimation was mainly caused by inaccurate extraretinal information derived from the pursuit-eye-movement system, which underestimated the distance at a constant ratio, irrespective of the velocity of tracking. Egocentric localization was not affected by the mode of eye movements, indicating that the egocentric localization system functions without interference from the inaccurate information from the pursuit-eye-movement system.     

Dyslexia: saccadic eye movements

This paper describes an extensive study of the parameters of saccadic eye movement in a group of 28 poor-reading children and a comparative normally reading group of 31 children. Ages ranged from 6.0 to 16.9 yr. Poor readers had normal intelligence but were lagging by at least two years in reading ability as compared to their peer age group. Parameters studied in refixation eye movements included saccadic latency, accuracy, velocity, and acceleration, as well as differences in latency for abduct and adduct movements. Shapes of saccades were also studied. It was shown that eye-movement performance in simple saccadic refixation was not distinguishably different for poor readers and normal readers, even though in reading, which requires higher processing, eye-movement characteristics are quite different. Maturational changes for various saccadic eye-movement parameters were also examined for the normal and poor reading groups.     

An acceleration illusion caused by underestimation of stimulus velocity during pursuit eye movements: Aubert-Fleischl revisited

When the eyes pursue a fixation point that sweeps across a moving background pattern, and the fixation point is suddenly made to stop, the ongoing motion of the background pattern seems to accelerate to a higher velocity. Experiment I showed that this acceleration illusion is not caused by the sudden change in (i) the relative velocity between background and fixation point, (ii) the velocity of the retinal image of the background pattern, or (iii) the motion of the retinal image of the rims of the CRT screen on which the experiment was carried out. In experiment II the magnitude of the illusion was quantified. It is strongest when background and eyes move in the same direction. When they move in opposite directions it becomes less pronounced (and may disappear) with higher background velocities. The findings are explained in terms of a model proposed by the first author, in which the perception of object motion and velocity derives from the interaction between retinal slip velocity information and the brain's 'estimate' of eye velocity in space. They illustrate that the classic Aubert-Fleischl phenomenon (a stimulus seems to be moving slower when pursued with the eyes than when moving in front of stationary eyes) is a special case of a more general phenomenon: whenever we make a pursuit eye movement we underestimate the velocity of all stimuli in our visual field which happen to move in the same direction as our eyes, or which move slowly in the direction opposite to our eyes.     

Real-time manipulation of visual displacement during manual aiming

This study examined the spatial and temporal limitations of the visual corrective process in the control of upper limb movements. Real-time calculation of kinematic data was used to trigger a prismatic displacement of the movement environment during manual aiming. Using an OptoTrak motion tracking system, a data acquisition unit, and a custom-made program, perturbations were triggered at peak acceleration, peak velocity, and the estimated time of peak deceleration. Movement outcome was significantly influenced only when the visual displacement occurred at peak acceleration. The results support models of visual control that posit that early visual information is required for accurate limb control.     

Speed-accuracy characteristics of saccadic eye movements

Speed-accuracy trade-off characteristic of horizontal saccadic eye movements were examined in this study. Unlike limb movements, saccadic eye movements are preprogrammed, unidimensional, and do not involve target impact. Hence, they provide an optimal test of the impulse variability account of the speed-accuracy trade-off in rapid movements. Subjects were required to alternately look at two target lights as fast and as accurately as possible for a period of 10 s. Target lights subtended angles of 5, 10, 15, and 20 degrees. By restricting target distances to less than 20 degrees of arc, the speed-accuracy relation was examined for single horizontal saccadic movements of the eye. movement of the dominant eye was tracked with an infra-red eye monitoring device. Fifty saccadic movements of the eye were recorded for each target distance and used to compute the average amplitude, duration, and velocity of eye movements, as well as, movement endpoint variability. An increase in both average velocity and movement endpoint variability with increasing movement amplitude was found. This, together with the unique features of the eye movement system, support the impulse variability account of the speed-accuracy trade-off in rapid movements.     

What you learn is what you see: using eye movements to study infant cross-situational word learning

Recent studies show that both adults and young children possess powerful statistical learning capabilities to solve the word-to-world mapping problem. However, the underlying mechanisms that make statistical learning possible and powerful are not yet known. With the goal of providing new insights into this issue, the research reported in this paper used an eye tracker to record the moment-by-moment eye movement data of 14-month-old babies in statistical learning tasks. Various measures are applied to such fine-grained temporal data, such as looking duration and shift rate (the number of shifts in gaze from one visual object to the other) trial by trial, showing different eye movement patterns between strong and weak statistical learners. Moreover, an information-theoretic measure is developed and applied to gaze data to quantify the degree of learning uncertainty trial by trial. Next, a simple associative statistical learning model is applied to eye movement data and these simulation results are compared with empirical results from young children, showing strong correlations between these two. This suggests that an associative learning mechanism with selective attention can provide a cognitively plausible model of cross-situational statistical learning. The work represents the first steps in using eye movement data to infer underlying real-time processes in statistical word learning.     

Response timing variability: coherence of kinematic and EMG parameters

A detailed kinematic and electromyographic (EMG) analysis of single degree of freedom timing responses is reported to (a) determine the coherence of kinematic and EMG variability to the reduced timing error variability exhibited with amplitude increments within a given criterion movement time and (b) understand the temporal organization of various movement parameters in simple responses. The data reveal that the variability of kinematic (time to peak acceleration, duration of acceleration phase, time to peak deceleration) and EMG (duration of agonist burst, duration of antagonist burst, time to antagonist burst) timing parameters decreased with increments of average velocity in a manner consistent with the variable timing error. In addition, the coefficient of variation for peak acceleration, peak deceleration, and integrated EMG of the agonist burst followed the same trend. Increasing average movement velocity also led to decreases in premotor and motor reaction times. Overall, the findings suggest a strong coherence between the variability of response outcome, kinematic, and EMG parameters.     

Response Timing Variability

A detailed kinematic and electromyographic (EMG) analysis of single degree of freedom timing responses is reported to (a) determine the coherence of kinematic and EMG variability to the reduced timing error variability exhibited with amplitude increments within a given criterion movement time and (b) understand the temporal organization of various movement parameters in simple responses. The data reveal that the variability of kinematic (time to peak acceleration, duration of acceleration phase, time to peak deceleration) and EMG (duration of agonist burst, duration of antagonist burst, time to antagonist burst) timing parameters decreased with increments of average velocity in a manner consistent with the variable timing error. In addition, the coefficient of variation for peak acceleration, peak deceleration, and integrated EMG of the agonist burst followed the same trend. Increasing average movement velocity also led to decreases in premotor and motor reaction times. Overall, the findings suggest a strong coherence between the variability of response outcome, kinematic, and EMG parameters.     

Instrument considerations in measuring fast eye movements

The dynamic limitations of eye movement recorders can distort the measurement of fast eye movements such as saccades and nystagmic quick phases. In this paper, the effects of the bandwidth and noise of recording methods and the problems incurred by digital sampling are discussed theoretically with respect to the measurement of peak velocity and duration of fast eye movements. As a practical example, a TV-based infrared corneal reflex system is examined and a method for calibrating it for peak velocity measurement is described.     

Perception and extrapolation of velocity and acceleration.

A moving target disappeared behind a screen and subjects predicted when the target passed behind a marker on the screen. When the target moved with constant velocity, predictions were extremely accurate, regardless of the spatial and temporal exposure and concealment of the target and regardless of its rate of velocity. When the target accelerated, accuracy of prediction decreased with increasing acceleration and with increasing target concealment. Analyses of the results suggest that the perception of velocity and acceleration is direct and accurate and that extrapolation of velocity and acceleration incorporates concrete and abstract characteristics of the motion that was seen. It is proposed that the motion perception system is tuned to accelerated rather than to constant velocity movement.     

Information entropy analysis of discrete aiming movements

Information entropy and mutual information were investigated in discrete movement aiming tasks over a wide range of spatial (20-160 mm) and temporal (250-1250 ms) constraints. Information entropy was calculated using two distinct analyses: (1) with no assumption on the nature of the data distribution; and (2) assuming the data have a normal distribution. The two analyses showed different results in the estimate of entropy that also changed as a function of task goals, indicating that the movement trajectory data were not from a normal distribution. It was also found that the information entropy of the discrete aiming movements was lower than the task defined indices of difficulty (ID) that were selected for the congruence with Fitts' law. Mutual information between time points of the trajectory was strongly influenced by the average movement velocity and the acceleration/deceleration segments of the movement. The entropy analysis revealed structure to the variability of the movement trajectory and outcome that has been masked by the traditional distributional analyses of discrete aiming movements.     

On-line eye movement recorders: The good,the bad,and the ugly

The first section of this paper describes three types of problems inherent in the use of on-line eye movement recorders (EMRs), These types of problems are (1) that software is difficult to develop, (2) that an on-line EMR is subject to all difficulties that are normally inherent in computer use and, (3) that with an on-line EMR. it is difficult to detect when the gaze is falling on an object that has an irregularly shaped boundary. It is argued that a large class of eye movement experimentation could be done as easily or more easily with an off-line EMR. The second section of the paper describes an off-line EMR laboratory currently in use by the author.     

Speed-Accuracy Characteristics of Saccadic Eye Movements

Speed-accuracy trade-off characteristic of horizontal saccadic eye movements were examined in this study. Unlike limb movements, saccadic eye movements are preprogrammed, unidimensional, and do not involve target impact. Hence, they provide an optimal test of the impulse variability account of the speed-accuracy trade-off in rapid movements. Subjects were required to alternately look at two target lights as fast and as accurately as possible for a period of 10 s. Target lights subtended angles of 5,10,15, and 20°. By restricting target distances to less than 20° of arc, the speed-accuracy relation was examined for single horizontal saccadic movements of the eye. Movement of the dominant eye was tracked with an infra-red eye monitoring device. Fifty saccadic movements of the eye were recorded for each target distance and used to compute the average amplitude, duration, and velocity of eye movements, as well as, movement endpoint variability. An increase in both average velocity and movement endpoint variability with increasing movement amplitude was found. This, together with the unique features of the eye movement system, support the impulse variability account of the speed-accuracy trade-off in rapid movements.