Changes In Retinal Image Research Articles

MSGANet-RAV: A multiscale guided attention network for artery-vein segmentation and classification from optic disc and retinal images

BackgroundRetinal and optic disc images are used to assess changes in the retinal vasculature. These can be changes associated with diseases such as diabetic retinopathy and glaucoma or induced using ophthalmodynamometry to measure arterial and venous pressure. Key steps toward automating the assessment of these changes are the segmentation and classification of the veins and arteries. However, such segmentation and classification are still required to be manually labelled by experts. Such automated labelling is challenging because of the complex morphology, anatomical variations, alterations due to disease and scarcity of labelled data for algorithm development. We present a deep machine learning solution called the multiscale guided attention network for retinal artery and vein segmentation and classification (MSGANet-RAV). MethodsMSGANet-RAV was developed and tested on 383 colour clinical optic disc images from LEI-CENTRAL, constructed in-house and 40 colour fundus images from the AV-DRIVE public dataset. The datasets have a mean optic disc occupancy per image of 60.6% and 2.18%, respectively. MSGANet-RAV is a U-shaped encoder-decoder network, where the encoder extracts multiscale features, and the decoder includes a sequence of self-attention modules. The self-attention modules explore, guide and incorporate vessel-specific structural and contextual feature information to segment and classify central optic disc and retinal vessel pixels. ResultsMSGANet-RAV achieved a pixel classification accuracy of 93.15%, sensitivity of 92.19%, and specificity of 94.13% on LEI-CENTRAL, outperforming several reference models. It similarly performed highly on AV-DRIVE with an accuracy, sensitivity and specificity of 95.48%, 93.59% and 97.27%, respectively. ConclusionThe results show the efficacy of MSGANet-RAV for identifying central optic disc and retinal arteries and veins. The method can be used in automated systems designed to assess vascular changes in retinal and optic disc images quantitatively.

Mice Preferentially Use Increases in Cerebral Cortex Spiking to Detect Changes in Visual Stimuli.

Whenever the retinal image changes, some neurons in visual cortex increase their rate of firing whereas others decrease their rate of firing. Linking specific sets of neuronal responses with perception and behavior is essential for understanding mechanisms of neural circuit computation. We trained mice of both sexes to perform visual detection tasks and used optogenetic perturbations to increase or decrease neuronal spiking primary visual cortex (V1). Perceptual reports were always enhanced by increments in V1 spike counts and impaired by decrements, even when increments and decrements in spiking were generated in the same neuronal populations. Moreover, detecting changes in cortical activity depended on spike count integration rather than instantaneous changes in spiking. Recurrent neural networks trained in the task similarly relied on increments in neuronal activity when activity has costs. This work clarifies neuronal decoding strategies used by cerebral cortex to translate cortical spiking into percepts that can be used to guide behavior.SIGNIFICANCE STATEMENT Visual responses in the primary visual cortex (V1) are diverse, in that neurons can be either excited or inhibited by the onset of a visual stimulus. We selectively potentiated or suppressed V1 spiking in mice while they performed contrast change detection tasks. In other experiments, excitation or inhibition was delivered to V1 independent of visual stimuli. Mice readily detected increases in V1 spiking while equivalent reductions in V1 spiking suppressed the probability of detection, even when increases and decreases in V1 spiking were generated in the same neuronal populations. Our data raise the striking possibility that only increments in spiking are used to render information to structures downstream of V1.

Spatial updating of attention across eye movements: A neuro-computational approach.

While we are scanning our environment, the retinal image changes with every saccade. Nevertheless, the visual system anticipates where an attended target will be next and attention is updated to the new location. Recently, two different types of perisaccadic attentional updates were discovered: predictive remapping of attention before saccade onset (Rolfs, Jonikaitis, Deubel, & Cavanagh, 2011) and lingering of attention after saccade (Golomb, Chun, & Mazer, 2008; Golomb, Pulido, Albrecht, Chun, & Mazer, 2010). We here propose a neuro-computational model located in lateral intraparietal cortex based on a previous model of perisaccadic space perception (Ziesche & Hamker, 2011, 2014). Our model can account for both types of updating of attention at a neural-systems level. The lingering effect originates from the late updating of the proprioceptive eye-position signal and the remapping from the early corollary-discharge signal. We put these results in relationship to predictive remapping of receptive fields and show that both phenomena arise from the same simple, recurrent neural circuit. Thus, together with the previously published results, the model provides a comprehensive framework for discussing multiple experimental observations that occur around saccades.

Dynamics of fixational eye position and microsaccades during spatial cueing: the case of express microsaccades

Microsaccades are systematically modulated by peripheral spatial cues, and these eye movements have been implicated in perceptual and motor performance changes in cueing tasks. However, an additional oculomotor factor that may also influence performance in these tasks, fixational eye position itself, has been largely neglected so far. Using precise eye tracking and real-time retinal-image stabilization, we carefully analyzed fixational eye position dynamics and related them to microsaccade generation during spatial cueing. As expected, during baseline fixation, microsaccades corrected for a foveal motor error away from the preferred retinal locus of fixation (the so-called ocular position "set point" of the oculomotor system). However, we found that this relationship was violated during a short period immediately after cue onset; a subset of cue-directed "express microsaccades" that were highly precise in time and direction, and that were larger than regular microsaccades, occurred. These movements, having <100-ms latencies from cue onset, were triggered when fixational eye position was already at the oculomotor set point when the cue appeared; they were thus error-increasing rather than error-decreasing. Critically, even when no microsaccades occurred, fixational eye position itself was systematically deviated toward the cue, again with ~100-ms latency, suggesting that the oculomotor system establishes a new set point at different postcue times. This new set point later switched to being away from the cue after ~200-300 ms. Because eye position alters the location of retinal images, our results suggest that both eye position and microsaccades can be associated with performance changes in spatial cueing tasks. NEW & NOTEWORTHY Covert spatial cueing tasks are a workhorse for studying cognitive processing in humans and monkeys, but gaze is not perfectly stable during these tasks. We found that minute fixational eye position changes, independent of the more studied microsaccades, are not random in cueing tasks and are thus not "averaged out" in analyses. These changes can additionally dictate microsaccade times. Thus, in addition to microsaccadic influences, retinal image changes associated with fixational eye position are relevant for performance in cueing tasks.

Analysis of Retinal Architectural Changes Using Intraoperative OCT Following Surgical Manipulations With Membrane Flex Loop in the DISCOVER Study.

PurposeWe investigate acute retinal alterations identified on intraoperative optical coherence tomography (iOCT) immediately following surgical intervention with the Finesse Flex Loop for vitreoretinal interface disorders.MethodsThe Determination of feasibility of Intraoperative Spectral domain microscope Combined/integrated OCT Visualization during En face Retinal and ophthalmic surgery (DISCOVER) study is a prospective multisurgeon intraoperative OCT clinical study. Subjects who had participated in the DISCOVER study and had undergone macular surgery with the membrane flex loop from August 2014 to July 2016 were identified. iOCT images and video sequences were evaluated at various surgical time points. Subjects were excluded if iOCT images were not obtained over the area of membrane peeling performed with the membrane flex loop. Qualitative characteristics of intraoperative changes in retinal images were analyzed, with particular focus on the retinal layers within the bed of membrane peeling performed specifically with the membrane flex loop.ResultsWe studied 34 eyes of 34 patients, with a mean age of 72.7 (± 6.4) years, 25 of whom were women. The intraoperative diagnosis was full thickness macular hole in 21 eyes (62%) and epiretinal membrane in 13 (38%). All eyes had successful indocyanine green–assisted inner limiting membrane (ILM) flap initiation using the membrane loop. Intraoperative OCT demonstrated expansion of the ellipsoid zone-to-RPE distance in 3 eyes (9%), definitive subretinal fluid accumulation in 1 eye (3%), and hyperreflectivity of the inner retinal layers associated with retinal hemorrhage in 10 eyes (29%). Retinal breaks were not observed in any eye.ConclusionsAcute retinal alterations after ILM peeling with the membrane flex loop were visualized at a frequency of less than 10%. Additional research is needed to understand the clinical impact, if any, of these architectural alterations.

Crowdsourcing to Evaluate Fundus Photographs for the Presence of Glaucoma.

To assess the accuracy of crowdsourcing for grading optic nerve images for glaucoma using Amazon Mechanical Turk before and after training modules. Images (n=60) from 2 large population studies were graded for glaucoma status and vertical cup-to-disc ratio (VCDR). In the baseline trial, users on Amazon Mechanical Turk (Turkers) graded fundus photos for glaucoma and VCDR after reviewing annotated example images. In 2 additional trials, Turkers viewed a 26-slide PowerPoint training or a 10-minute video training and passed a quiz before being permitted to grade the same 60 images. Each image was graded by 10 unique Turkers in all trials. The mode of Turker grades for each image was compared with an adjudicated expert grade to determine accuracy as well as the sensitivity and specificity of Turker grading. In the baseline study, 50% of the images were graded correctly for glaucoma status and the area under the receiver operating characteristic (AUROC) was 0.75 [95% confidence interval (CI), 0.64-0.87]. Post-PowerPoint training, 66.7% of the images were graded correctly with AUROC of 0.86 (95% CI, 0.78-0.95). Finally, Turker grading accuracy was 63.3% with AUROC of 0.89 (95% CI, 0.83-0.96) after video training. Overall, Turker VCDR grades for each image correlated with expert VCDR grades (Bland-Altman plot mean difference=-0.02). Turkers graded 60 fundus images quickly and at low cost, with grading accuracy, sensitivity, and specificity, all improving with brief training. With effective education, crowdsourcing may be an efficient tool to aid in the identification of glaucomatous changes in retinal images.

Temporal impulse response function of the visual system estimated from ocular following responses in humans

Early visual processing functions as a set of spatiotemporal image filters. Our ability to sense changes in retinal images is determined by these filters along the temporal axis. In this study, we developed a paradigm to identify the kernel of the temporal filters based on ocular following responses (OFRs) to two-frame apparent motion stimuli. We first conducted two experiments to acquire fundamental data. In the first experiment, in which a quarter wavelength step of a sinusoidal grating was presented with various inter-stimulus intervals (ISIs), we found that OFRs were reversed by the ISI, which is consistent with previous findings. In the second experiment, a quarter wavelength step of a sinusoidal grating was applied with various durations of the initial image frame (motion onset delays; MODs); we found that longer exposure to the initial image reduced OFRs. Parameters of motion energy model involving temporal filters were optimized so that the model could reproduce the dependence of OFRs on ISIs and MODs. We were then able to successfully obtain quantitative estimates of the biphasic temporal filters with optimal frequencies in 6–8Hz. This method is completely objective and will thus be applicable to a wide range of human subjects and model animals.

Perisaccadic Updating of Visual Representations and Attentional States: Linking Behavior and Neurophysiology.

During natural vision, saccadic eye movements lead to frequent retinal image changes that result in different neuronal subpopulations representing the same visual feature across fixations. Despite these potentially disruptive changes to the neural representation, our visual percept is remarkably stable. Visual receptive field remapping, characterized as an anticipatory shift in the position of a neuron’s spatial receptive field immediately before saccades, has been proposed as one possible neural substrate for visual stability. Many of the specific properties of remapping, e.g., the exact direction of remapping relative to the saccade vector and the precise mechanisms by which remapping could instantiate stability, remain a matter of debate. Recent studies have also shown that visual attention, like perception itself, can be sustained across saccades, suggesting that the attentional control system can also compensate for eye movements. Classical remapping could have an attentional component, or there could be a distinct attentional analog of visual remapping. At this time we do not yet fully understand how the stability of attentional representations relates to perisaccadic receptive field shifts. In this review, we develop a vocabulary for discussing perisaccadic shifts in receptive field location and perisaccadic shifts of attentional focus, review and synthesize behavioral and neurophysiological studies of perisaccadic perception and perisaccadic attention, and identify open questions that remain to be experimentally addressed.

Shape aftereffects reflect shape constancy operations: Appearance matters

One of the oldest known visual aftereffects is the shape aftereffect, wherein looking at a particular shape can make subsequent shapes seem distorted in the opposite direction. After viewing a narrow ellipse, for example, a perfect circle can look like a broad ellipse. It is thought that shape aftereffects are determined by the dimensions of successive retinal images. However, perceived shape is invariant for large retinal image changes resulting from different viewing angles; current understanding suggests that shape aftereffects should not be impacted by the operations responsible for this viewpoint invariance. By viewing adaptors from an angle, with subsequent frontoparallel tests, we establish that shape aftereffects are not solely determined by the dimensions of successive retinal images. Moreover, by comparing performance with and without stereo surface slant cues, we show that shape aftereffects reflect a weighted function of retinal image shape and surface slant information, a hallmark of shape constancy operations. Thus our data establish that shape aftereffects can be influenced by perceived shape, as determined by constancy operations, and must therefore involve higher-level neural substrates than previously thought.

Auto-detection of Longitudinal Changes in Retinal Images for Monitoring Diabetic Retinopathy

A computer-aided retinal image analysis could provide an immediate detection and monitoring of abnormalities present in the retinal image. It allows to diagnose some retinal diseases prior to specialist inspection. This paper presents automatic system which can aid in the detection and monitoring of diabetic retinopathy (DR). The method proposed here is based on the preliminary automatic registration of retinal images, and the detection of changes in retinal images. This is done by comparing the registered retinal images. A novel algorithm is developed to achieve accurate registration. It ensures that the detected changes reflect only the real changes, and avoids any artifacts associated with the registration procedure itself. The special facts about retinal images are considered while performing image differencing. The present work in this paper is motivated by the need for automated, objective, quantitative approaches to detect the appearance of lesions and to detect longitudinal changes for monitoring DR. This system will considerably reduce the overall workload of ophthalmologists. General Terms Image Processing

Shape aftereffects reflect shape constancy operations: Appearance matters.

One of the oldest known visual aftereffects is the shape aftereffect, wherein looking at a particular shape can make subsequent shapes seem distorted in the opposite direction. After viewing a narrow ellipse, for example, a perfect circle can look like a broad ellipse. It is thought that shape aftereffects are determined by the dimensions of successive retinal images. However, perceived shape is invariant for large retinal image changes resulting from different viewing angles; current understanding suggests that shape aftereffects should not be impacted by the operations responsible for this viewpoint invariance. By viewing adaptors from an angle, with subsequent frontoparallel tests, we establish that shape aftereffects are not solely determined by the dimensions of successive retinal images. Moreover, by comparing performance with and without stereo surface slant cues, we show that shape aftereffects reflect a weighted function of retinal image shape and surface slant information, a hallmark of shape constancy operations. Thus our data establish that shape aftereffects can be influenced by perceived shape, as determined by constancy operations, and must therefore involve higher-level neural substrates than previously thought.

Remapping of border ownership in the visual cortex.

We see objects as having continuity although the retinal image changes frequently. How such continuity is achieved is hard to understand, because neurons in the visual cortex have small receptive fields that are fixed on the retina, which means that a different set of neurons is activated every time the eyes move. Neurons in areas V1 and V2 of the visual cortex signal the local features that are currently in their receptive fields and do not show "remapping" when the image moves. However, subsets of neurons in these areas also carry information about global aspects, such as figure-ground organization. Here we performed experiments to find out whether figure-ground organization is remapped. We recorded single neurons in macaque V1 and V2 in which figure-ground organization is represented by assignment of contours to regions (border ownership). We found previously that border-ownership signals persist when a figure edge is switched to an ambiguous edge by removing the context. We now used this paradigm to see whether border ownership transfers when the ambiguous edge is moved across the retina. In the new position, the edge activated a different set of neurons at a different location in cortex. We found that border ownership was transferred to the newly activated neurons. The transfer occurred whether the edge was moved by a saccade or by moving the visual display. Thus, although the contours are coded in retinal coordinates, their assignment to objects is maintained across movements of the retinal image.

Encoding of naturalistic optic flow by motion sensitive neurons of nucleus rotundus in the zebra finch (Taeniopygia guttata)

The retinal image changes that occur during locomotion, the optic flow, carry information about self-motion and the three-dimensional structure of the environment. Especially fast moving animals with only little binocular vision depend on these depth cues for maneuvering. They actively control their gaze to facilitate perception of depth based on cues in the optic flow. In the visual system of birds, nucleus rotundus neurons were originally found to respond to object motion but not to background motion. However, when background and object were both moving, responses increased the more the direction and velocity of object and background motion on the retina differed. These properties may play a role in representing depth cues in the optic flow. We therefore investigated, how neurons in nucleus rotundus respond to optic flow that contains depth cues. We presented simplified and naturalistic optic flow on a panoramic LED display while recording from single neurons in nucleus rotundus of anaesthetized zebra finches. Unlike most studies on motion vision in birds, our stimuli included depth information. We found extensive responses of motion selective neurons in nucleus rotundus to optic flow stimuli. Simplified stimuli revealed preferences for optic flow reflecting translational or rotational self-motion. Naturalistic optic flow stimuli elicited complex response modulations, but the presence of objects was signaled by only few neurons. The neurons that did respond to objects in the optic flow, however, show interesting properties.

Visual attention and stability.

In the present review, we address the relationship between attention and visual stability. Even though with each eye, head and body movement the retinal image changes dramatically, we perceive the world as stable and are able to perform visually guided actions. However, visual stability is not as complete as introspection would lead us to believe. We attend to only a few items at a time and stability is maintained only for those items. There appear to be two distinct mechanisms underlying visual stability. The first is a passive mechanism: the visual system assumes the world to be stable, unless there is a clear discrepancy between the pre- and post-saccadic image of the region surrounding the saccade target. This is related to the pre-saccadic shift of attention, which allows for an accurate preview of the saccade target. The second is an active mechanism: information about attended objects is remapped within retinotopic maps to compensate for eye movements. The locus of attention itself, which is also characterized by localized retinotopic activity, is remapped as well. We conclude that visual attention is crucial in our perception of a stable world.

Unsupervised learning of head-centered representations in a network of spiking neurons

Movement planning based on visual information requires a transformation from a retina-centered into a heador body-centered frame of reference. It has been shown that such transformations can be achieved via basis function networks [1,2]. We investigated whether basis functions for coordinate transformations can be learned by a biologically plausible neural network. We employed a model network of spiking neurons that learns invariant representations based on spatio-temporal stimulus correlations [3]. The model consists of a three-stage network of leaky integrate-and-fire neurons with biologically realistic conductances. The network has two input layers, corresponding to neurons representing the retinal image and neurons representing the direction of gaze. These inputs are represented in the map layer via excitatory or modulatory connections, respectively, that exhibit Hebbian-like spiketiming dependent plasticity (STDP). Neurons within the map layer are connected via short-range lateral excitatory connections and unspecific lateral inhibition. We trained the network with stimuli corresponding to typical viewing situations when a visual scene is explored by saccadic eye movements, with gaze direction changing on a faster time scale than object positions in space. After learning, each neuron in the map layer was selective for a small subset of the stimulus space, with excitatory and modulatory connections adapted to achieve a topographic map of the inputs. Neurons in the output layer with a localized receptive field in the map layer were selective for positions in head-centered space, invariant to changes in retinal image due to changes in gaze direction. Our results show that coordinate transformations via basis function networks can be learned in a biologically plausible way by exploiting the spatio-temporal correlations between visual stimulation and eye position signals under natural viewing conditions.

Unsupervised learning of gain-field like interactions to achieve head-centered representations

Event Abstract Back to Event Unsupervised learning of gain-field like interactions to achieve head-centered representations Sebastian Thomas-Philipp1*, Frank Michler2 and Thomas Wachtler3 1 Ludwig-Maximilian-University, Computational Neuroscience, Department Biologie II, Germany 2 Philipps-Universitat, Neurophysik, Germany 3 Ludwig-Maximilians University Munich, Fakultat fur Biologie, Germany Movement planning based on visual information requires a transformation from a retina-centered into a head- or body-centered frame of reference. It has been shown that such transformations can be achieved via basis function networks [1,2]. We investigated whether basis functions for coordinate transformations can be learned by a biologically plausible neural network. We employed a model network of spiking neurons that learns invariant representations based on spatio-temporal stimulus correlations [3]. The model consists of a three-stage network of leaky integrate-and-fire neurons with biologically realistic conductances. The network has two input layers, corresponding to neurons representing the retinal image and neurons representing the direction of gaze. These inputs are represented in the map layer via excitatory or modulatory connections, respectively, that exhibit Hebbian-like spike-timing dependent plasticity. Neurons within the map layer are connected via short range lateral excitatory connections and unspecific lateral inhibition. We trained the network with stimuli corresponding to typical viewing situations when a visual scene is explored by saccadic eye movements, with gaze direction changing on a faster time scale than object positions in space. After learning, each neuron in the map layer was selective for a small subset of the stimulus space, with excitatory and modulatory connections adapted to achieve a topographic map of the inputs. Neurons in the output layer with a localized receptive field in the map layer were selective for positions in head-centered space, invariant to changes in retinal image due to changes in gaze direction. Our results show that coordinate transformations via basis function networks can be learned in a biologically plausible way by exploiting the spatio-temporal correlations between visual stimulation and eye position signals under natural viewing conditions. Acknowledgements:Supported by DFG Forschergruppe 560 and BCCN Munich

How color break-up occurs in the human-visual system: The mechanism of the color break-up phenomenon

— In the present set of experiments, we examined the mechanisms underlying color break-up (CBU), a phenomenon observed when images produced with a color-sequential projector are viewed. The perceived position of CBU was measured during fast eye movement, saccade with static and briefly flashed stimuli. Results showed that CBU did not simply correspond to the locus of the stimulus on the retina during saccades, because the width of the CBU perception was narrower than the distance of the eye movements. This effect was thought to be related to visual stability, which allows objects to be perceived as stationary even when the eyes move and the retinal image changes position. Visual stability is assumed to operate by compensating for the change in retinal image position using eye-position signals; however, this compensation is imperfect during saccades. Thus, incomplete compensation results in a CBU perception that is of a narrower width than the amplitude of the saccade. In conclusion, CBU cannot be simulated with moving video cameras because it results largely from the mechanisms of visual perception. Large inter-individual differences in perception of CBU were also found. This observation also supports the idea that CBU depends on the mechanism of perception.

Virtual-Reality Techniques Resolve the Visual Cues Used by Fruit Flies to Evaluate Object Distances

Insects can estimate distance or time-to-contact of surrounding objects from locomotion-induced changes in their retinal position and/or size [1–8]. Freely walking fruit flies (Drosophila melanogaster) use the received mixture of different distance cues to select the nearest objects for subsequent visits [9, 10]. Conventional methods of behavioral analysis fail to elucidate the underlying data extraction. Here we demonstrate first comprehensive solutions of this problem by substituting virtual for real objects; a tracker-controlled 360° panorama converts a fruit fly's changing coordinates into object illusions that require the perception of specific cues to appear at preselected distances up to infinity. An application reveals the following: (1) en-route sampling of retinal-image changes accounts for distance discrimination within a surprising range of at least 8–80 body lengths (20–200 mm). Stereopsis and peering are not involved. (2) Distance from image translation in the expected direction (motion parallax) outweighs distance from image expansion, which accounts for impact-avoiding flight reactions to looming objects. (3) The ability to discriminate distances is robust to artificially delayed updating of image translation. Fruit flies appear to interrelate self-motion and its visual feedback within a surprisingly long time window of about 2 s. The comparative distance inspection practiced in the small fruit fly deserves utilization in self-moving robots.

Simulation of Night Myopia in Pseudophakic Eyes

Purpose: Eyes with an intraocular lens (IOL) implanted that have no accommodative ability are likely to have night myopia caused by Purkinje's shift and chromatic aberration. We evaluated the changes in retinal images caused by night myopia in various IOL-implanted eyes by simulation using model eyes. Methods: A polymethylmethacrylate (PMMA) IOL, soft acrylic IOLs, and high-refractive-index silicone IOLs were prepared, and inserted into the model eye. The image plane of the model eye was determined as the line image at which a slit light of 560 nm was best focused, which was regarded as emmetropic condition under photopic vision. A slit light of 505 nm was then directed into the model eye. Its line image on the image plane was blurred, which was regarded as myopic condition under scotopic vision. Under each condition. The modulation transfer functions (MTF) were calculated and the line images were photographed. Results: The degree of reduction of MTF and blurring of the line image under the night myopic condition was in the order of PMMA<soft acrylic<high refractive index silicone IOL. Conclusion: Eyes with a soft acrylic or high-refractive-index silicone IOL may have more intense night myopia caused by chromatic aberration.

Human ocular vergence movements induced by changing size and disparity.

Human subjects viewed an electronically generated bright square. Horizontal movements of the two eyes were recorded with the scleral coil method. The dynamic properties of vergence movements induced by movement of the bright square were investigated for the following three kinds of stimulus motion: (a) both the size and the binocular disparity of the square changed together, in such a way as to exactly mimic the retinal image changes produced by a real object's motion in depth; (b) the changing-size component in (a) was present with no disparity component; (c) the changing-disparity component in (a) was present with no size component. The gain and phase of the ocular vergence responses to these three stimuli were computed. Ocular vergence movements were induced by changing size in all five subjects. Responses during binocular viewing were higher and less variable than responses during monocular viewing. Size oscillations induced ocular vergence oscillations with a phase lead of up to 65 deg relative to target size for frequencies of stimulation below 1.0 Hz. Vergence oscillation amplitudes were of the order of 10 min of arc and maximal for frequencies of 0.4-0.7 Hz. Ocular vergence movements were not induced by changes in target size in one dimension nor by flickering a stationary square. Ocular vergence movements induced by size changes were entirely transient with no sustained component: vergence responses to disparity were sustained. When the stimulus combined size change with disparity change in the ratio characteristic of a real moving object, vergence tracking was more accurate and less noisy than when the eyes were stimulated with the disparity component alone. The ocular vergence response induced by the combination of size change with disparity change was accurately predicted by linearly adding the vergence response produced by the size change alone to the vergence response produced by the disparity change alone: combined stimulation produced no evidence of non-linear interaction between responses to size change and to disparity change. The properties of vergence responses induced by changing size and by changing disparity showed several close correlations with the corresponding data on psychophysical sensitivity for motion-in-depth sensation. We suggest that responses to changing size contribute to the accuracy with which ocular vergence tracks real objects moving in depth.