Dual Purkinje Image Eyetracker Research Articles

Poster Session: Fixational Eye Movements in Myopia.

During fixation, an incessant drift of the eye keeps the image impinging on the retina always in motion. Previous work indicated that luminance modulations from ocular drift serve important visual functions in emmetropes (Intoy & Rucci, 2020; Clark et al 2022). However, it remains unknown how ocular drift varies under myopia, a visual impairment commonly caused by eye elongation. We measured eye movements in 19 individuals with varying degrees of myopia (-0.25D to -6.5D) using a digital Dual-Purkinje Image eye-tracker, a recently developed system with sub-arcminute resolution. Subjects observed stimuli monocularly with vision corrected via a Badal optometer. They engaged in two high-acuity tasks: (a) resolution of a 20/20 line of an eye chart (5 evenly spaced tumbling E optotypes); and (b) a more natural task where subjects were presented with images of distant faces (1⁰) and asked to report the image's gaze direction. We show ocular drift characteristics differ in myopes relative to emmetropes. Drift was faster and less curved in myopic observers. On the retina, these changes result in luminance modulations that amplify low spatial frequencies at the expense of high spatial frequencies, so that high-frequency signals are effectively weaker in myopes These results are consistent with the proposal that fine spatial vision strongly relies on oculomotor-induced luminance modulations and emphasize the importance of considering fine eye movements in myopia.

High-resolution oculomotor measurements via a digital Dual Purkinje Image eye-tracker

High-resolution oculomotor measurements via a digital Dual Purkinje Image eye-tracker

Contrast Sensitivity of Vertical Vergence Eye Alignment

Purpose: The purpose of the study was to measure the response of vertical vergence eye movement to bandpass filtered noise of various spatial frequencies and contrasts. The results of our previous experiment in which we measured the contrast sensitivities using vertical vergence response to sine-wave gratings suggested that vertical vergence responses were robust for middle spatial frequencies (0.5 to 4 cycles per degree). However, the responses to higher and lower spatial frequencies were weaker. The weak responses at higher spatial frequencies could have been due to matching ambiguity of the gratings in the two eyes. So, to eliminate this ambiguity we used bandpass filtered noise as the stimulus to measure the contrast sensitivity of vertical vergence eye alignment. Methods: Subjects were 2 adults with normal vision. A Dual Purkinje image eye tracker was used for eye tracking. Vertical vergence was measured in response to full-screen bandpass filtered noise patterns of all orientations with one-octave bandwidth. A block of trials comprised combinations of 8 contrasts [0%, 1%, 2%,4%,8%,16%,32%,64%] and 8 spatial frequencies [0.125, 0.25, 0.50, 1, 2, 4, 8 ,16 cpd] presented in shuffled order. Various Oculometric functions (AUC, Percent Correct, Weibull fit, Vergence amplitudes) were created to determine the oculometric CSFs. The psychophysical CSF was obtained with a 2AFC staircase method for the same spatial frequencies using bandpass filtered noise patterns as stimuli. Results: Vertical vergence responses were robust for middle spatial frequencies. The responses for higher spatial frequencies were weaker or absent, similar to the results with sine-wave gratings. Conclusion: The results of this study suggest that matching ambiguity was not the primary limiting factor in the response of vertical vergence at higher spatial frequencies, despite being easily detected psychophysically. Further research is needed to understand why vertical vergence does not show a response at higher spatial frequencies.

Monocular microsaccades: Do they really occur?

Small saccades, known as microsaccades, occur frequently during fixation. Several recent studies have argued that a considerable fraction of these movements are present in the traces from one eye only. This claim contrasts with the findings of older reports, which concluded that microsaccades, like larger saccades, are virtually always binocular events. Here we examined the characteristics of small saccades by means of two of the most established high-resolution eye-tracking techniques available. A binocular Dual Purkinje Image eye-tracker was used to record eye movements while observers fixated, with their head immobilized, on markers displayed on a monitor. A specially designed eye-coil system was used to measure eye movements during normal head-free viewing, while subjects fixated on markers at various distances. Monocular microsaccades were virtually absent in both datasets. In the head-fixed data, not a single monocular microsaccade was observed. In the head-free data, only one event appeared to be monocular out of more than a thousand saccades. Monocular microsaccades do not seem to occur during normal head-free or head-immobilized fixation.

Bayesian microsaccade detection.

Microsaccades are high-velocity fixational eye movements, with special roles in perception and cognition. The default microsaccade detection method is to determine when the smoothed eye velocity exceeds a threshold. We have developed a new method, Bayesian microsaccade detection (BMD), which performs inference based on a simple statistical model of eye positions. In this model, a hidden state variable changes between drift and microsaccade states at random times. The eye position is a biased random walk with different velocity distributions for each state. BMD generates samples from the posterior probability distribution over the eye state time series given the eye position time series. Applied to simulated data, BMD recovers the “true” microsaccades with fewer errors than alternative algorithms, especially at high noise. Applied to EyeLink eye tracker data, BMD detects almost all the microsaccades detected by the default method, but also apparent microsaccades embedded in high noise—although these can also be interpreted as false positives. Next we apply the algorithms to data collected with a Dual Purkinje Image eye tracker, whose higher precision justifies defining the inferred microsaccades as ground truth. When we add artificial measurement noise, the inferences of all algorithms degrade; however, at noise levels comparable to EyeLink data, BMD recovers the “true” microsaccades with 54% fewer errors than the default algorithm. Though unsuitable for online detection, BMD has other advantages: It returns probabilities rather than binary judgments, and it can be straightforwardly adapted as the generative model is refined. We make our algorithm available as a software package.

Asymmetry in saccadic latency during smooth pursuit: A signature of visual spatial attention?

Previous studies reported faster saccadic reaction times (SRTs) for targets presented in the same compared to the opposite direction of pursuit, suggesting that the locus of visual spatial attention might be ahead of the pursuit target. We revisited these findings to investigate if task instruction or saccadic amplitude influenced the difference in SRTs between forward and backward saccades during ongoing pursuit. The pursuit target was a single letter from the set, CDHKNORSVZE, which traversed the screen horizontally in sinusoidal motion with a peak-to-peak amplitude of 10 degrees and a frequency of 0.3 Hz. Every 500 ms the target changed to a different letter. Each time the letter crossed the straight-ahead position, it jumped either forward or backward with an amplitude of 2.5 or 4 degrees. A dual Purkinje image eye tracker was used to track subjects' (N = 4) eye movements at a sampling rate of 357 Hz. Two different tracking conditions were run in blocks, each lasting for 24 seconds and consisting of 8 pursuit cycles with 16 target jumps. In the 'track only' condition subjects were instructed to hold their gaze on the center of the changing letter throughout the course of stimulus motion. In the 'read' condition the subjects were encouraged only to read the changing string of letters. In both viewing conditions, saccadic reaction times were on average 30-40 ms less for saccades to targets in the direction of pursuit compared to the saccades to targets that jumped in the opposite direction. This difference could not be explained by a difference in the amplitudes of forward vs. backward saccades. The results concur with those of previous studies and suggest that attention remains ahead of the pursuit target. Meeting abstract presented at VSS 2016

An eye-movement-defined hierarchy of visual stimuli

Reverse correlation of stimulus velocity with eye velocity can be used to infer pursuit latency in a task in which a subject attempts to follow a target undergoing random motion. Using a binocular dual Purkinje image eye tracker equipped with stimulus deflectors, we measured tracking responses to four-dimensional motion, produced by delivering an independent random walk to each eye, analyzing the results in terms of horizontal and vertical vergence and version. High-contrast luminance-defined stimuli produce the shortest latencies (around 100 msec); varying texture contrast increases the delay by approximately 50 msec / log unit. Nulling the first-order luminance contrast in a second-order constrast-modulated target produces a dramatic increase in latency (over 100 msec additional delay), and abolishes the vertical vergence response - the only one of the four types of movement that cannot be executed voluntarily. We propose a model of the system in which a fast reflexive system responds to a limited class of stimuli, while a slower voluntary system is capable of tracking anything that can be seen. Meeting abstract presented at OSA Fall Vision 2012

Temporal Encoding of Spatial Information during Active Visual Fixation

Humans and other species continually perform microscopic eye movements, even when attending to a single point. These movements, which include drifts and microsaccades, are under oculomotor control, elicit strong neural responses, and have been thought to serve important functions. The influence of these fixational eye movements on the acquisition and neural processing of visual information remains unclear. Here, we show that during viewing of natural scenes, microscopic eye movements carry out a crucial information-processing step: they remove predictable correlations in natural scenes by equalizing the spatial power of the retinal image within the frequency range of ganglion cells' peak sensitivity. This transformation, which had been attributed to center-surround receptive field organization, occurs prior to any neural processing and reveals a form of matching between the statistics of natural images and those of normal eye movements. We further show that the combined effect of microscopic eye movements and retinal receptive field organization is to convert spatial luminance discontinuities into synchronous firing events, beginning the process of edge detection. Thus, microscopic eye movements are fundamental to two goals of early visual processing: redundancy reduction and feature extraction.

An Eye-Movement-Defined Hierarchy of Visual Stimuli

Reverse correlation of stimulus velocity with eye velocity can be used to infer pursuit latency in a task in which a subject attempts to follow a target undergoing random motion. Using a binocular dual Purkinje image eye tracker equipped with stimulus deflectors, we measured tracking responses to four-dimensional motion, produced by delivering an independent random walk to each eye, analyzing the results in terms of horizontal and vertical vergence and version. High-contrast luminance-defined stimuli produce the shortest latencies (around 100 msec); varying texture contrast increases the delay by approximately 50 msec/log unit. Nulling the first-order luminance contrast in a second-order constrast-modulated target produces a dramatic increase in latency (over 100 msec additional delay), and abolishes the vertical vergence response - the only one of the four types of movement that cannot be executed voluntarily. We propose a model of the system in which a fast reflexive system responds to a limited class of stimuli, while a slower voluntary system is capable of tracking anything that can be seen.

Miniature eye movements measured simultaneously with ophthalmic imaging and a dual-Purkinje image eye tracker

Background: Scanning laser ophthalmoscopes with adaptive optics correction of ocular aberrations provide retinal images of unprecedented resolution, allowing for real-time imaging of photoreceptors. Eye movements made by the subject/patient during recording produce distortions that must be corrected before multiple frames can be added together to achieve noise reduction or to build a mosaic image from different retinal areas. These distortions also provide a high spatial and temporal resolution record of the miniature eye movements made during fixation. Here we report simultaneous measurements of fixation eye movements with an Adaptive Optics Scanning Laser Ophthalmoscope (AOSLO) and a dual-Purkinje image (dPi) eye tracker in order to cross-validate these two methods of recording miniature eye movement. Method: Foveas of three subjects were imaged with a one degree square scan using the Houston AOSLO, at a resolution of 8 pixels per arc minute. A Generation V dPi tracker from SRI was placed in front of the AOSLO, and eye movements were recorded at the same time from the same eye being imaged by the AOSLO. AOSLO movies were analyzed off line to extract retinal image motion. The resulting traces were then overlaid on the dPi recordings for comparison. Results: The two methods produced records that agreed to within about one arc minute, with more significant disagreements occurring after eye blinks. Microsaccades in the dPi record were accompanied by overshoots that have previously been associated with lens wobble. AOSLO traces also showed saccade-related overshoots, but of much smaller amplitude. Conclusions: Eye movement recordings measured with dual Purkinje image trackers predict retinal image motion to a precision of about 1 arc minute, except for 10–20 milliseconds following each saccade and 500 – 1000 milliseconds following each eye blink. Retinal image motion measured directly from AOSLO recordings can be recovered to a precision of just a few arc seconds.

The effects of preceding moving stimuli on the initial part of smooth pursuit eye movement

We examined whether there are any adaptive effects on the pursuit initiation after a prolonged exposure to moving visual stimuli. The eye movements of six human subjects were recorded with the scleral search-coil technique or a Dual Purkinje Image Eye-tracker system. A random-dot image appeared on a CRT monitor and moved coherently in one direction (rightward or leftward) at 10 deg/s for 4 s, while the subject fixated on a stationary target (conditioning stimulus). The screen was blanked for 0.2 s, and then the target stepped to the right or left of the center and moved 10 deg/s leftward or rightward. We measured change in the eye position over the open-loop period of the pursuit initiation. When the pursuit target moved in the same direction as the preceding visual stimulus, a significant reduction in the initial tracking responses (55.9% decrease on average) was found. We then studied in detail the properties of the motion adaptation in pursuit initiation by varying the visual conditions systematically and obtained the following findings. When the subjects tracked the target that moved at 10 deg/s, the pursuit initiation was affected not only by the conditioning stimulus of the same speed as the target, but also by those of different speeds. Further, the conditioning stimulus moving at 10 deg/s affected the pursuit initiation not only when the target moved with the same speed but also when it moved at different speeds (more remarkable for slower speeds). The effect of conditioning stimuli on the pursuit initiation was larger when the duration of the conditioning period was longer. The effect of conditioning stimuli decayed as the duration of the blank period became longer. The findings from the present study are consistent with the properties of neurons in the middle temporal area of monkeys.

The effect of retinal stabilization on anorthoscopic percepts under free-viewing conditions

Outline shapes were moved back and forth at .5 Hz behind a simulated narrow (10′ or 15′) slit without visible borders. Under free-viewing conditions observers reported spontaneous transitions between periods in which they saw a horizontally moving shape and periods in which they saw only vertically moving contour segments. Eye movements were monitored with a dual-Purkinje image eye-tracker. On half the trials we used retinal stabilization to eliminate any pursuit dependent retinal painting. With no stabilization, we observed low amplitude spontaneous pursuit of the horizontal figure motion when a shape was seen, and occasionally lower amplitude pursuit even when no shape was seen. Pursuit amplitudes increased when the display was stabilized. However, stabilization did not alter either the duration or frequency of the episodes of shape perception. Our findings indicate that under the free-viewing conditions we employed, spontaneous smooth pursuit and pursuit dependent painting did not contribute to the formation or maintenance of anorthoscopic percepts.

Perception onset time during fixations in free viewing

In this study, we investigated when visual perception begins in fixations. During picture viewing, the picture was degraded at the beginning of selected saccades and changed back to the original after varying intervals. Participants manually responded whenever they detected changes. The change-backs were undetected when they occurred <6 msec after the end of the saccade, marked by the peak of the overshoot in dual Purkinje image eyetracker data, and detection reached asymptote 32 msec after that marker. Eye velocity at the change-back time also affected detection likelihood. Apparently, perception begins around the time at which the eyes stop rotating at the end of a saccade, giving a psychological justification for measuring fixation durations from then. This also specifies the deadline for gaze-contingent display changes to occur without detectable image motion. Investigators using the dual Purkinje image eyetracker should consider the peak of the overshoot as the fixation onset time and measure intrafixational presentation times from then.

Retinal image motion during deliberate fixation: implications to laser safety for long duration viewing.

With the easy attainability of hand-held laser devices and the burgeoning light emitting diode (LED) technology, safety standards for long-term viewing of continuous light sources are being scrutinized. One concern is with quantifying the effect of head and eye movements on the distribution of energy over the retina. This experiment describes target motion over the retina as a result of head and eye movements during a deliberate fixation task. Volunteers deliberately fixated, with (fettered) and without (unfettered) head and chin rest support, on LEDs that subtended a 0.1 min of arc visual angle. A Dual Purkinje Image Eyetracker measured eye position during each 100-s fixation trial. The data showed an elliptical retinal energy distribution, oriented on the temporal/ nasal retinal axis, with a major axis 1.5 times greater than the minor. The average half-maximum diameter for the major axis was 40 microm for the fettered and 107 microm for the unfettered condition. The retinal area illuminated from head and eye movements showed a positive linear relationship with time. Peak retinal radiant exposure at 100 s was reduced by a factor of 10 as a result of eye movements and by a factor of 30 as a result of head and eye movements. Although the net result was a reduction in radiant exposure from a "no-movement" baseline, the distribution of energy over the retina was well within the 200 microm foveal boundary. The data suggest that the laser permissible exposure limits for long term viewing of a small continuous wave source include a factor that adjusts for this reduction in retinal radiant exposure with time.

Residual Vision in Multiple Retinal Locations within a Scotoma: Implications for Blindsight

There is an important new proposal that "blindsight"-the ability to detect and identify visual stimuli by forcedchoice guessing and in the absence of conscious awareness when they fall in blind regions of the visual field-is a function of residual "islands" of undamaged visual cortex. This stands in contrast to the widely accepted view that blindsight is exclusively a function of secondary visual pathways. According to the new view, residual vision in blindsight should be patchy. Thus, when apparently wide areas of residual vision in blindsight are found, these may be due to eye-movements that allow stimuli to pass over retinal locations corresponding to islands of sparing. We tested this hypothesis by examining the distribution of residual vision in blindsight when the effects of eye movements on the retinal location of stimuli were minimized. We report a series of experiments that examined twealternate forcedchoice discrimination in the blind field of the subject GY. Using a dual-Purkinje image eye-tracker we applied three methods of minimizing the effects of retinal slippage due to eye-movements on discrimination performance: fixation stability-dependent trials, software image stabilization, and post hoc rejection of trials in which saccadic eye-movements were detected. In the first experiment, GY's discrimination performance was significantly above chance in 8 of 15 locations tested. In the subsequent experiments the subject knew the location of the target in each block of trials, and this resulted in improvements to performance in a further three locations. Increasing the luminance of the stimulus display (while maintaining 95% target contrast), and increasing the temporal discriminability of the forced choice produced performance above chance in all but two of the locations tested. The consistent chance performance observed in two locations in the lower visual field nevertheless implies that GY's blindsight does not extend over the whole of his scotoma. Nevertheless, abolishing, or minimizing, the effects of eye-movements did not result in a loss of detection in all the widely separated regions tested, and we thus conclude that GY's blindsight cannot adequately be explained in terms of islands of spared vision. Islands may account for residual vision in scotomata in some patients, but cannot be a universal account of the phenomenon of blindsight.

Fourth Purkinje image signals reveal eye-lens deviations and retinal image distortions during saccades

Saccadic eye movements of various sizes and directions were registered simultaneously with the scleral search coil and a fifth-generation Dual Purkinje image eyetracker. Comparison of the search coil and the Purkinje image tracker records reveal considerable dynamic deviations during and immediately after the saccade, which we ascribe to the movements of the eye lens relative to the optical axis of the eye. Due to the increased stiffness of the tissues these deviations are smaller in older subjects. Also, they are larger at near accommodation. The size of the retinal image displacement which results from the lens movement proper can be as large as 0.5 deg, which may not be negligible in a number of visual tasks.