Appendix A: Eye-Movement Recording Systems and Criteria

Nystagmus in infancy and childhood outlines the understanding, evaluation, and treatments of nystagmus in infancy and childhood. Aligning this condition with advanced concepts of developmental brain-eye diseases and summarizing novel treatment paradigms, the authors provide an authoritative resource for both clinicians and scientists in the care of infants and children with nystagmus. The chapters comprised here offer valuable coverage in all relevant areas related to nystagmus: algorithms for examination; descriptions of diagnostic techniques; medical, surgical, and alternative treatments of the visual system in infants and children; methodologies for investigation, including analysis software, models of the ocular motor system, and current hypotheses on the pathophysiology of ocular motor oscillations. Unlike earlier works on this topic, emphasis is placed on the motor mechanisms that cause the various types of nystagmus rather than the diagnosis or treatment of the afferent visual deficits that may accompany them. The study of each type of nystagmus using accurate eye-movement recordings serves as the foundation for differential diagnosis and treatment options. Each chapter summarizes the results of ocular motor research in a narrative manner, identifying the important ideas and observations that point to underlying neurophysiological mechanisms. Based on insights from the authors' combined 75 years of clinical experience, Nystagmus in Infancy and Childhood is a valuable clinical reference for ophthalmologists, neurologists, and other specialists in the treatment of this condition.

OUP UNCORRECTED PROOF – REVISES, 09/06/12, NEWGEN appendix a eye-movement recording systems and criteria A.1 RECORDING METHODS 263 A.1.1 Contact Electrooculography 263 A.1.2 Infrared Reflection 264 A.1.3 Scleral Search Coil 265 A.1.4 High-Speed Video Oculography 266 A.2 RESEARCH CRITERIA 268 A.3 CLINICAL CRITERIA 268 A.4 CALIBRATION TECHNIQUES 268 A.4.1 Adults and Children 268 A.4.1.1 Infantile Nystagmus 269 A.4.1.2 Fusion Maldevelopment Nystagmus 269 A.4.2 Infants 269 THE STUDY of ocular motor mechanisms (normal or abnormal) for the diagnosis and treatment of ocular motor disorders requires accurate recording of eye movements. Th is includes a reliable, easy-to-use recording system; accurate monocular calibration and linearization paradigms; and, in subjects with nystagmus or other ocular motor oscillations, knowledge of the foveation portions of the eye-movement waveforms. We describe the historical development of eye-movement recording systems, the advantages and disadvantages of those commonly used today, the requirements for accurate calibration of those systems, and the use of eye-movement recordings in basic research, including clinical sett ings. Despite the fact that human hearing is inferior to that of the owl, and the human sense of smell is far poorer than a dog’s, our visual acuity is excelled by few other species. However, if we processed our entire visual field simultaneously at our maximal resolution, we would need so many optic nerve fibers to carry visual information back to the brain that there would be room for litt le else. Evolution has solved this problem by making the resolution of the retina—the light-sensitive neural layer of the eye—inhomogeneous. Visual acuity in the central 1° of the visual field is maximal, but it falls off rapidly as one moves toward the periphery. What keeps us from ever being aware of this is the nearly incessant motion of our eyes, controlled by the interconnected control systems that direct our gaze to an object of interest and keep it fi xated in the face of target and body movement. Massive processing in the visual areas of the brain integrates the discontinuous flow of visual images along with efference copy of motor commands to the extraocular muscles, into the clear, stable perception of the world experienced by both normals and those with some ocular motor oscillations. What follows are descriptions of the more common eye-movement recording technologies used in the past and present. Technical descriptions, engineering, and physics of these and other methods will not be discussed in this volume (see also Abel and Dell’Osso1). Rather, emphasis will be on the abilities of different types of systems and the calibration requirements to provide accurate eye-movement data in both the basic and clinical research sett ings. Historically, the instrumentation for recording all types of eye movements was used • 10_Hertle_Appendix_A.indd 259 259 9/6/2012 9:46:00 PM OUP UNCORRECTED PROOF – REVISES, 09/06/12, NEWGEN originally to record vestibular nystagmus. Purkinje noted eye movements by visual observation in 1825 and E. Darwin by palpations of the eyes in 1794.2 Studies of eye movements by visual observation were described by Java1 in 1879. 3 The experimenter stood behind the subject and observed the movement of the eyes in a mirror. Specially designed optical instruments were used to magnify the mirrored images that then could be studied in detail. More accurate descriptions, based on the observation of afterimages, were also made at the end of the eighteenth century. Using this method, Wells described the slow and fast phases of vestibular nystagmus.4 The occurrence of saccades during reading was fi rst reported by Javal and Lamare, who used a rubber tube connected to the conjunctiva and both ears. 3 With this device, each eye movement caused a sound that was heard. Hering used a similar acoustic device in combination with the technique of afterimages. 5 The earliest mechanical methods of recording eye movements were proposed by Raehlmann in 1878, who used one end of a lever attached to the globe and with the other end of the lever, recorded the transmitted eye motions, on a moving smoked drum (Fig. A.1).2–4,6,7 The lever technique was modified by Gradenigo in l909, by Buys and Coppez in 1909, and Ohm in 1914. In their studies, one polished end of the lever touched the anesthetized cornea, while the other end of the moving lever made the record on a moving paper FIGURE A.1 Lever device to document horizontal eye movements during reading. 260 tape or “kymograph”8 (Fig. A.2). The technique further improved when small cups resembling contact lenses were attached to the cornea. In 1899, Orschansky fi xed a small mirror to the cup on the eye and used a beam of light to project the reflected eye movements onto a screen.9,10 The method of recording eye movement by reflected light was further advanced with the use of special contact lenses. Unfortunately, this method could injure the eye and was too heavy to measure the large accelerations occurring during saccades. To overcome this problem, Javal recorded the reflection of a light beam from a litt le mirror attached to the conjunctiva, a method that was not successfully applied before von Romberg and Ohm used it to measure ocular torsion. Th is technique was, however, still too invasive.11 Photographic analysis of nystagmus was introduced by Dodge and Cline in 1901 and FIGURE A.2 Model of Kymograph of the type used to record eye movements. Time recordings of physiological phenomena became possible when a galvanometer’s needle was put into contact with a loop of paper coated with a thin layer of smoke black and stretched over a metal drum. A precise mechanical clockwork allows the drum to be rotated at a calibrated speed, and the movements of the galvanometer’s pen scratch out the smoke black, leaving a record of amplitude as a function of time. Mechanical phenomena, such as eye movements, could be recorded along time, by rigging levers, axes, membranes, springs, and strings. The time axis was calibrated and measured by using electromagnetic tuning forks att ached to inscribing pens. • EY E-MOV EM ENT R ECOR DI NG S YSTEMS A N D CR ITER I A 10_Hertle_Appendix_A.indd 260 9/6/2012 9:46:01 PM OUP UNCORRECTED PROOF – REVISES, 09/06/12, NEWGEN 1903. 3 In 1913 Coppez used early cinematography. Electrical recording of eye movements was fi rst reported by Schott in 1922 by modifying electrocardiography.12 Schott and Meyers measured electrical potentials with skin electrodes attached near the eye.12 Mowrer et al. discovered that the electrical potential is primarily caused by the electrical dipole between cornea and retina, which moves with the eye. Th is technique has been the basis of modern electronystagmography (ENG) or electrooculography (EOG)13 (Fig. A.3). The eyes are the origin of a steady electric potential field, which can also be detected in total darkness and if the eyes are closed. The electric signal that can be derived using two pairs of contact electrodes placed on the skin around each eye with a ground or reference electrode placed on the forehead. Jung applied this method to record horizontal and vertical components of the eye position simultaneously14 (Fig. A.4). Previously, recording techniques had been restricted to one movement direction only. Moreover, the EOG allows recording of eye movements while the eyes are closed, of particular interest for sleep research. Noncontact optical methods are currently the most popular. The use of infrared light reflected (IR) from the eye, which is sensed by specially designed optical sensors remains common. A voltage is generated from the difference in reflection between the sclera and iris as the eye moves and is the basic output to extract eye rotation information (Fig. A.5). These IR devices measure the intensity of these reflections by photosensitive elements placed at different locations in front the eye. The fi rst system was developed by Torok et al.15 Video-based eye trackers typically use one or multiple Purkinje images and the center of the pupil as features to track eye movement over time. These optical methods, particularly those based on video recording, are now widely used and are favored for eye-movement analysis. They are especially useful in infants and children, being noninvasive and inexpensive (Fig. A.6). These so-called double Purkinje image (DPI) eye trackers reach high resolution, accuracy, and bandwidth. The high accuracy of the DPI eye tracker during steady fi xation is due to the fact that it uses the angular differences between light reflections that are insensitive to small translations between the eye and the tracker. Videooculography (VOG), defi ned as the use of these methods for dynamic measure of eye movements, became feasible with the rapid development computer-based automatic image processing. Th is progress is mainly reflected in the frame rates being processed online and in the robustness and the accuracy of the marker detection algorithms. Both improve with the increase in computational power. Since the measurement of two-dimensional gaze direction in VOG is primarily based on the localization of the pupil, the two-dimensional VOG + + – – FIGURE A.3 The eye acts as a dipole in which the anterior pole is positive and the posterior pole is negative (arrow). The cornea (relative positive charge) approaches one canthal electrode while moving away from the other canthal electrode to which the retina moves near (relative negative charge), resulting in recordable changes in the potentials between the two electrodes. FIGURE A.4 Electrode placement for contact oculography. Pairs of electrodes are placed to the left and right and top and bottom of each eye. If the eye is moved from the center position toward one electrode, this electrode “sees” the positive side (the cornea) and the opposite electrode “sees” the negative side (the retina). Consequently, a potential difference occurs between the electrodes. Nystagmus in Infancy and Childhood • 261 10_Hertle_Appendix_A.indd 261 9/6/2012 9:46:02 PM OUP UNCORRECTED PROOF – REVISES, 09/06/12, NEWGEN IR Emitter IR Detector FIGURE A.5 Principle of infrared-reflectance (IR), eye-movement recording. A voltage is generated as the eye moves in front of an infrared detector due to the difference in reflectance between the sclera and iris as the eye moves. Th is signal from the detector as a result of the measured difference in reflectance is amplified and fi ltered to produce the eye-movement recording. Standard push-pullconnected detectors (top panel) and single detector (bottom panel). works reliably in head-mounted systems and with stabilized head positions. To compute the three-dimensional eye position, the orientation of the iris signature can be used. Th is signature must be scanned along a circular path close to the limbus, in order to be insensitive to changes of the pupil diameter. Direct polar cross-correlation of the iris signature at the actual eye position with that of a reference position can be used to measure ocular torsion. Th is works well while gaze is pointing straight ahead, but geometric distortions of the iris occurring at eccentric gaze positions lead to large errors. None of the recording methods mentioned thus far are able to quantify horizontal, vertical, 262 FIGURE A.6 Remote video-based, eye-movement recording system. Th is method is based on tracking of the position of eye-fi xed markers in a two-dimensional image. and torsional eye movements simultaneously. Vertical and horizontal movement components could be quantified by the EOG, IR, or the DPI tracker, but these devices cannot measure ocular torsion. Von Romberg and Ohm measured pure ocular torsion in primary position with their mirror system.16 By the nineteenth century, the technique of afterimages had provided important fi ndings about ocular torsion during fi xation. Th is field of research became of increasing interest when the magnetic search-coil technique, developed by Robinson and Collewijn et al., was extended by Collewijn et al. and Kasper and Hess to cover three-dimensional movements.17,18 The method is based on the voltages induced in coils by two or three orthogonal, rapidly alternating magnetic fields. The coils are embedded in a soft silicone annulus that adheres elastically to the eyeball. One coil is sufficient to measure gaze direction (Fig. A.7). Two coils with different orientations must be molded in the annulus to measure gaze direction and ocular torsion simultaneously. The search-coil method combines high spatial and temporal resolution and is so far the most precise method for measuring ocular torsion. Like other methods based on contact lenses, the search-coil technique has the main disadvantage of being invasive (Fig. A.8). Therefore, considerable effort was made to evaluate the • EY E-MOV EM ENT R ECOR DI NG S YSTEMS A N D CR ITER I A 10_Hertle_Appendix_A.indd 262 9/6/2012 9:46:02 PM OUP UNCORRECTED PROOF – REVISES, 09/06/12, NEWGEN FIGURE A.7 Scleral search coil. The coils are embedded in a soft silicone annulus that adheres elastically to the eyeball. three-dimensional eye position on the basis of photographic images of the eye. All photographic methods are based on the detection and localization of eye-fi xed markers (pupil, limbus, iris signatures, episcleral blood vessels) in image coordinates. The eye position with respect to the head can be computed from these image coordinates if the camera is fi rmly attached to the head. Otherwise, head-fi xed markers can be used to compensate for relative translations between head and camera. Howard and Evans described a method for computing three-dimensional angular eye positions from the image coordinates of the markers.19 A.1 RECORDING METHODS A.1.1 Contact Electrooculography The simplest method for measuring human eye movements is based on the feature that the human eye is an electrical dipole. The axis of this dipole and the optical axis of the human eye are roughly collinear. The retina is more negative than the cornea. The potential difference of about 0.4–1.0 millivolts results from the electrical activity of photoreceptors and neurons in the retina. Changes of this potential induced by sudden light stimulus have been used for decades to monitor the electrical activity of the retina (electroretinogram [ERG]). However, the EOG measures the eye dipole as it rotates. Th is causes FIGURE A.8 Scleral search coil. The silicone annulus rests on the eye at the corneo-scleral limbus on the conjunctiva similar to the older “scleral” contact lenses. small differences between the electrical potential at the skin surface next to the eye depending on eye position. A rightward eye movement will increase the surface potential at the temporal canthus and decrease it at the nasal canthus of the right eye (Fig. A.4). The potential differences can be measured with a contact electrode configuration. The voltages are usually referenced to a third electrode that is generally placed at the forehead or one of the mastoid processes or on the earlobe.2,20–22 To simultaneously record vertical eye movements, two additional electrodes must be placed below and above the eye. Vertical EOG signals are less reliable than horizontal signals due to lid artifacts. The resolution of both horizontal and vertical EOG signals is limited by electromagnetic field noise in the environment, thermal noise generated by the input resistance of the amplifier and the contact resistance of the skin electrodes, and capacitive noise due to electrical activity of muscles and neurons.2,20–22 To lower the contact resistance, the skin should be cleaned with alcohol. Electrodes should be made of relatively nonpolarizeable material such as silver-silver chloride or gold and applied with a conductive paste. Subjects should be instructed to avoid any movements except eye movements. Changes of the dark adaptation level induce slow drift s of the corneo-retinal potential that are superimposed Nystagmus in Infancy and Childhood • 263 10_Hertle_Appendix_A.indd 263 9/6/2012 9:46:03 PM OUP UNCORRECTED PROOF – REVISES, 09/06/12, NEWGEN on the EOG signal. Since both the EOG and ERG measure the corneo-retinal potential, the standards of ERG recordings are also recommended for EOG recordings. The spatial resolution of EOG is ~1°, temporal resolution ~40 Hz, vertical recording is confounded by blink artifact, noise is 1° or more, setup is slow, calibration is needed, and cost is ~$500.00. A.1.2 Infrared Reflection These devices measure the differences in intensity of infrared light reflected from across the surface of the eye at a fi xed location from the eye.2,23–26 Light intensity is measured with photo diodes that have a high temporal resolution (Figs. A.9 and A.10). The distance between eye and photoreceptors is in the range of 24 mm. At such small distances, the differences in the intensity between the different photodiodes depend mainly on the position of iris and pupil, which reflect less light than the surrounding sclera. IR devices are very sensitive to relative translations of the photodiodes and the eye because they do not evaluate the angle, but only the intensity of the reflection. For an eye radius of 1.25 cm, a translational error of 1 mm will lead to an eye-position error of almost 5°. The system must therefore be fi rmly attached to the head. IR devices have a much lower noise level than EOG, but they suffer from eyelid artifacts that critically depend on the position of the photodiodes. These lid artifacts may increase dramatically if FIGURE A.9 Infrared reflectance eye movement recording system embedded in a goggle system. Th is shows the combination of vertical and horizontal detectors for each eye. 264 FIGURE A.10 Infrared reflectance eye-movement recording system embedded in a goggle system specially designed for use in infants. the device is not properly adjusted in front of the eye. Lid artifacts are more pronounced for vertical than for horizontal eye movements. Moreover, the position of the photodiodes is also critical for the system linearity. Due to these features, optimal adjustment of the device requires that the experimenter carefully controls the eyeposition signal of the IR device and compares it with the eye movements (Figs. A.11 and A.12). The spatial resolution is ~0.1°, temporal resolution is 100–500 Hz, and vertical recording is confounded by blink artifact and the intrinsic difficulty in distinguishing lid movement from eye movement. Setup is fast but calibration is necessary. Linearity is a problem with nonlinearity occurring at 15°–20° and cost is moderate, ~ $4000. FIGURE A.11 Children positioned in head and chin rest with infrared goggles in place and a viewing stimulus screen for accurate calibration (side view). • EY E-MOV EM ENT R ECOR DI NG S YSTEMS A N D CR ITER I A 10_Hertle_Appendix_A.indd 264 9/6/2012 9:46:04 PM OUP UNCORRECTED PROOF – REVISES, 09/06/12, NEWGEN FIGURE A.12 Children positioned in head and chin rest with infrared goggles in place and a viewing stimulus screen for accurate calibration (front view). IR systems are second only to EOG in their range of applications. Because they can resolve fi ne detail with low noise, they are excellent for conditions where subtle features of the eye movement are important. IR systems were the key to the accurate analyses of saccadic trajectories and the analysis of small corrective saccades within nystagmus waveforms. Also, being noninvasive, they provided a major advantage for the study of patients and children. A.1.3 Scleral Search Coil The scleral search-coil system measures the voltages in one or two coils induced by two or three rapidly oscillating magnetic fields2,17,18,27–31 (Fig. A.13). The coils are molded in a soft contact annulus that is att ached to the eyeball. Th ree pairs of large coils, mounted in a cubic frame, generate the magnetic fields. The subject’s head is positioned at its center. The field coils should be large, because the homogeneity of the magnetic field is crucial for the precision of the measurement. With pairs of squareshaped coils, arranged in a cubic configuration, the inhomogeneity inside of a central test cube stays below 5% when the edge length of the test cube approaches one-fi ft h of the edge length of the field coil. Th is means that when using field coils with an edge length of 1.5 m, subjects should not move by more than 7 cm. FIGURE A.13 Th ree-foot scleral search-coil, magnetic field system with dichoptic stimuli apparatus (monocular stimulus screens and mirrors). Around the head of the subject an alternating horizontal and vertical magnetic field (spatially and temporally in quadrature) is generated and consequently an alternating voltage will be induced in the scleral contact lens with embedded coil. The voltage induced by one of the magnetic fields in the scleral search coil is proportional to the projection of the coil vector (defi ned as the vector orthogonal to the effective coil plane) onto the magnetic field vector. Thus, the three voltages induced by three orthogonal magnetic fields form the vector components of the coil vector expressed in field coordinates. A dual search coil for recording three-dimensional eye orientation provides six voltages, corresponding to the two three-dimensional coil vectors of the directional and the torsional coil (Fig. A.14). Methods to compute the three-dimensional eye orientations from these six signals are then FIGURE A.14 Torsional coil on eye. The principle of the scleral search-coil technique is based upon the magnetic induction of a small coil embedded in a flexible ring of silicone rubber, which adheres to the limbus of the human eye concentric with the cornea. Nystagmus in Infancy and Childhood • 265 10_Hertle_Appendix_A.indd 265 9/6/2012 9:46:05 PM OUP UNCORRECTED PROOF – REVISES, 09/06/12, NEWGEN employed. Search-coil recordings require a calibration procedure that is often based on multiple alignments to targets at various positions. The calibration parameters are computed by minimizing the errors between the calibrated gaze vector and target vectors. 2,17,18,27–31 Systems with three magnetic fields can be objectively calibrated, that is, their calibration does not rely on accurate fi xation of targets at different positions, as most other recording techniques. Only a single fi xation target is needed in order to determine the orientation of the coil with respect to the eye. Another important advantage of threefield systems over two-field systems is that the orientation of the coil vector can be determined without knowledge of the actual inductance of the scleral search coil. Th is is because changes in inductance will have the same proportional effect on all three voltages and can easily be eliminated by normalization. 2,17,18,27–31 With the search-coil technique, the inherent system noise of horizontal and vertical eye position has been estimated to be on the order of 0.5 min of arc (0.0083°). The system resolution is a very important parameter; it determines the smallest eye movement that can be detected. However, to compare the metrics of eye movements between different subjects or with a stimulus-defi ned requirement the accuracy is more important than the system noise. The system accuracy of search coils depends mainly on the quality of the calibration. Due to its large signal-to-noise ratio and reliability, the search-coil technique has been the generally accepted reference standard for eye movement recordings for 30 years. However, the disadvantages, connected with the invasiveness of the method, have also been recognized. The search coil not only measures eye movements but also affects them. Some authors have found that saccades last longer (by about 8%) and become slower (by about 5%) when subjects wear search coils in both eyes than when they do not. 32,33 It was also shown that the eye torsion, when evaluated with the search coil, depends on the orientation of exit point of the connecting line from the search coil and, with the nasal exiting orientation of a commercial eye coil (Skalar), ocular torsion depended more on eye elevation than 266 with a modified exit point that minimized the contact between wire and eyelids. Other disadvantages of the scleral search coil are that wearing the coil may lead to drying, and temporal deformations of the cornea, and reduced visual acuity in the eye with the search coil. Therefore, the manufacturer of the search coil limits wearing time to 30 minutes. The coil spatial resolution is ~0.01°; temporal resolution is at least 1000 Hz. Vertical and torsional recordings are also possible and linearity is good, although setup is slow and calibration is needed. A reasonable coil system can be bought for about $15,000; each eye coil costs about $100. A typical eye coil lasts for two subjects. There is a small risk of a corneal abrasion from the contact lens. The estimate of the risk is about 1/400 subjects. There is also a small risk of transmitt ing very serious diseases if the lab reuses lens between patients. Certain biologic agents (prions—such as found in “Mad Cow”) are very difficult to kill. Th is risk can be avoided if the lab simply uses a new scleral contact lens rather than “recycling” them. Only about 30 minutes of continuous recording is usually possible at one sett ing. Eyecoil systems are usually research tools. To use an eye-coil system subjects must sign a consent form because of the risk of corneal abrasion. Th is technology is certainly the most expensive of all, because of the cost of the eye coils. A.1.4 High-Speed Video Oculography Video-based, eye-movement recordings have become more and more popular because of the rapid progress made in electronic data processing.21,24,34–39 High-speed video oculography (VOG) has become affordable, the robustness of the algorithms improved, and the range of applications expanded (Fig. A.15). Nowadays, commercial companies produce VOG devices that can be used in a functional magnetic resonance imaging scanner (MeyeTrack R , SMI R , Berlin, Germany). Most fundamental VOG techniques are based on tracking of the position of eye-fi xed markers in a two-dimensional image. These positions have to be expressed in head-fi xed coordinates. Since head-fi xed markers are difficult to obtain with high precision, one • EY E-MOV EM ENT R ECOR DI NG S YSTEMS A N D CR ITER I A 10_Hertle_Appendix_A.indd 266 9/6/2012 9:46:10 PM OUP UNCORRECTED PROOF – REVISES, 09/06/12, NEWGEN FIGURE A.15 Video-based, eye-movement system showing programming computer interface displaying calibration parameters and actual pupil location within the camera system. strategy of VOG systems is to attach the video camera as fi rmly as possible to the head. As long as the system is not compensated for relative translation between camera and head, the accuracy of VOG has a problem very similar to that of IR. A translation of 1 mm will result in an error of about 5°. Head-fi xed devices cause a problem under head free conditions, because the stability of the head mount is not sufficient. Because of this problem, actual VOG systems can make highly accurate measurements of eye position, only as long as the head is stable in space. A VOG method of compensating for head translation uses the relative position of the corneal reflex of an infrared LED (Eyelink II R , SR Research, Osgode, Canada) (Fig. A.16). One difficulty with this method is that using the corneal reflection adds more noise. For eye movements (a) (b) FIGURE A.16 Head-mounted eye trackers: (a) SMI EyeLink I and (b) Arrington Research ViewPoint PC-60 BS007. The shutter-glasses have been att ached to the head-mount and the cameras are recording the eyes from below. of about 12°–15° the reflection reaches the edge of the cornea and can no longer be used for compensation. Moreover, this approach relies on the topography of the cornea, which varies between subjects. Therefore, it seems to be useful when compensating for large translations, but it may be unable to provide very high accuracy. Since the pupil position is detected and evaluated in image coordinates, the nonlinearity of the VOG (in contrast to IR) systems is well defi ned by the geometry of the image projection. With parallel projection, the angular eccentricity of the eye can be approximated by the inverse sinus of the ratio of the eccentricity of the pupil center and the eye radius, both expressed in image coordinates. The main aim of the VOG calibration is therefore to determine the location of the center of rotation of the eye and the radius of the eyeball. The resolution of the two-dimensional VOG defi ned by the standard deviation of system noise measured with an artificial eye is about 0.01°. The VOG of two-dimensional measurements of ocular torsion also reach accuracy values that are similar to those of coil measurements. The spatial resolution is 1 part in 1024 and now with highspeed cameras, temporal resolution is as high as 1000 Hz. VOG can record vertical and torsion movements.21,24,34–39 Setup is rapid and calibration may be less difficult. Goggles effectively black out vision, so maintaining a “light-tight” lab is not crucial; this may be important in some applications and for some systems. Systems cost from $18, 000 to $50,000. As Figure A.17 illustrates, in addition to the study of humans with nystagmus, we have used both IR and video systems successfully to study canine eye movements in both normal dogs and those with nystagmus. The dog on the left was an achiasmatic Belgian sheepdog with infantile and seesaw nystagmus; the dog on the right was a Briard with Leber congenital amaurosis; and the dog at the bottom was a normal Brittany, highly trained for upland bird hunting. Eizenman and his students have developed new algorithms for video-based systems that may make patient-cooperative calibration unnecessary.40,41, 42 Th is promises to be a major improvement in the ability to obtain accurate eyemovement data from infants and uncooperative Nystagmus in Infancy and Childhood • 267 10_Hertle_Appendix_A.indd 267 9/6/2012 9:46:10 PM OUP UNCORRECTED PROOF – REVISES, 09/06/12, NEWGEN FIGURE A.17 Use of infrared (eyeglass-frame type, top left panel and goggle type, top right panel) and high-speed digital video (bottom panel) systems to study canine eye movements. children in the clinic. Its accuracy in patients with nystagmus is yet to be tested. A.2 RESEARCH CRITERIA At present, IR, search-coil, and digital video systems meet the criteria for accurate eye-movement data. Each has its own set of advantages and limitations, and a well-equipped laboratory usually has several of the systems available that can be tailored to each study. In all systems, accurate, monocular calibration and zeroing is a necessity. A.3 CLINICAL CRITERIA In the clinic or clinical laboratory, IR, searchcoil, and digital video systems have also been used, but the search-coil system is usually too 268 invasive for studying most patients and children. Accurate differential diagnoses of nystagmus and detection of fi xating-eye changes due to strabismus require accurate, monocular calibration and zeroing. A.4 CALIBRATION TECHNIQUES A.4.1 Adults and Children For both adults and children each eye must be calibrated at several target positions while the other eye is occluded; for search coils, each eye must be zeroed while the other is occluded (each eye coil will have been precalibrated on a protractor jig before inserting it into the eye). The calibration and zeroing values can then be used post hoc on the data collected to ensure accuracy and linearity. • EY E-MOV EM ENT R ECOR DI NG S YSTEMS A N D CR ITER I A 10_Hertle_Appendix_A.indd 268 9/6/2012 9:46:12 PM OUP UNCORRECTED PROOF – REVISES, 09/06/12, NEWGEN A . 4 .1.1 I N FA N T I L E N Y S TA G M US For infantile nystagmus syndrome (INS) patients, monocular calibration allows analysis of the fi xating eye for foveation quality and detection of microstrabismus, both static and time varying. Application of foveation-quality formulae like the NAFX require that the data being analyzed come from the fi xating eye only; data from a deviated eye would be meaningless in the determination of bestpossible acuity. A . 4 .1. 2 F US I O N M A L D E V E L O P M E N T N Y S TA G M US For fusion maldevelopment nystagmus (FMNS) patients, monocular calibration is necessary for the same reasons as for INS patients. In these patients, switching from one fi xating eye to the other varies with both gaze angle and time; without an accurate method of determining which eye is fi xating at any particular instant, analysis would become hopelessly confounded. A.4.2 Infants For infants with either INS or FMNS, the difficulties in calibration may, in some cases, preclude acquiring accurate data. This is usually a problem in children between the ages of 2 to 4 years old who may be uncooperative. For this group, a noninvasive, self-calibrating system would be invaluable (see earlier discussion of recently developed soft ware algorithms that may solve this problem). REFERENCES 1. Abel LA, Dell’Osso LF. Ocular motility recording and nystagmus. In: Webster J, ed. Encyclopedia of Medical Devices and Instrumentation. 2nd ed. Hoboken, NJ: John Wiley and Sons; 2006:137–149. 2. Eggert T. Eye movement recordings: methods. Dev Ophthalmol 2007;40:15–34. 3. Wade NJ, Tatler BW, Heller D. Dodge-ing the issue: Dodge, Javal, Hering, and the measurement of saccades in eye-movement research. Perception 2003;32(7):793–804. 4. Wade NJ. William Charles Wells (1757–1817) and vestibular research before Purkinje and Flourens. J Vestib Res 2000;10(3):127–137. 5. Kennedy EA, Bonivtch AR, Manoogian SJ, Stitzel JD, Herring IP, Duma SM. The effects of extraocular muscles on static displacements of the human eye. Biomed Sci Instrum 2006;42:372–377. 6. Wade NJ. Porterfield and Wells on the motions of our eyes. Perception 2000;29(2):221–239. 7. Huey E. Preliminary experiments in the physiology and psychology of reading. Am J Psychol 1898;9:575–586. 8. De Leo A. The origin of graphic recording of psycho-physiological phenomena in Germany. Physis Riv Int Stor Sci 2006;43(1–2):345–362. 9. Hoff HE, Geddes LA, Guillemin R. The anemograph of Ons-en-Bray: an early self-registering predecessor of the kymograph with translations of original description and a biography of the inventor. J Hist Med Allied Sci 1957;12(4):424–448. 10. Ono H. On Wells’s (1792) law of visual direction. Percept Psychophys 1981;30(4):403–406. 11. Aust W. Der Nachbildablauf bei Schielamblyopen am Ohm-Rombergschen Gerat. [The afterimage process in squint amblyopia with the Ohm-Romberg apparatus]. Albrecht Von Graefes Arch Klin Exp Ophthalmol 1966;170(2):175–200. 12. Hobson JA, Pace-Schott EF, Stickgold R, Kahn D. To dream or not to dream? Relevant data from new neuroimaging and electrophysiological studies. Curr Opin Neurobiol 1998;8(2):239–244. 13. Smith CR. Measurement of nystagmus using electronystagmography (ENG). J Speech Hear Disord 1967;32(2):133–138. 14. Jung R. Cerebral control of eye movements and motion perception. Introduction. Bibl Ophthalmol 1972;82:1–6. 15. Torok N. Measuring eye movements. Science 1960;131(3404):940. 16. Stormer A. Ernst von Romberg; zu seinem 100. Geburtstag am 5. November 1965. [Ernst von Romberg; on the 100th anniversary of his birth on November 5, 1965]. Munch Med Wochenschr 1965;107(53):2683–2686. 17. Robinson DA. A method of measuring eye movement using a scleral search coil in a magnetic field. IEEE Trans Biomed Eng 1963;10:137–145. 18. Ferman L, Collewijn H, Van den Berg AV. A direct test of Listing’s law—II. Human ocular torsion measured under dynamic conditions. Vision Res 1987;27(6):939–951. 19. Howard IP, Evans JA. The measurement of eye torsion. Vision Res 1963;61:447–455. Nystagmus in Infancy and Childhood • 269 10_Hertle_Appendix_A.indd 269 9/6/2012 9:46:15 PM OUP UNCORRECTED PROOF – REVISES, 09/06/12, NEWGEN 20. Hoff HE, Geddes LA. Graphic registration before Ludwig; the antecedents of the kymograph. Isis 1959;50(159):5–21. 21. Henn V, Straumann D. Th ree-dimensional eye movement recording for clinical application. J Vestib Res 1999;9(3):157–162. 22. Rubin AM, Zafar SS. The assessment and management of the dizzy patient. Otolaryngol Clin North Am 2002;35(2):255–273. 23. Ciuff reda KJ, Bahill AT, Kenyon RV, Stark L. Eye movements during reading: case reports. Am J Optom Physiol Opt 1976;53(8):389–395. 24. Ukai K, Saida S, Ishikawa N. Use of infrared TV cameras built into head-mounted display to measure torsional eye movements. Jpn J Ophthalmol 2001;45(1):5–12. 25. Traisk F, Bolzani R, Ygge J. A comparison between the magnetic scleral search coil and infrared reflection methods for saccadic eye movement analysis. Graefes Arch Clin Exp Ophthalmol 2005;243(8):791–797. 26. De Luca M, Spinelli D, Zoccolott i P, Zeri F. Measuring fi xation disparity with infrared eye-trackers. J Biomed Opt 2009;14(1):014013. 27. Ferman L, Collewijn H, Jansen TC, Van den Berg AV. Human gaze stability in the horizontal, vertical and torsional direction during voluntary head movements, evaluated with a three-dimensional scleral induction coil technique. Vision Res 1987;27(5):811–828. 28. Quinn KJ, Rude SA, Brett ler SC, Baker JF. Chronic recording of the vestibulo-ocular reflex in the restrained rat using a permanently implanted scleral search coil. J Neurosci Methods 1998;80(2):201–208. 29. Stahl JS, van Alphen AM, De Zeeuw CI. A comparison of video and magnetic search coil recordings of mouse eye movements. J Neurosci Methods 2000;99(1–2):101–110. 30. Foeller P, Tychsen L. Eye movement training and recording in alert macaque monkeys: 1. Operant visual conditioning; 2. Magnetic search coil and head restraint surgical implantation; 3. Calibration and recording. Strabismus 2002;10(1):5–22. 31. Houben MM, Goumans J, van der Steen J. Recording three-dimensional eye movements: scleral search coils versus video oculography. Invest Ophthalmol Vis Sci 2006;47(1):179–187. 270 • 32. Bergamin O, Zee DS, Roberts DC, Landau K, Lasker AG, Straumann D. Three-dimensional Hess screen test with binocular dual search coils in a three-field magnetic system. Invest Ophthalmol Vis Sci 2001;42(3):660–667. 33. Bergamin O, Ramat S, Straumann D, Zee DS. Influence of orientation of exiting wire of search coil annulus on torsion aft er saccades. Invest Ophthalmol Vis Sci 2004;45(1):131–137. 34. Allum JH, Honegger F, Troescher M. Principles underlying real-time nystagmus analysis of horizontal and vertical eye movements recorded with electro-, infrared-, or video-oculographic techniques. J Vestib Res 1998;8(6):449–463. 35. Eckert AM, Gizzi M. Video-oculography as part of the ENG battery. Br J Audiol 1998;32(6):411–416. 36. Zhu D, Moore ST, Raphan T. Robust pupil center detection using a curvature algorithm. Comput Methods Programs Biomed 1999;59(3):145–157. 37. Geisler C, Bergenius J, Brantberg K. Nystagmus fi ndings in healthy subjects examined with infrared videonystagmoscopy. ORL J Otorhinolaryngol Relat Spec 2000;62(5):266–269. 38. van der Geest JN, Frens MA. Recording eye movements with video-oculography and scleral search coils: a direct comparison of two methods. J Neurosci Methods 2002;114(2):185–195. 39. Schreiber K, Haslwanter T. Improving calibration of 3-D video oculography systems. IEEE Trans Biomed Eng 2004;51(4):676–679. 40. Guestrin ED, Eizenman M. General theory of remote gaze estimation using the pupil center and corneal reflections. IEEE Trans Biomed Eng 2006;53(6):1124–1133. 41. Kang JJ, Eizenman M, Guestrin ED, Eizenman E. Investigation of the cross-ratios method for point-of-gaze estimation. IEEE Trans Biomed Eng 2008;55(9):2293–2302. 42. Model D, Eizenman M. An automatic personal calibration procedure for advanced gaze estimation systems. IEEE Trans Biomed Eng 2010;57(5):1031–1039. EY E-MOV EM ENT R ECOR DI NG S YSTEMS A N D CR ITER I A 10_Hertle_Appendix_A.indd 270 9/6/2012 9:46:16 PM