The retinal and perceived locus of fixation in the human visual system

Bidirectional Photovoltage-Driven Oxide Transistors for Neuromorphic Visual Sensors

Biological vision is one of the most important parts of the human perception system. However, emulating biological visuals is challenging because it requires complementary photoexcitation and photoinhibition. Here, the study presents a bidirectional photovoltage-driven neuromorphic visual sensor (BPNVS) that is constructed by monolithically integrating two perovskite solar cells (PSCs) with dual-gate ion-gel-gated oxide transistors. PSCs act as photoreceptors, converting external visual stimuli into electrical signals, whereas oxide transistors generate neuromorphic signal outputs that can be adjusted to produce positive and negative photoresponses. This device mimics the human vision system's ability to recognize colored and color-mixed patterns. The device achieves a static color recognition accuracy of 96% by utilizing the reservoir computing system for feature extraction. The BPNVS mem-reservoir chip is also proposed for handing object movement and dynamic color recognition. This work is a significant step forward in neuromorphic sensing and complex pattern recognition.

The preferred retinal loci when the eyes converge

The preferred retinal locus (PRL) is the position on the retina to which humans direct stimuli during fixation. In healthy normal eyes, it has been shown to be very stable across time and between different tasks. Previous measurements of the PRL have been made under monocular viewing conditions. The current study examines where the PRLs in the two eyes' retinas are when subjects fixate binocularly and whether they shift when the demand for the eyes to converge is changed. Our apparatus allows us to see exactly where binocular stimuli fell on the two retinas during binocular fixation. Thus, our technique bypasses some of the issues involved in measuring binocular alignment with subjective techniques and previous objective techniques that use conventional eye trackers. These results show that PRLs shift slightly but systematically as the demand for convergence increases. The shifts cause under-convergence (also called exo fixation disparity) for near targets. They are not large enough to cause a break in binocular fusion. The fixation disparity we observed with increasing vergence demand is similar to fixation disparity observed in previous reports.

Differences of ocular oscillations and neuro-retinal structures in patients with nystagmus caused by GPR143 and FRMD7 gene variants

PURPOSE

Mutations of G protein-coupled receptor 143 (GPR143) and FERM domain containing 7 (FRMD7) may result in congenital nystagmus (CN) in the first 6 months of life. We aimed to compare the differences in ocular oscillations between patients with these two gene mutations as well as the functional and structural changes in their retinas and visual pathways.

METHODS

Medical records were retrospectively reviewed to identify patients of congenital nystagmus with confirmed mutations in either GPR143 or FMRD7 genes from January 2018 to May 2023. The parameters of the ocular oscillations were recorded using Eyelink 1000 Plus. The retinal structure and function were evaluated using optical coherence tomography and multi-focal electroretinography (mERG). The visual pathway and optical nerve projection were evaluated using visual evoked potentials. The next-generation sequencing technique was used to identify the pathogenic variations in the disease-causing genes for CN.

RESULTS

Twenty nystagmus patients of GPR143 and 21 patients of FMRD7 who had been confirmed by molecular testing between January 2018 and May 2023 were included. Foveal hypoplasia was detected only in patients with the GPR143 pathogenic variant. mERG examination showed a flat response topography in the GPR143 group compared to the FRMD7 group. VEP showed that bilateral amplitude inconsistency was detected only in the patients with GPR143 gene mutation. The amplitude and frequency of the ocular oscillations were not found to differ between patients with two different genetic mutations.

CONCLUSIONS

Although the etiology and molecular mechanisms are completely different between CN patients, they may have similar ocular oscillations. A careful clinical examination and electrophysiological test will be helpful in making a differential diagnosis. Our novel identified variants will further expand the spectrum of the GPR143 and FRMD7 variants.

Peak Cone Density Predicted from Outer Segment Length Measured on Optical Coherence Tomography

PURPOSE

To compare peak cone density predicted from outer segment length measured on optical coherence tomography with direct measures of peak cone density from adaptive optics scanning light ophthalmoscopy.

METHODS

Data from 42 healthy participants with direct peak cone density measures and optical coherence tomography line scans available were used in this study. Longitudinal reflectivity profiles were analyzed using two methods of identifying the boundaries of the ellipsoid and interdigitation zones to estimate maximum outer segment length: peak-to-peak and the slope method. These maximum outer segment length values were then used to predict peak cone density using a previously described geometrical model. A comparison between predicted and direct peak cone density measures was then performed.

RESULTS

The mean bias between observers for estimating maximum outer segment length across methods was less than 2 µm. Cone density predicted from the peak-to-peak method against direct cone density measures showed a mean bias of 6,812 cones/mm2 with 50% of participants displaying a 10% difference or less between predicted and direct cone density values. Cone density derived from the slope method showed a mean bias of -17,929 cones/mm2 relative to direct cone density measures, with only 41% of participants demonstrating less than a 10% difference between direct and predicted cone density values.

CONCLUSION

Predicted foveal cone density derived from peak-to-peak outer segment length measurements using commercial optical coherence tomography show modest agreement with direct measures of peak cone density from adaptive optics scanning light ophthalmoscopy. The methods used here are imperfect predictors of cone density, however, further exploration of this relationship could reveal a clinically relevant marker of cone structure.

High-resolution eye-tracking via digital imaging of Purkinje reflections

Reliably measuring eye movements and determining where the observer looks are fundamental needs in vision science. A classical approach to achieve high-resolution oculomotor measurements is the so-called dual Purkinje image (DPI) method, a technique that relies on the relative motion of the reflections generated by two distinct surfaces in the eye, the cornea and the back of the lens. This technique has been traditionally implemented in fragile and difficult to operate analog devices, which have remained exclusive use of specialized oculomotor laboratories. Here we describe progress on the development of a digital DPI, a system that builds on recent advances in digital imaging to enable fast, highly precise eye-tracking without the complications of previous analog devices. This system integrates an optical setup with no moving components with a digital imaging module and dedicated software on a fast processing unit. Data from both artificial and human eyes demonstrate subarcminute resolution at 1 kHz. Furthermore, when coupled with previously developed gaze-contingent calibration methods, this system enables localization of the line of sight within a few arcminutes.

Binocularity Principles of PRL Development in Patients With Macular Disease

Purpose

We tested the hypothesis that binocularity requirements for correspondence play a role in establishing the preferred retinal locus (PRL) in macular degeneration.

Methods

Monocular PRL locations in 202 eyes of 101 patients with macular degeneration (79 ± 10 years) were recorded with the MP1 microperimeter. Corresponding PRLs were those with similar polar angle and distance from former fovea in the better eye (BE) and the worse eye (WE).

Results

On average, the PRL in the BE was in the foveal proximity at 1.1 ± 0.99 degrees for 55 patients (foveal-driven PRL) and eccentrically at 6.9 ± 3.4 degrees for 46 patients with central lesions involving the fovea (peripheral-driven PRL). For the foveal-driven PRL group, the PRL in the BE was not affected by the status of the WE. In 100% of cases, the monocular PRL in the WE was in a corresponding location either on functioning retina or onto the lesion, or would fall onto the lesion during binocular viewing. For the peripheral-driven PRL group, the PRL location depended on the lesion size in both eyes to maximize correspondence and/or the function of peripheral vision during binocular viewing. In this group, PRL correspondence status was different for those with equal, unequal, or extensive lesions in both eyes.

Conclusions

Binocularity requirements for correspondence play an important role in determining the PRL location. We formulated two principles based on whether the BE has foveal sparing (foveal-driven PRL) or central lesions affecting the fovea (peripheral-driven PRL). The PRL should be evaluated in the framework of binocular viewing.

Retinal magnification factors at the fixation locus derived from schematic eyes with four individualized surfaces

Retinal magnification factors (RMFs) allow the conversion of angles to lengths in retinal images. In this work, we propose paraxial and non-paraxial RMF calculation methods that incorporate the individual topography and separation of the anterior and posterior surfaces of the cornea and crystalline lens, assuming homogeneous ocular media. Across 34 eyes, the two RMF methods differ by 0.1% on average, due to surface tilt, decenter, and lack of rotational symmetry in the non-paraxial modeling, which results in up to 2.2% RMF variation with retinal meridian. Differences with widely used individualized RMF calculation methods are smallest for eyes with ∼24 mm axial length, and as large as 7.5% in a 29.7 mm long eye (15D myope). To better model the capture of retinal images, we propose the tracing of chief rays, instead of the scaling of posterior nodal or principal distances often used in RMF definitions. We also report that RMF scale change is approximately proportional to both refractive error and axial separation between the ophthalmoscope's exit pupil and the eye's entrance pupil, resulting in RMF changes as large as 13% for a 1cm displacement in a 15D myopic eye. Our biometry data shows weak correlation and statistical significance between surface radii and refractive error, as well as axial length, whether considering all eyes in the study, or just the high myopes, defined as those with refractive error sphere equivalent ≤ -4D. In contrast, vitreous thicknesses show a strong correlation (r ≤ -0.92) and significance (p ≤ 10-13) with refractive error when considering all eyes or just high myopes (r ≤ -0.95; p ≤ 10-5). We also found that potential RMF change with depth of cycloplegia and/or residual accommodation is smaller than 0.2%. Finally, we propose the reporting of individual ocular biometry data and a detailed RMF calculation method description in scientific publications to facilitate the comparison of retinal imaging biomarker data across studies.

Rethinking Model-Based Gaze Estimation

Over the past several years, a number of data-driven gaze tracking algorithms have been proposed, which have been shown to outperform classic model-based methods in terms of gaze direction accuracy. These algorithms leverage the recent development of sophisticated CNN architectures, as well as the availability of large gaze datasets captured under various conditions. One shortcoming of black-box, end-to-end methods, though, is that any unexpected behaviors are difficult to explain. In addition, there is always the risk that a system trained with a certain dataset may not perform well when tested on data from a different source (the "domain gap" problem.) In this work, we propose a novel method to embed eye geometry information in an end-to-end gaze estimation network by means of a "geometric layer". Our experimental results show that our system outperforms other state-of-the-art methods in cross-dataset evaluation, while producing competitive performance over within dataset tests. In addition, the proposed system is able to extrapolate gaze angles outside the range of those considered in the training data.

Impact of Reference Center Choice on Adaptive Optics Imaging Cone Mosaic Analysis

Purpose

Foveal center marking is a key step in retinal image analysis. We investigated the discordance between the adaptive optics (AO) montage center (AMC) and the foveal pit center (FPC) and its implications for cone mosaic analysis using a commercial flood-illumination AO camera.

Methods

Thirty eyes of 30 individuals (including 15 healthy and 15 patients with rod–cone dystrophy) were included. Spectral-domain optical coherence tomography was used to determine the FPC, and flood-illumination AO imaging was performed with overlapping image frames to create an AO montage. The AMC was determined by averaging the (0,0) coordinates in the four paracentral overlapping AO image frames. Cone mosaic measurements at various retinal eccentricities were compared between corresponding retinal loci relative to the AMC or FPC.

Results

AMCs were located temporally to the FPCs in 14 of 15 eyes in both groups. The average AMC–FPC discordance was 0.85° among healthy controls and 0.33° among patients with rod-cone dystrophy (P < 0.05). The distance of the AMC from the FPC was a significant determinant of the cone density (β estimate = 218 cells/deg²/deg; 95% confidence interval [CI], 107–330; P < 0.001) and inter-cone distance (β estimate = 0.28 arcmin/deg; 95% CI, 0.15–0.40; P < 0.001), after adjustment for age, sex, axial length, spherical equivalent, eccentricity, and disease status.

Conclusions

There is a marked mismatch between the AMC and FPC in healthy eyes that may be modified by disease process such as rod–cone dystrophy. We recommend users of AO imaging systems carefully align the AO montage with a foveal anatomical landmark, such as the FPC, to ensure precise and reproducible localization of the eccentricities and regions of interest for cone mosaic analysis.

Fixational eye movements in passive versus active sustained fixation tasks

Human fixational eye movements are so small and precise that high-speed, accurate tools are needed to fully reveal their properties and functional roles. Where the fixated image lands on the retina and how it moves for different levels of visually demanding tasks is the subject of the current study. An Adaptive Optics Scanning Laser Ophthalmoscope (AOSLO) was used to image, track and present a variety of fixation targets (Maltese cross, disk, concentric circles, Vernier and tumbling-E letter) to healthy subjects. During these different passive (static) or active (discriminating) tasks under natural eye motion, the landing position of the target on the retina was tracked in space and time over the retinal image directly with high spatial (<1 arcmin) and temporal (960 Hz) resolution. We computed both the eye motion and the exact trajectory of the fixated target's motion over the retina. We confirmed that compared to passive tasks, active tasks elicited a partial inhibition of microsaccades, leading to longer drift periods compensated by larger corrective saccades. Consequently, the overall fixation stability during active tasks was on average 57% larger than during passive tasks. The preferred retinal locus of fixation was the same for each task and did not coincide with the location of the peak cone density.