Video-based Eye Tracker Research Articles

Multimodal oculomotor assessment reveals prodromal markers of Parkinson’s disease in non-manifesting LRRK2 G2019S mutation carriers

Oculomotor behaviour changes in patients with Parkinson’s disease (PD) are a promising source of prodromal disease markers. Capitalizing on this phenomenon to facilitate early diagnosis requires oculomotor assessment in prodromal cohorts. We examined oculomotor behaviour in non-manifesting LRRK2 G2019S mutation carriers (LRRK2-NM), who have heightened PD risk.Seventeen LRRK2-NM participants, 47 patients with idiopathic PD, and 63 healthy age-matched control participants completed an interleaved pro- and antisaccade task while undergoing video-based eye-tracking. We analyzed between-group differences in saccade, pupil, blink, and fixation acquisition behaviour. Patients with PD showed previously demonstrated abnormalities (saccade hypometria, antisaccade errors). Relative to controls, LRRK2-NM participants and patients with PD both displayed increased short-latency prosaccades and reduced pupil velocity, plus altered fixation acquisition—less preemptive returning of gaze to the future fixation point location. Interestingly, the effect on blink probability was opposite—higher than controls in LRRK2-NM participants but lower in patients with PD. Future longitudinal studies must confirm the viability of these features as prodromal PD markers.

Microsaccades Tracking by Secondary Speckle Pattern Analysis.

Here we propose a not pupil-dependent microsaccades tracking technique and a novel detection method. We present a proof of concept for detecting microsaccades using a non-contact laser-based photonic system recording and processing the temporal changes of speckle patterns scattered from an eye sclera. The data, simultaneously recorded by the speckle-based tracker (SBT) and the video-based eye tracker (Eyelink), was analyzed by the frequently used detection method of Engbert and Kliegl (E&K) and by advanced machine learning detection (MLD) techniques. We detected 93% of microsaccades in the SBT data out of microsaccades detected in the Eyelink data with the E&K method. By utilizing MLD, a precision of 86% was achieved. The findings of our study demonstrate a potential improvement in measuring tiny eye movements, such as microsaccades, using speckle-based eye tracking and, thus, an alternative to video-based eye tracking for detecting microsaccades.

Oculomotor behaviors in youth with an eating disorder: findings from a video-based eye tracking task

BackgroundThe oculomotor circuit spans many cortical and subcortical areas that have been implicated in psychiatric disease. This, combined with previous findings, suggests that eye tracking may be a useful method to investigate eating disorders. Therefore, this study aimed to assess oculomotor behaviors in youth with and without an eating disorder.MethodsFemale youth with and without an eating disorder completed a structured task involving randomly interleaved pro-saccade (toward at a stimulus) and anti-saccade (away from stimulus) trials with video-based eye tracking. Differences in saccades (rapid eye movements between two points), eye blinks and pupil were examined.ResultsYouth with an eating disorder (n = 65, Mage = 17.16 ± 3.5 years) were compared to healthy controls (HC; n = 65, Mage = 17.88 ± 4.3 years). The eating disorder group was composed of individuals with anorexia nervosa (n = 49), bulimia nervosa (n = 7) and other specified feeding or eating disorder (n = 9). The eating disorder group was further divided into two subgroups: individuals with a restrictive spectrum eating disorder (ED-R; n = 43) or a bulimic spectrum eating disorder (ED-BP; n = 22). In pro-saccade trials, the eating disorder group made significantly more fixation breaks than HCs (F(1,128) = 5.33, p = 0.023). The ED-BP group made the most anticipatory pro-saccades, followed by ED-R, then HCs (F(2,127) = 3.38, p = 0.037). Groups did not differ on rate of correct express or regular latency pro-saccades. In anti-saccade trials, groups only significantly differed on percentage of direction errors corrected (F(2, 127) = 4.554, p = 0.012). The eating disorder group had a significantly smaller baseline pupil size (F(2,127) = 3.60, p = 0.030) and slower pro-saccade dilation velocity (F(2,127) = 3.30, p = 0.040) compared to HCs. The ED-R group had the lowest blink probability during the intertrial interval (ITI), followed by ED-BP, with HCs having the highest ITI blink probability (F(2,125) = 3.63, p = 0.029).ConclusionsThese results suggest that youth with an eating disorder may have different oculomotor behaviors during a structured eye tracking task. The oculomotor behavioral differences observed in this study presents an important step towards identifying neurobiological and cognitive contributions towards eating disorders.

Impairment of Visual Fixation and Preparatory Saccade Control in Borderline Personality Disorder With and Without Comorbid Attention-Deficit/Hyperactivity Disorder

BackgroundBorderline personality disorder (BPD) is associated with heightened impulsivity, evidenced by increased substance abuse, self-harm, and suicide attempts. Addressing impulsivity in individuals with BPD is a therapeutic objective, but its underlying neural basis in this clinical population remains unclear, partly due to its frequent comorbidity with attention-deficit/hyperactivity disorder (ADHD). MethodsWe used a response inhibition paradigm—the interleaved pro-/antisaccade task—among adolescents diagnosed with BPD with and without comorbid ADHD (n = 25 and n = 24, respectively) during concomitant video-based eye tracking. We quantified various eye movement response parameters reflective of impulsive action during the task, including delay to fixation acquisition, fixation breaks, anticipatory saccades, and direction errors with express saccade (saccade reaction time: 90–140 ms) and regular saccade latencies (saccade reaction time > 140 ms). ResultsIndividuals with BPD exhibited deficient response preparation, as evidenced by reduced visual fixation on task cues and greater variability of saccade responses (i.e., saccade reaction time and peak velocity). The ADHD/BPD group shared these traits and made more anticipatory responses and direction errors with express saccade latencies and reduced error correction. ConclusionsSaccadic deficits in BPD and ADHD/BPD stemmed not from an inability to execute antisaccades but rather from inadequate preparation for the upcoming task set. These distinctions may arise due to abnormal signaling in cortical areas like the frontal eye fields, posterior parietal cortex, and anterior cingulate cortex. Understanding these mechanisms could provide insights into targeted interventions focusing on task set preparation to manage response inhibition deficits in BPD and ADHD/BPD.

Altered pupil light and darkness reflex and eye-blink responses in late-life depression

BackgroundLate-life depression (LLD) is a prevalent neuropsychiatric disorder in the older population. While LLD exhibits high mortality rates, depressive symptoms in older adults are often masked by physical health conditions. In younger adults, depression is associated with deficits in pupil light reflex and eye blink rate, suggesting the potential use of these responses as biomarkers for LLD.MethodsWe conducted a study using video-based eye-tracking to investigate pupil and blink responses in LLD patients (n = 25), older (OLD) healthy controls (n = 29), and younger (YOUNG) healthy controls (n = 25). The aim was to determine whether there were alterations in pupil and blink responses in LLD compared to both OLD and YOUNG groups.ResultsLLD patients displayed significantly higher blink rates and dampened pupil constriction responses compared to OLD and YOUNG controls. While tonic pupil size in YOUNG differed from that of OLD, LLD patients did not exhibit a significant difference compared to OLD and YOUNG controls. GDS-15 scores in older adults correlated with light and darkness reflex response variability and blink rates. PHQ-15 scores showed a correlation with blink rates, while MoCA scores correlated with tonic pupil sizes.ConclusionsThe findings demonstrate that LLD patients display altered pupil and blink behavior compared to OLD and YOUNG controls. These altered responses correlated differently with the severity of depressive, somatic, and cognitive symptoms, indicating their potential as objective biomarkers for LLD.

Regulation of pupil size in natural vision across the human lifespan.

Vision is mediated by light passing through the pupil, which changes in diameter from approximately 2 to 8 mm between bright and dark illumination. With age, mean pupil size declines. In laboratory experiments, factors affecting pupil size can be experimentally controlled. How the pupil reflects the change in retinal input from the visual environment under natural viewing conditions is unclear. We address this question in a field experiment (N = 83, 43 female, 18-87 years) using a custom-made wearable video-based eye tracker with a spectroradiometer measuring near-corneal spectral irradiance. Participants moved in and between indoor and outdoor environments varying in spectrum and engaged in a range of everyday tasks. Our data confirm that light-adapted pupil size is determined by light level, with a better model fit of melanopic over photopic units, and that it decreased with increasing age, yielding steeper slopes at lower light levels. We found no indication that sex, iris colour or reported caffeine consumption affects pupil size. Our exploratory results point to a role of photoreceptor integration in controlling steady-state pupil size. The data provide evidence for considering age in personalized lighting solutions and against the use of photopic illuminance alone to assess the impact of real-world lighting conditions.

Using Deep Learning to Increase Eye-Tracking Robustness, Accuracy, and Precision in Virtual Reality.

Algorithms for the estimation of gaze direction from mobile and video-based eye trackers typically involve tracking a feature of the eye that moves through the eye camera image in a way that covaries with the shifting gaze direction, such as the center or boundaries of the pupil. Tracking these features using traditional computer vision techniques can be difficult due to partial occlusion and environmental reflections. Although recent efforts to use machine learning (ML) for pupil tracking have demonstrated superior results when evaluated using standard measures of segmentation performance, little is known of how these networks may affect the quality of the final gaze estimate. This work provides an objective assessment of the impact of several contemporary ML-based methods for eye feature tracking when the subsequent gaze estimate is produced using either feature-based or model-based methods. Metrics include the accuracy and precision of the gaze estimate, as well as drop-out rate.

Potential of a laser pointer contact lens to improve the reliability of video-based eye-trackers in indoor and outdoor conditions.

Many video-based eye trackers rely on detecting and tracking ocular features, a task that can be negatively affected by a number of individual or environmental factors. In this context, the aim of this study was to practically evaluate how the use of a scleral contact lens with two integrated nearinfrared lasers (denoted CLP) could improve the tracking robustness in difficult lighting conditions, particularly outdoor ones. We assessed the ability of the CLP (on a model eye) to detect the lasers and to deduce a gaze position with an accuracy better than 1° under four lighting conditions (1 lx, 250 lx, 50 klux and alternating 1lx /250 lx) on an artificial eye. These results were compared to the ability of a commercial eye tracker (Pupil Core) to detect the pupil on human eyes with a confidence score equal to or greater than 0.9. CLP provided good results in all conditions (tracking accuracy and detection rates). In comparison, the Pupil Core performed well in all indoor conditions (99% detection) but failed in outdoor conditions (9.85% detection). In conclusion, the CLP presents strong potential to improve the reliability of video-based eyetrackers in outdoor conditions by providing easy trackable feature.

Automated Analysis Pipeline for Extracting Saccade, Pupil, and Blink Parameters Using Video-Based Eye Tracking.

The tremendous increase in the use of video-based eye tracking has made it possible to collect eye tracking data from thousands of participants. The traditional procedures for the manual detection and classification of saccades and for trial categorization (e.g., correct vs. incorrect) are not viable for the large datasets being collected. Additionally, video-based eye trackers allow for the analysis of pupil responses and blink behaviors. Here, we present a detailed description of our pipeline for collecting, storing, and cleaning data, as well as for organizing participant codes, which are fairly lab-specific but nonetheless, are important precursory steps in establishing standardized pipelines. More importantly, we also include descriptions of the automated detection and classification of saccades, blinks, "blincades" (blinks occurring during saccades), and boomerang saccades (two nearly simultaneous saccades in opposite directions where speed-based algorithms fail to split them), This is almost entirely task-agnostic and can be used on a wide variety of data. We additionally describe novel findings regarding post-saccadic oscillations and provide a method to achieve more accurate estimates for saccade end points. Lastly, we describe the automated behavior classification for the interleaved pro/anti-saccade task (IPAST), a task that probes voluntary and inhibitory control. This pipeline was evaluated using data collected from 592 human participants between 5 and 93 years of age, making it robust enough to handle large clinical patient datasets. In summary, this pipeline has been optimized to consistently handle large datasets obtained from diverse study cohorts (i.e., developmental, aging, clinical) and collected across multiple laboratory sites.

What is a blink? Classifying and characterizing blinks in eye openness signals

Blinks, the closing and opening of the eyelids, are used in a wide array of fields where human function and behavior are studied. In data from video-based eye trackers, blink rate and duration are often estimated from the pupil-size signal. However, blinks and their parameters can be estimated only indirectly from this signal, since it does not explicitly contain information about the eyelid position. We ask whether blinks detected from an eye openness signal that estimates the distance between the eyelids (EO blinks) are comparable to blinks detected with a traditional algorithm using the pupil-size signal (PS blinks) and how robust blink detection is when data quality is low. In terms of rate, there was an almost-perfect overlap between EO and PS blink (F1 score: 0.98) when the head was in the center of the eye tracker’s tracking range where data quality was high and a high overlap (F1 score 0.94) when the head was at the edge of the tracking range where data quality was worse. When there was a difference in blink rate between EO and PS blinks, it was mainly due to data loss in the pupil-size signal. Blink durations were about 60 ms longer in EO blinks compared to PS blinks. Moreover, the dynamics of EO blinks was similar to results from previous literature. We conclude that the eye openness signal together with our proposed blink detection algorithm provides an advantageous method to detect and describe blinks in greater detail.

Saccadic movements during an exploratory visual search task in patients with glaucomatous visual field loss.

To evaluate the saccadic movements of patients with visual field loss due to primary open-angle glaucoma. Thirteen patients with good visual acuity (0.2 logMAR or better) (seven patients with primary open-angle glaucoma 65 ± 13 years) and six controls (51 ± 6 years) yielded a comprehensive ophthalmological examination, including Humphrey Visual Field tests (SITA-Standard 24-2), and performed a monocular, exploratory digital visual search task that quantifies the duration for finding the number "4" on a random array of digits distributed on the screen. After individual adjustments of the angle and distance positioning, the screen was spatially matched with the 24-2 visual field, and divided into five areas for analysis. During the task, saccades were simultaneously recorded in the same eye with a video-based eye tracker. The patients with primary open-angle glaucoma showed a significantly higher number of saccades/screen (median ± interquartile range, 59.00 ± 29.00 vs. 32.50 ± 19.75 saccades (p=0.027) and visual search time per screen (38.50 ± 60.14 vs. 23.75 ± 8.90 seconds (p=0.035) than the controls did. Although the univariate analysis indicated a significant correlation with visual field mean deviation (coefficient=26.19 (p=0.02), only the visual search time/screen was significantly associated with the number of saccades/screen in the multivariate regression model (coefficient=0.55 (p<0.001). Overall, no significant correlation was observed between the sectorial number of saccades and the sensitivity of the five visual field areas. The patients with primary open-angle glaucoma show impaired search performance and showed a higher number of saccades needed to find stimuli when performing the exploratory visual task.

Invited Session IV: Extended reality--applications in vision science and beyond: Recent developments in head-mounted eye tracking for the understanding of natural(istic) behavior.

The scientific investigation of eye movements in natural or simulated and "naturalistic" environments has historically been operating at the limit of what head-worn eye tracking technology is capable of. In this presentation, I will review efforts by myself and collaborators at RIT to push these limits, and to increase the scope of scientific inquiry into natural visual and motor behavior. This talk will end with a brief discussion of emerging methods, most of which are aimed at resolving long standing limitations to video-based eye tracking. Most notably, as a consequence of USB transfer limits and stringent power budgets, video based eye trackers are restricted to either a high spatial resolution of the eye image which improves the spatial accuracy of the final gaze estimate, or a high temporal sampling rate (i.e., a high number of eye frames per second), but not both.

Delayed oculomotor response associates with optic neuritis in youth with demyelinating disorders

IntroductionImpairment in visual and cognitive functions occur in youth with demyelinating disorders such as multiple sclerosis, neuromyelitis optica spectrum disorder, and myelin oligodendrocyte glycoprotein antibody-associated disease. Quantitative behavioral assessment using eye-tracking and pupillometry can provide functional metrics for important prognostic and clinically relevant information at the bedside. MethodsChildren and adolescents diagnosed with demyelinating disorders and healthy, age-matched controls completed an interleaved pro- and anti-saccade task using video-based eye-tracking and underwent spectral-domain optical coherence tomography examination for evaluation of retinal nerve fiber layer and ganglion cell inner plexiform layer thickness. Low-contrast visual acuity and Symbol Digit Modalities Test were performed for visual and cognitive functional assessments. We assessed saccade and pupil parameters including saccade reaction time, direction error rate, pupil response latency, peak constriction time, and peak constriction and dilation velocities. Generalized Estimating Equations were used to examine the association of eye-tracking parameters with optic neuritis history, structural metrics, and visual and cognitive scores. ResultsThe study included 36 demyelinating disorders patients, aged 8–18 yrs. (75% F; median = 15.22 yrs., SD = 2.8) and 34 age-matched controls (65% F; median = 15.26 yrs., SD = 2.3). Surprisingly, pro- and anti-saccade performance was comparable between patients and controls, whereas pupil control was altered in patients. Oculomotor latency measures were strongly associated with the number of optic neuritis episodes, including saccade reaction time, pupil response latency, and peak constriction time. Peak constriction time was associated with both retinal nerve fiber layer and ganglion cell inner plexiform layer thickness. Pupil response latency and peak constriction time were associated with visual acuity. Pupil velocity for both constriction and dilation was associated with Symbol Digit Modalities Test scores. ConclusionThe strong associations between oculomotor measures with history of optic neuritis, structural, visual, and cognitive assessments in these cohorts demonstrates that quantitative eye-tracking can be useful for probing demyelinating injury of the brain and optic nerve. Future studies should evaluate their utility in discriminating between demyelinating disorders and tracking disease progression.

Eye detection and coarse localization of pupil for video-based eye tracking systems

A video-based eye tracking system generally captures NIR images, each of which contains one or two eyes of a subject. The subject’s point of gaze is then determined using 3D eye model and pupil centre corneal reflection technique. Eye detection and pupil localization play significant roles in video-based eye tracking systems. However, face rotation, wearing glasses, eye-shape variation and illumination variation make it difficult to detect an eye and localize a pupil accurately in the images captured by video-based eye tracking systems. In this paper, we proposed an eye detector that adopts a coarse-to-fine strategy. The eye detector consists of three classifiers: an ATLBP-THACs feature-based cascade classifier, a branch CNN and a multi-task CNN. We also proposed a method for coarse pupil localization. Coarse localization is an important step for pupil localization since it can provide initial pupil coordinates for a fine pupil localization method. Given a downscaled eye image, a shallow CNN is used to estimate the location of seven landmarks. On this basis, pupil center and radius are estimated. A method for small dim target enhancement is used to increase the contrast between pupil and background. The main goal of pupil enhancement is to make it easier to binarize an eye image. At last, component filtering is made by utilizing the estimated pupil center and radius. We collected a dataset named neepuEYE dataset that consists of 5500 NIR eye images from 109 people. The images can be used to generate augmented samples for training an eye detector since they contain eyes with different shape, orientation and pupil localization. Experimental results show that our eye detector is a fast and robust detector. Furthermore, our method for coarse pupil localization can obtain not only high detection rate but also high localization speed.

Using a video-based eye tracker to analyse the binocular near-reflex dynamics response.

This study presents a novel video-based eye-tracking system for analysing the dynamics of the binocular near-reflex response. The system enables the simultaneous measurement of convergence, divergence and pupillary size during accommodation and disaccommodation to aid the comprehensive understanding of the three-component near-reflex. A high-speed (90 Hz) video-based eye tracker was used to capture changes in eye gaze and pupil radius in 15 participants in response to altering stimulus conditions. An offline analysis involved separating the gaze vector components and pupil radius, which were fitted to a hyperbolic tangent function to characterise the dynamics of the near-reflex process. Significant differences in the temporal parameters of the pupil radius were observed between the near-to-far and far-to-near vision changes, with faster miosis compared with mydriasis. Additionally, differences in response times were found between gaze angle components, with longer convergence times compared to changes in the vertical direction (saccades). The steady-state values of the gaze components and pupil radius were in line with theoretical expectations and previous reports. The proposed system provides a non-invasive, portable and cost-effective method for evaluating near-reflex dynamics under natural viewing conditions using a video-based eye tracker. The sampling rate ensures the accurate assessment of vergence eye movements and pupillary dynamics. By simultaneously measuring eye convergence, divergence and pupil size, the system offers a more comprehensive assessment of the near-reflex response. This makes it a valuable tool for clinical diagnosis, research studies and investigating the effects of near work on the visual system.

Noise estimation for head-mounted 3D binocular eye tracking using Pupil Core eye-tracking goggles.

Head-mounted, video-based eye tracking is becoming increasingly common and has promise in a range of applications. Here, we provide a practical and systematic assessment of the sources of measurement uncertainty for one such device - the Pupil Core - in three eye-tracking domains: (1) the 2D scene camera image; (2) the physical rotation of the eye relative to the scene camera 3D space; and (3) the external projection of the estimated gaze point location onto the target plane or in relation to world coordinates. We also assess eye camera motion during active tasks relative to the eye and the scene camera, an important consideration as the rigid arrangement of eye and scene camera is essential for proper alignment of the detected gaze. We find that eye camera motion, improper gaze point depth estimation, and erroneous eye models can all lead to added noise that must be considered in the experimental design. Further, while calibration accuracy and precision estimates can help assess data quality in the scene camera image, they may not be reflective of errors and variability in gaze point estimation. These findings support the importance of eye model constancy for comparisons across experimental conditions and suggest additional assessments of data reliability may be warranted for experiments that require the gaze point or measure eye movements relative to the external world.

Motor synchronization and impulsivity in pediatric borderline personality disorder with and without attention-deficit hyperactivity disorder: an eye-tracking study of saccade, blink and pupil behavior.

Shifting motor actions from reflexively reacting to an environmental stimulus to predicting it allows for smooth synchronization of behavior with the outside world. This shift relies on the identification of patterns within the stimulus - knowing when a stimulus is predictable and when it is not - and launching motor actions accordingly. Failure to identify predictable stimuli results in movement delays whereas failure to recognize unpredictable stimuli results in early movements with incomplete information that can result in errors. Here we used a metronome task, combined with video-based eye-tracking, to quantify temporal predictive learning and performance to regularly paced visual targets at 5 different interstimulus intervals (ISIs). We compared these results to the random task where the timing of the target was randomized at each target step. We completed these tasks in female pediatric psychiatry patients (age range: 11-18 years) with borderline personality disorder (BPD) symptoms, with (n = 22) and without (n = 23) a comorbid attention-deficit hyperactivity disorder (ADHD) diagnosis, against controls (n = 35). Compared to controls, BPD and ADHD/BPD cohorts showed no differences in their predictive saccade performance to metronome targets, however, when targets were random ADHD/BPD participants made significantly more anticipatory saccades (i.e., guesses of target arrival). The ADHD/BPD group also significantly increased their blink rate and pupil size when initiating movements to predictable versus unpredictable targets, likely a reflection of increased neural effort for motor synchronization. BPD and ADHD/BPD groups showed increased sympathetic tone evidenced by larger pupil sizes than controls. Together, these results support normal temporal motor prediction in BPD with and without ADHD, reduced response inhibition in BPD with comorbid ADHD, and increased pupil sizes in BPD patients. Further these results emphasize the importance of controlling for comorbid ADHD when querying BPD pathology.

Altered Gaze Control During Emotional Face Exploration in Patients With Amyotrophic Lateral Sclerosis.

Up to 50% of patients with amyotrophic lateral sclerosis (ALS) present with cognitive problems and behavioral dysfunctions including recognition of human faces presenting different emotions. We investigated whether impaired processing of emotional faces is associated with abnormal scan paths during visual exploration. Cognitively unimpaired patients with ALS (n = 45) and matched healthy controls (n = 37) underwent neuropsychological assessment and video-based eye tracking. Eye movements were recorded while participants visually explored faces expressing different emotions (neutral, disgusted, happy, fearful, and sad) and houses mimicking faces. Compared with controls, patients with ALS fixated significantly longer to regions which are not relevant for emotional information when faces expressed fear (p = 0.007) and disgust (p = 0.006), whereas the eyes received less attention in faces expressing disgust (p = 0.041). Fixation duration in any area of interest was not significantly associated with the cognitive state or clinical symptoms of disease severity. In cognitively unimpaired patients with ALS, altered gaze patterns while visually exploring faces expressing different emotions might derive from impaired top-down attentional control with possible involvement of subliminal frontotemporal areas. This may account for indistinctness in emotion recognition reported in previous studies because nonsalient features retrieve more attention compared with salient areas. Current findings may indicate distinct emotion processing dysfunction of ALS pathology, which may be different from, for example, executive dysfunction.

Subclinical involvement of eye movements detected by video-based eye tracking in myasthenia gravis.

Ocular abnormalities in myasthenia gravis (MG) are characterized by severely limited movements and rapid saccades. Data about eye motility of MG patients whose ocular movements are apparently normal are lacking. Our study assessed the eye movement parameters in MG patients without clinical eye motility dysfunctions and investigated the effects of neostigmine administration on the eye motility in these patients. In this longitudinal study, we screened all patients diagnosed with MG referring to the Neurologic Clinic of the University of Catania between October 1, 2019, and June 30, 2021. Ten age- and sex-matched healthy controls were enrolled. Patients underwent eye movement recording using the EyeLink1000 Plus® eye tracker at baseline and after 90min from the intramuscular administration of neostigmine (0.5mg). A total of 14 MG patients with no clinical signs of ocular motor dysfunction (64.3% men, with a mean age of 50.4 ± 14.4years) were enrolled. At baseline, saccades in MG patients showed slower velocities and longer latencies compared to controls. Moreover, the fatigue test induced a reduction in saccadic velocity and an increase in latencies. After neostigmine administration, the ocular motility analysis showed shorter saccadic latencies and a significant improvement of velocities. Eye motility is impaired even in MG patients with no clinical evidence of ocular movement disturbance. Video-based eye tracking may detect subclinical involvement of eye movements in patients with MG.

Classification and staging of Parkinson's disease using video-based eye tracking

Introduction83% of those diagnosed with Parkinson's Disease (PD) eventually progress to PD with mild cognitive impairment (PD-MCI) followed by dementia (PDD) - suggesting a complex spectrum of pathology concomitant with aging. Biomarkers sensitive and specific to this spectrum are required if useful diagnostics are to be developed that may supplement current clinical testing procedures. We used video-based eye tracking and machine learning to develop a simple, non-invasive test sensitive to PD and the stages of cognitive dysfunction. MethodsFrom 121 PD (45 Cognitively Normal/45 MCI/20 Dementia/11 Other) and 106 healthy controls, we collected video-based eye tracking data on an interleaved pro/anti-saccade task. Features of saccade, pupil, and blink behavior were used to train a classifier to predict confidence scores for PD/PD-MCI/PDD diagnosis. ResultsThe Receiver Operator Characteristic Area Under the Curve (ROC-AUC) of the classifier was 0.88, with the cognitive-dysfunction subgroups showing progressively increased AUC, and the AUC of PDD being 0.95. The classifier reached a sensitivity of 83% and a specificity of 78%. The confidence scores predicted PD motor and cognitive performance scores. ConclusionBiomarkers of saccade, pupil, and blink were extracted from video-based eye tracking to create a classifier with high sensitivity to the landscape of PD cognitive and motor dysfunction. A complex landscape of PD is revealed through a quick, non-invasive eye tracking task and our model provides a framework for such a task to be used as a supplementary screening tool in the clinic.