Measuring three dimensions of eye movement in dynamic situations by means of videooculography

Developing a Diagnostic Decision Support System for Benign Paroxysmal Positional Vertigo Using a Deep-Learning Model

Background: Diagnosis of benign paroxysmal positional vertigo (BPPV) depends on the accurate interpretation of nystagmus induced by positional tests. However, difficulties in interpreting eye-movement often can arise in primary care practice or emergency room. We hypothesized that the use of machine learning would be helpful for the interpretation. Methods: From our clinical data warehouse, 91,778 nystagmus videos from 3467 patients with dizziness were obtained, in which the three-dimensional movement of nystagmus was annotated by four otologic experts. From each labeled video, 30 features changed into 255 grid images fed into the input layer of the neural network for the training dataset. For the model validation, video dataset of 3566 horizontal, 2068 vertical, and 720 torsional movements from 1005 patients with BPPV were collected. Results: The model had a sensitivity and specificity of 0.910 ± 0.036 and 0.919 ± 0.032 for horizontal nystagmus; of 0.879 ± 0.029 and 0.894 ± 0.025 for vertical nystagmus; and of 0.783 ± 0.040 and 0.799 ± 0.038 for torsional nystagmus, respectively. The affected canal was predicted with a sensitivity of 0.806 ± 0.010 and a specificity of 0.971 ± 0.003. Conclusions: As our deep-learning model had high sensitivity and specificity for the classification of nystagmus and localization of affected canal in patients with BPPV, it may have wide clinical applicability.

Knowing what the brain is seeing in three dimensions: A novel, noninvasive, sensitive, accurate, and low-noise technique for measuring ocular torsion

Torsional eye movements are rotations of the eye around the line of sight. Measuring torsion is essential to understanding how the brain controls eye position and how it creates a veridical perception of object orientation in three dimensions. Torsion is also important for diagnosis of many vestibular, neurological, and ophthalmological disorders. Currently, there are multiple devices and methods that produce reliable measurements of horizontal and vertical eye movements. Measuring torsion, however, noninvasively and reliably has been a longstanding challenge, with previous methods lacking real-time capabilities or suffering from intrusive artifacts. We propose a novel method for measuring eye movements in three dimensions using modern computer vision software (OpenCV) and concepts of iris recognition. To measure torsion, we use template matching of the entire iris and automatically account for occlusion of the iris and pupil by the eyelids. The current setup operates binocularly at 100 Hz with noise <0.1° and is accurate within 20° of gaze to the left, to the right, and up and 10° of gaze down. This new method can be widely applicable and fill a gap in many scientific and clinical disciplines.

Nystagmus as a sign of labyrinthine disorders--three-dimensional analysis of nystagmus

In order to diagnose the pathological condition of vertiginous patients, a detailed observation of nystagmus in addition to examination of body equilibrium and other neurotological tests are essential. How to precisely record the eye movements is one of the goals of the researchers and clinicians who are interested in the analysis of eye movements for a long time. For considering that, one has to think about the optimal method for recording eye movements. In this review, the author introduced a new method, that is, an analysis of vestibular induced eye movements in three-dimensions and discussed the advantages and limitations of this method.

Vestibular function testing

PURPOSE OF REVIEW

This review provides an overview of vestibular function testing and highlights the new techniques that have emerged during the past 5 years.

RECENT FINDINGS

Since the introduction of video-oculography as an alternative to electro-oculography for the assessment of vestibular-induced eye movements, the investigation of the utricle has become a part of vestibular function testing, using unilateral centrifugation. Vestibular evoked myogenic potentials have become an important test for assessing saccular function, although further standardization and methodological issues remain to be clarified. Galvanic stimulation of the labyrinth also is an evolving test that may become useful diagnostically.

SUMMARY

A basic vestibular function testing battery that includes ocular motor tests, caloric testing, positional testing, and earth-vertical axis rotational testing focuses on the horizontal semicircular canal. Newer methods to investigate the otolith organs are being developed. These new tests, when combined with standard testing, will provide a more comprehensive assessment of the complex vestibular organ.

Effects of unilateral ocular motor nerve palsies on smooth pursuit eye movements in adult patients

AIM

The aim of this study was to investigate the influence of single ocular muscle weakness on smooth pursuit eye movements.

METHODS

Infrared video recordings of horizontal and vertical eye movements were obtained from 14 adult patients with either unilateral abducens nerve palsy or trochlear nerve palsy. During the recordings, subsequent series of horizontal, vertical and oblique ramp stimuli of 10 degrees/s constant target velocity and +/-10 degrees amplitude were presented under monocular viewing conditions.

RESULTS

In both forms of ocular nerve palsies, similar changes of pursuit eye movements were observed in the pulling plane of the paretic muscles. The movements of the covered paretic eye showed the lowest amplitude and gain values as well as the lowest numbers of catch-up saccades. The highest amplitude and gain values were calculated from the movements of the covered sound eye. The highest numbers of saccades, however, were produced by the fixating paretic eye.

CONCLUSIONS

We conclude that the fixating paretic eye compensates for the paresis by raising the pursuit gain and the number of catch-up saccades. In the covered paretic eye, however, monocular adaptation is connected with a symmetric low pursuit gain and a reduced number of saccades in the pulling plane of the paretic muscle.

Retinal nerve fiber layer assessment using optical coherence tomography with active optic nerve head tracking

PURPOSE

To develop an eye-motion-tracking optical coherence tomographic (OCT) method and assess its effect on image registration and nerve fiber layer (NFL) thickness measurement reproducibility.

METHODS

A system capable of tracking common fundus features based on reflectance changes was integrated into a commercial OCT unit (OCT II; Carl Zeiss Meditec, Inc., Dublin, CA) and tested on healthy subjects and patients with glaucoma. Twenty successive peripapillary NFL scans were obtained with tracking and 20 without tracking, for 40 images in each session for each eye. Subjects participated in one session on three different days. Composite OCT scans and composite fundus images were generated for assessment of eye tracking. NFL thickness measurement reproducibility was also assessed.

RESULTS

Seven healthy and nine glaucomatous eyes of 16 subjects were recruited. A qualitative assessment of composite OCT scans and composite fundus images showed little motion artifact or blurring along edges and blood vessels during tracking; however, those structures were less clearly defined when tracking was disengaged. There was no significant reproducibility difference with and without tracking in both intra- and intersession NFL measurement SD calculations in any location. The mean retinal pixel SD was significantly smaller with tracking than without (490.9 +/- 19.3 microm vs. 506.4 +/- 31.8 microm, P = 0.005, paired t-test).

CONCLUSIONS

A retinal-tracking system was successfully developed and integrated into a commercial OCT unit. Tracking OCT improved the consistency of scan registration, but did not influence NFL thickness measurement reproducibility in this small sample study.

Use of preoperative assessment of positionally induced cyclotorsion: a video-oculographic study

PURPOSE

Positionally induced cyclotorsion could be an important factor concerning correction of astigmatism in refractive surgery. The method of binocular three dimensional infrared video-oculography (3D-VOG) was used to determine a possible influence of body position on cyclotorsion.

METHODS

38 eyes (19 healthy subjects, median value of age 25.5) with normal binocular vision were examined using 3D-VOG. This method records ocular motions and positions of both eyes simultaneously in the x, y, and z axis. Cycloposition of the eyes was recorded first in a seated position (both eyes open, test 1), then in a supine position (right eye closed, test 2), occlusion of both eyes (test 3), both eyes open (test 4). Cyclovergence was calculated as the difference between the right and the left eye positions.

RESULTS

The range of cyclotorsion of the right and left eye in all four tests was between 1.13 degrees excyclotorsion and 0.34 degrees incyclotorsion. There was no statistically significant difference of the median values for torsion for the four test situations. Concerning the influence of body position on cyclotorsion, a statistically significant difference between the different test positions and settings did not exist. Median values for right/left torsion/cyclovergence were: 0.17/0.04/0.02 (test 1), -0.31/-0.71/-0.16 (test 2), -1.09/-0.60/0.82 (test 3), 0.28/0.28/-0.82 (test 4).

CONCLUSIONS

Cyclotorsion does not significantly change between seated and supine position in subjects with normal binocular vision and stable fixation. In these subjects, an erroneous refractive surgery due to incorrect measurement of the axis of astigmatism in the seated position and performing the refractive surgery in the supine position, is very unlikely.

A Comparative Study on the Observation of Spontaneous Nystagmus with Frenzel Glasses and an Infrared CCD Camera

OBJECTIVES

To compare the usefulness of a CCD camera with infrared illumination (IR-CCD camera) over Frenzel glasses (F Glasses) for the observation of spontaneous nystagmus, the incidence and direction of nystagmus, and the frequency, amplitude and slow phase of spontaneous nystagmus.

METHODS

One hundred vertiginous patients, fifty-three females and forty-seven males participated in this study. Before undergoing routine neurotological examination, their eye movements were recorded by electronystagmogram (ENG) in conjunction with observations of eye movements under F glasses and through an IR-CCD camera. The data was collected from patients who exhibited spontaneous nystagmus either under F glasses or the IR-CCD camera.

RESULTS

Thirty-three patients showed spontaneous nystagmus under F glasses. On the other hand, under the IR-CCD camera, all patients examined exhibited spontaneous nystagmus. The frequency of nystagmus was not significantly different between these two systems. However, the amplitude and slow phase velocity exhibited significantly larger values under the IR-CCD camera in patients with spontaneous nystagmus both under the IR-CCD camera and F glasses.

CONCLUSION

From these observations and evidence, the IR-CCD camera can be recommended as a more useful system and powerful tool for neurotological examination than F glasses.

Compact scanning laser ophthalmoscope with high-speed retinal tracker

The effectiveness of image stabilization with a retinal tracker in a multifunction, compact scanning laser ophthalmoscope (TSLO) was demonstrated in initial human subject tests. The retinal tracking system uses a co confocal reflectometer with a closed-loop optical servo system to lock onto features in the fundus. The system is multifarious and modular to allow configuration for many research a clinical applications. Adult volunteers were tested without mydriasis to optimize the tracking instrumentation and to characterize imaging performance. The retinal tracking system achieves a bandwidth of greater than 1 kHz, which permits tracking at rates that greatly exceed the maximum rate of motion of the human eye. The TSLO system stabilized images to an accuracy of 0.05 deg in all test subjects during ordinary saccades with a velocity up to approximately 500 deg/s. Feature lock was maintained for minutes despite subject eye blinking. Even when nearly 1000 frames were coadded, image blur was minimal. Successful frame coaddition allowed image acquisition with decreased noise in low-light applications. The retinal tracking system significantly enhances the imaging capabilities of the scanning laser ophthalmoscope.

Image stabilization for scanning laser ophthalmoscopy

A scanning laser ophthalmoscope with an integrated retinal tracker (TSLO) was designed, constructed, and tested in human subjects without mydriasis. The TSLO collected infrared images at a wavelength of780 nm while compensating for all transverse eye movements. An active, high-speed, hardware-based tracker was able to lock onto many common features in the fundus, including the optic nerve head, blood vessel junctions, hypopigmentation, and the foveal pit. The TSLO has a system bandwidth of ~1 kHz and robustly tracked rapid and large saccades of approximately 500 deg/sec with an accuracy of 0.05 deg. Image stabilization with retinal tracking greatly improves the clinical potential of the scanning laser ophthalmoscope for imaging where fixation is difficult or impossible and for diagnostic applications that require long duration exposures to collect meaningful information.

Adaptation of torsional eye alignment in relation to smooth pursuit and saccades

The long-term fusion of vertical or horizontal disparities by vergence eye movements is known to evoke persistent changes in vertical and horizontal eye alignment. Adaptive changes in response to torsional disparities have not been well studied. Torsional eye position was measured binocularly with a video system before and after 90 min training periods in which subjects attempted to fuse cyclodisparities. Subjects trained with either a single cyclodisparity presented at a single vertical eye position or with cyclodisparities that varied smoothly from an incyclodisparity to an excyclodisparity as a function of either vertical or horizontal eye position. All five subjects showed persistent changes in binocular torsional eye alignment following both types of training. Incyclodisparities were more easily fused during training and the training aftereffect was greater in that direction. The training aftereffect was observed in relation to both saccades and smooth pursuit under both open-loop and closed-loop viewing conditions. During saccades, the dynamics of the cyclovergence training aftereffect more closely resembled the dynamics of cyclofusional movements than the dynamics of the saccades with which they were associated.

Adaptation of torsional eye alignment in relation to head roll

The coordination of head tilt, ocular counter-roll and vertical vergence is maintained by adaptive mechanisms; the desired outcome being clear single vision. A disruption or imbalance in otolith-ocular pathways may result in diplopia which stimulates these adaptive processes. In the present experiment, dove prisms were used to create cyclodisparities that varied with head tilt about a naso-occipital axis (roll). A stimulus for incyclovergence was presented with the head rolled 45 degrees to one side and a stimulus for an excyclovergence was presented with the head rolled 45 degrees to the other side. At the end of 1 h of training, all subjects demonstrated a change in open-loop cyclovergence that would help to correct for the cyclodisparities experienced during the closed-loop training period. The change appeared to be a simple gain change in the ocular counter-roll of one or both eyes.

Image analysis of quick phase eye movements in nystagmus with high-speed video system

There are several methods of measuring horizontal, vertical and torsional nystagmus for the diagnosis of vertigo. Using Frentzel glasses, electronystagmography (ENG) and video-oculography (VOG) are common methods for observing nystagmus. In this study, a high-speed video camera is used to analyse high-speed eye movements, including three components of nystagmus, in order to monitor precisely the amplitude and velocity of the quick phase of nystagmus. To confirm the accuracy of the analysis, images of optokinetic nystagmus (OKN) evoked by a stimulating device were analysed. We obtained a high linearity with regard to the relation between the horizontal nystagmus amplitude detected with the high-speed video system and ENG (R2= 0.99). On comparison of the maximum velocities detected with 30Hz and with 250 Hz at each amplitude, the velocities calculated at 30Hz were smaller than those calculated at 250 Hz. At an amplitude of 5 degrees , the 30 Hz velocity decreased by 50%, while a decrease of 35% was observed at 15 degrees. The vertical nystagmus data obtained using high-speed VOG was different in appearance to that obtained using ENG. The torsional component of nystagmus was also measured, and was recorded in the form of a large number of sampling points.

Retinal micromovements, the visual line, and Donders' law

PURPOSE

To report the relationship of the retinal micromovements to the visual line and to confirm the validity of Donders' Law.

METHODS

Two video cameras suspended from a headband were used to record eye (video-oculography) and head movements. Eye positions in held gaze and following various trajectories to a target were recorded in five normal, young subjects. The videotapes were analyzed off-line using a computer algorithm.

RESULTS

Retinal micromovements cause the visual line to trace a zigzag pathway across the foveola, which has an approximate diameter of 350 microm (about 2 degrees). The mean micromovement was about 10 microm in 33.3 msec. The cumulative effect of successive micromovements may move the visual line across the foveola from edge to edge depending on the elapsed time. When the visual line reaches the edge of the foveola it changes its direction. When the eye resets to the same target by different trajectories, the visual line may alight up to about 350 microm from its original location anywhere within the foveola.

CONCLUSIONS

Donders' Law is upheld because for each direction of gaze, and regardless of the trajectory used to reach that direction of gaze, the retina has a constant orientation to an index head plane at any given moment in time. Failure to consider that the micromovements cause a shift in the position of the visual line within the foveola may account for the exceptions to Donders' Law found by contemporary researchers using invasive recording techniques.

Evaluation of a video tracking device for measurement of horizontal and vertical eye rotations during locomotion

We have evaluated a video-based method for measuring binocular horizontal and vertical eye movements of human subjects by comparing it with the magnetic search coil technique. This video tracking system (VTS) uses multiple infrared light sources and small video cameras to simultaneously measure the positions of reflected corneal images and the center of the pupil. The system has a linear range of approximately +/- 40 degrees horizontally and +/- 30 degrees vertically, a sampling rate of 120 Hz (180 Hz with the head fixed), and system noise with standard deviation of < 0.04 degree. The binocular eye-tracking system is light-weight (190 g), being mounted on goggles that, with the eyes in primary position, permit a field of view of 60 degrees horizontally and vertically. The VTS is insensitive to translations of the tracker relative to the eyes. By placing the video preprocessing unit on a cart, eye movements may be recorded while subjects walk through distances up to 100 feet. In comparison with the magnetic search coil technique, the VTS generally provides reliable measurements of horizontal and vertical eye position; eye velocity is noisier than corresponding coil signals, but superior to electro-oculography.

Three-dimensional eye movement analysis in benign paroxysmal positioning vertigo and nystagmus

Benign paroxysmal positioning vertigo (BPPV) and nystagmus (BPPN) is the commonest type of rotational vertigo, but its origin is not fully understood. In this study we used search coils to measure 3-D eye position in 5 patients with BPPN. We present for the first time a complete three-dimensional description of the eye movements induced in BPPN. By calculating the eye movement direction in head co-ordinates we are able to relate the elicited nystagmus very precisely to a particular canal and thereby localize the lesion.

Three-dimensional (3-D) eye movement analysis in patients with positioning nystagmus

Benign paroxysmal positioning vertigo (BPPV), and nystagmus (BPPN) is the commonest type of rotational vertigo. Typical BPPN is generally believed to arise from one posterior semicircular canal. If this is true, the syndrome would offer the unique possibility to study vestibular responses when just one single semicircular canal is excited. In this study we used search coils to measure 3-D eye positions in 3 patients with BPPN. We present a complete 3-D description of the eye movements induced in BPPN. We found that in our patients the eyes rotate rather precisely in the plane of one posterior semicircular canal, suggesting that BPPN in our cases is indeed solely produced by the posterior semicircular canal.