Screen-based Visual Impairment Simulation: An Exploratory Investigation of Acuity Loss Simulation Using Gaussian Blur Filters

Screen-based Visual Impairment Simulation: An Exploratory Investigation of Acuity Loss Simulation Using Gaussian Blur Filters | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Screen-based Visual Impairment Simulation: An Exploratory Investigation of Acuity Loss Simulation Using Gaussian Blur Filters Wai Lam Leung, Oliver Runswick, Claire Heard, Tim Rakow This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7957115/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Purpose Accessible and scalable tools for simulating visual impairment are essential for studying accessibility and human-computer interaction under visual constraints. This study assessed whether Gaussian blur applied to electronic displays can serve as a valid and practical alternative to traditional simulation spectacles for inducing visual acuity loss. Method Thirty-seven participants completed and compared standard visual-acuity assessments using ETDRS charts and a screen-based tumbling-E test displayed on monitors with three pixel densities (94, 170, and 218 PPI). Tests were conducted under normal vision, moderate impairment, and severe impairment induced by simulation spectacles. Participants then performed 29 screen-based tumbling-E tests incorporating varying Gaussian-blur standard deviations (σ) to discover the relationship between σ and measured visual acuity loss. Results Gaussian blur strength exhibited a strong log-linear relationship with measured visual-acuity loss, with minimal influence from monitor pixel density. A pooled regression model explained over 90% of the variance in visual acuity and accurately predicted impairment levels across display types at different σ. Conclusion Screen-based Gaussian blur offers a valid and reproducible method for simulating visual acuity loss. The derived regression model enables estimation of the blur level (σ) corresponding to specific degrees of impairment, facilitating consistent and controllable simulation across studies. Translational Relevance This approach supports remote experimentation and low-cost accessibility testing, providing researchers and designers with a scalable tool to investigate perceptual and attentional processes under degraded visual conditions. Psychology visual acuity simulation low vision accessibility psychophysics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction International Agency for the Prevention of Blindness demonstrated that there are 1.1 billion people globally with vision loss, and it is projected to increase to 1.7 billion by 2050 ( 1 ). Various daily activities can be difficult for these individuals, including managing finances online, online shopping, navigation using apps, video calls, reading digital documents ( 2 – 4 ), and even assistive technologies designed to aid accessibility because many developers lack knowledge of accessibility ( 5 ). Consequently, researchers have sought to understand the nature and consequences of vision impairment and to inform, develop and test mitigations for its adverse consequences. One cost-effective methodology used to understand the impacts of vision impairment and the accessibility of products (e.g., websites, navigation systems, accessibility tools) is simulating the loss of visual function in sighted individuals ( 6 – 8 ). Simulation methods cannot replicate the lived experiences of individuals with vision impairment because sighted individuals can foveate the artificially degraded images and use their normally functioning visual pathways. However, simulations can facilitate experimental control by employing within-subject designs with different simulated degrees of impairment and increase recruitment by allowing less restrictive inclusion criteria. To this end, simulation methods have been used to understand the impact of vision loss in tasks ranging from everyday living to playing sports ( 9 – 11 ). A common approach to simulating vision impairment is via simulation spectacles that use different filters to induce varying degrees of acuity loss, a reduction in sharpness and clarity of vision ( 6 , 12 , 13 ) valuable for research, simulation spectacles have some limitations. Remote testing without the researcher present is impractical, participants are still able to move their eye behind the simulation spectacle without the impairment remaining in the same place in the visual field, and eye tracking is generally not possible because the eye is not visible. Modern alternatives include the use of virtual reality simulations with gaze-contingent simulation of impairment ( 9 ), but these approaches again require in-person testing and significant efforts in stimuli creation. For these reasons and others, we explored and tested an alternative approach to simulating vision loss, which involves applying blur to screen images. Such simulation via an electronic display could address the abovementioned limitations: reduce physical labour load for data collection, facilitate larger sample sizes by allowing remote testing with many individuals participating concurrently and could allow for gaze tracking when this is valuable for the research question. Additionally, compared to simulation spectacles, electronic displays are better for producing consistent and accurate simulations of visual defects ( 6 ). The current experiment tests this image-blur method for simulating (different degrees of) visual acuity loss. Loss of visual acuity is the most common form of visual impairment; it is associated with several common pathologies ( 14 ) and ageing ( 15 – 17 ); and is highly correlated with other impairments such as contrast sensitivity ( 18 ). The World Health Organisation defines vision impairment through measures of acuity where mild, moderate and severe impairment are defined as 0.3 to 0.5, 0.5 to 1.0, and 1.0 to 1.3 logMAR (logarithm of the Minimum Angle of Resolution, a scale used to measure visual acuity, where lower values indicate better vision), respectively. We selected Gaussian blur as the sole method for simulating acuity loss because this image-processing technique is widely available and should be relatively straightforward for researchers to implement. Gaussian blur is an image processing technique that reduces image noise and detail by averaging the colour or intensity values (for greyscale images) of each pixel with its surrounding pixels using the Gaussian function, which smooths out the images to create a blurry and softened effect. Gaussian blur has previously been applied via more complex algorithms to simulate impairment in visual field and acuity ( 6 ). The degree (width) of blurring is adjusted by changing the value of the standard deviation (σ) of the Gaussian function, often referred to as the radius or sigma, where the larger the standard deviation, the greater the blur. In the following equation that defines image blur, x and y are the coordinates of a point in the 2D plane, and e is Euler’s number: $$\:G\left(x,y\right)=\:\frac{1}{2\pi\:{\sigma\:}^{2}}{e}^{-\frac{{x}^{2}+{y}^{2}}{2{\sigma\:}^{2}}}$$ The primary aim of our experiment was to determine the appropriate σ for the Gaussian function to simulate specific levels of acuity loss. Therefore, this experiment aimed to understand the relationship between Gaussian blur across different values of σ and the resulting functional impairment as measured by vision tests. By developing a regression model that maps the σ value to an estimated level of visual impairment, we provide guidance for using simple Gaussian blur in well-controlled research on visual impairment. Because variations in electronic displays could impact the precision of the simulation, we also explored the impact of these display variations, focusing primarily on differences in screen pixel density. Our goal was to explore whether a single regression model could be used to identify simulation settings to be used across different types of screens or whether substantially different models are required for different displays. To achieve our primary objective, we first must ensure that visual acuity can be measured accurately using computer screens. To this end, participants completed visual acuity tests using the ETDRS (Early Treatment Diabetic Retinopathy Study) on a physical chart and a newly developed vision test emulated the ETDRS letters on three types of computer screens (94/170/218 PPI [pixel per inch]). To further ensure the reliability of capturing visual acuity using screens, test results from screen-based vision tests were compared to physical testing at two distances, three and four metres, and under three different vision conditions. Participants completed those screen-based and physical vision tests without wearing simulation spectacles and with simulation spectacles. These spectacles simulated moderate and severe acuity loss to ensure the screen-based vision test can still measure visual acuity in low vision individuals to the same extent as physical charts Following this initial validation phase, we then applied different levels of Gaussian blur to the screens during the screen-based vision test to map the degree of acuity loss per increment of blur. Regression models were developed from these data to map out the relationship between blur strength (σ) and visual acuity that can be used to estimate the value of σ for specific levels of simulated acuity loss. This mapping process was conducted independently on three computer screens to examine whether differences in PPI impact the relationship between Gaussian blur applied on screen and visual acuity, in order to determine whether the simulation technique can be used generally across a range of screens with vary PPI or if each screen requires its own calibrated relationship between blur and visual acuity. Overall, the current experiment aims to investigate whether screen-based acuity tests are a valid approach to vision testing, which in turn allows for the generation of regression models that can be used to estimate the appropriate level of Gaussian blur for specific levels of simulated acuity loss for future in-person and remote simulation studies. Method Participants A total of 37 volunteers participated (8 male, 28 female, and 1 ‘prefer not to say’) with ages ranging from 18 to 23 years ( M = 18.8, SD = 1.0). Participants were first-year psychology undergraduates, recruited via an online participant management platform. They participated in the study for course credit. Informed written consent was obtained from all participants before the study, ensuring they were fully aware of the procedures and their rights. The study received ethical approval from the local ethics committee (Ethics Registration Number: MRSP-23/24-40842). Material ETDRS charts. Two early treatment diabetic retinopathy study (ETDRS) charts by Precision Vision were used, one designed for testing at 4 metres and one at 3 metres. The testing range of both charts is from the logMAR score of 1.0 to -0.4. Simulation software. Simulation software was developed to apply Gaussian blur from the SciPy python package ( 19 ) onto tumbling E eye tests that measure logMAR scores. The letter size adheres to a conventional 4-metre (testing distance) ETDRS chart, where the letter for logMAR = 1.0 is of the size of 5.82 x 5.82 cm. Compared to the ETDRS charts , four rows of letters above the logMAR = 1.0 row and eight rows below the logMAR = -0.4 are added using the standard letter size increment/decrement of 0.1 log units (≈ 1.26x) ( 20 ). LogMAR score calculations are automated and exported into a CSV file. Simulation spectacles. Two pairs of simulation spectacles, adapted from “Cambridge simulation glasses” ( 21 ) created by Runswick and colleagues using Hampshire frost filters were acquired for the experiment ( 11 ). The level four (logMAR mean = 1.21, SD = ± 0.23 [Range: 0.88–1.60]) and six (logMAR mean = 1.84, SD = ± 0.18 [Range: 1.60–2.50]) spectacles are used and labelled as medium and severe acuity loss spectacles, respectively. Monitors . Three different monitors were used, each varying in pixels per inch (PPI), a measure of pixel density in digital displays and a higher PPI indicates more pixels per inch. The monitors used were: 1) iMac with Retina 5k display at 218 PPI, 2) Acer Predator Helios 16 at approximately 170 PPI, and 3) Samsung CF396 24” monitor at 94 PPI. All screens are adjusted to 330 lux units to adhere to the same emitted lux by the ETDRS charts, measured by a lux meter hovering directly above the screens. Procedure The experiment was segmented into two parts, the first part being the validation phase , where participants completed tests on the ETDRS charts and simulation software with habitual vision, moderate, and severe vision induced by simulation spectacles to validate the accuracy of doing vision tests on multiple monitors compared to conventional charts at varying levels of acuity loss. In the second part, the mapping phase , participants completed a series of vision tests on the simulation software with Gaussian blurs with incremental strength (i.e., standard deviation; σ) applied on the screen to map out the relationship between blur strength and acuity loss. The mapping phase is of mixed design, with the independent variable of monitors (218/170/94 PPI) being the between-subject factor and Gaussian blur strength (σ = 0 to 10 with the increment of 0.5 and σ = 20 to 90 with the increment of 10) being the within-subject factor, and the dependent variable being the logMAR measured for each Gaussian blur strength level. Validation phase. Participants were instructed to wear or not wear simulation spectacles and complete binocular vision tests on the ETDRS charts (4 and 3 metres) and simulation software until all vision testing methods ( ETDRS charts at 3 metres, 4 metres, or simulation software ) were completed on all simulation spectacles levels (within-subjects: no, medium, and high severity). Overall, each participant completed 3 simulation spectacles x 3 testing methods = 9 vision tests (see Fig. 1 ). The order in which those combinations of simulation spectacles and testing methods were completed was randomised. For the ETDRS charts trials, participants were instructed to read aloud from the top row to the bottom. The test terminates when errors are made on three or more letters in a given row, and then the logMAR score is calculated. If participants cannot read the top row of the charts, their testing distance is halved, and a 0.3 logMAR score is added to their total score. For the simulation software trials, participants were tested at a 1-metre distance using a tumbling E vision test on their randomly allocated monitor. Since the selected letter size (5.82 x 5.82 cm) is according to the 4 metres ETDRS chart , using the testing-distance principle described above, a 0.6 logMAR score was added to the participants’ logMAR score due to halving twice from 4 metres. Participants used four directional arrow keys on a keyboard to indicate which direction the tumbling E is facing, based on the direction of the opening of the letter E (e.g., press up-arrow if the E’s opening is facing upward). Each test is terminated when three or more incorrect responses are given in a row. Mapping Phase. Participants completed a series of tumbling E vision tests on the simulation software with Gaussian blur applied on the screen. The settings, scoring, and instructions were identical to those used in the validation phase . In a pilot study ( N = 10), a logarithmic trend was observed in the relationship between the change in logMAR score and Gaussian blur strength (σ). Specifically, increases in σ at higher levels of blur resulted in minimal changes to the logMAR score, whereas increments in σ at lower levels of blur produced a steep increase in the logMAR score. Therefore, to avoid fatigue effects from testing an excessive number of levels, the tested blur strength was from σ = 0 to 10 with an increment of 0.5, and from σ = 20 to 90 with an increment of 10. This resulted in 29 separate vision tests, each with a different level of Gaussian blur applied; their presentation order was randomised (see Fig. 2 ). Results The data for all 37 participants were analysed using Bland-Altman analysis, which involves plotting differences in logMAR scores between the simulation software against ETDRS charts at 3 metres and (separately) at 4 metres (Fig. 3 , 4 , 5 & Table 1 ). This analysis assesses the agreement between the simulation software and the ETDRS charts in measuring visual acuity to identify systematic biases or variability across different conditions and distances. The 95% CI for the mean difference in logMAR shows that the systematic biases in the control and severe simulation spectacles condition are relatively small, with narrow confidence intervals across monitors and ETDRS charts (see Table 1 ). However, there is a noticeable positive bias for medium simulation spectacles across monitors and ETDRS charts , particularly for the ETDRS chart at 3 metres with the highest mean difference at 0.09 [95% CI: 0.05, 0.13] logMAR. Nevertheless, the limits of agreement indicate acceptable variability, aligning with previous research comparing ETDRS charts to other validated vision tests ( 22 ). Importantly, this finding falls below the 0.2 logMAR threshold typically considered a meaningful change ( 23 ). Moreover, this trend is consistent across monitors , suggesting good agreement across monitors , though with a slightly larger variability of measured logMAR for the medium simulation spectacles across all three monitors . Table 1 Bland-Altman Agreement: Mean Difference in logMAR Between Simulation Software on 218, 170 and 94 PPI screens and ETDRS Charts (3 and 4 Metres) Pixel Per Inch (PPI) Simulation Spectacles Mean difference [95% CI] ETDRS (3-metres) ETDRS (4-metres) 218 PPI (n = 13) Control -0.058 [-0.10, -0.01] 0.015 [-0.15, 0.18] Medium 0.090 [0.05, 0.13] 0.064 [0.004, 0.12] Severe 0.002 [-0.05, 0.05] 0.005 [-0.04, 0.05] 170 PPI (n = 12) Control -0.019 [-0.06, 0.02] -0.017 [-0.07, 0.04] Medium 0.088 [0.04, 0.13] 0.084 [0.04, 0.12] Severe 0.053 [0.01, 0.09] 0.047 [0.006, 0.08] 94 PPI (n = 12) Control -0.026 [-0.07, 0.01] -0.03 [-0.07, 0.009] Medium 0.033 [-0.02, 0.09] 0.018 [-0.05, 0.08] Severe 0.012 [-0.03, 0.05] -0.007 [-0.03, 0.02] The logMAR score change was calculated by subtracting the logMAR score obtained at σ = 0 from logMAR scores obtained from σ > 0. A mixed-effects log-linear regression model was fit by REML, where the dependent variable was the logMAR score change, and the independent variables were monitor’s PPI and Gaussian blur strength (σ) with random effects of different participants accounted for. Note that the PPI variable was treated as categorical in all analyses since the difference between screens may involve more than just a difference in pixel density (e.g., contrast). A total of 1023 data points were analysed after 13 were removed due to misinput during the experiment by participants (i.e., stuck keys on the keyboard and accidental key holds that led to completion of trials despite no input from participants). The mixed-effects model’s formula is as follows: $$\:logMAR\:Change\sim\:PPI\:+\text{log}\left(Gaussian\:Blur\:\sigma\:\right)+(PPI\:\times\:\text{log}\left(Gaussian\:Blur\:\sigma\:\right))+(1\left|Participant\right)$$ The conditional R² indicated that fixed and random effects explained 95.19% of the variance, while the marginal R² showed that the fixed effects alone explained 90.71%. A type III ANOVA with Satterthwaite’s method found significant effects for the main effect of monitor PPI, F (2, 44.88) = 3.91, p = 0.02, \(\:{\omega\:}^{2}\) = 0.15, 95% CI [0.00, 1.00], the main effect of blur, F (1, 983.22) = 19590.73, p < .001, \(\:{\omega\:}^{2}\) = 0.95, 95% CI [0.95, 1.00], and the interaction term, F (2, 983.22) = 12.13, p < .001, \(\:{\omega\:}^{2}\) = 0.02, 95% CI [0.01, 1.00] (see Fig. 6 ). Pairwise comparison between monitors with Bonferroni correction found that the screen with 94 PPI Samsung and 218 PPI iMac monitors had significantly higher increase of logMAR compared to the 170 PPI Acer monitor, t (35.5) = 4.299, p < .001, SE = 0.03, 95% CI [0.07, 0.27], and, t (35.5) = 4.63, p < .001, SE = 0.03, 95% CI [0.27, 0.08], respectively. However, the 94 PPI Samsung and 218 PPI iMac monitors did not significantly differ from each other, t (35.6) = -0.25, p = 1.00, SE = 0.03, 95% CI [-0.10, 0.08]. As the interaction term was significant in the omnibus test, it suggests that blur may perceived differently based on display type; therefore, a comparison of slopes was conducted to examine the extent of such effect. The analysis found that the slope of log(Gaussian Blur σ) for the 94 PPI Samsung monitor was b = 0.416, SE = 0.005, 95% CI [0.40,0.42], df = 984, for the 170 PPI Acer monitor was b = 0.381, SE = 0.005, 95% CI [0.37, 0.39], df = 983, and for the 218 PPI iMac monitor was b = 0.404, SE = 0.004, 95% CI [0.39, 0.41], df = 983. Such findings suggests that the log slopes of σ differ across screen levels, but the differences are quite small in absolute terms. Specifically, the difference between 94 PPI Samsung and 170 PPI Acer monitors is only 0.035, and between the 170 PPI Acer and 218 PPI iMac monitors is 0.023. Given that the slopes are within a narrow range and the confidence intervals are quite narrow, it suggests that the relationship between log(Gaussian Blur σ) and logMAR change is relatively consistent across screen levels. Practically speaking, when Gaussian Blur’s σ is at 2, 10, and 20, the logMAR change for each screen according to the slope analysis corresponds fairly well to the classification of mild, moderate, and severe vision impairment, respectively (Table 2 ). Table 2 Comparison of Predicted logMAR Change Across Different levels and Screen Types σ 94 PPI Samsung 170 PPI Acer 218 PPI iMac σ = 2 0.288 0.264 0.280 σ = 10 0.958 0.877 0.930 σ = 20 1.246 1.141 1.210 Coupled with the small effect size of the interaction (ω² = 0.02), these findings suggest that screen PPI or monitor type is unlikely to be of practical concern for monitors of comparable specification to those that we tested. However, it is important to note that as σ increases, the magnitude of differences in logMAR change also increases. Nonetheless, this effect remains of limited practical concern for simulations ranging across mild to severe acuity loss, because the differences do not exceed the 0.2 logMAR threshold for meaningful changes in visual acuity ( 23 ). Therefore, we opted for a pooled model that disregards the fixed effect of screen PPI or monitor type to simplify future implementation of this method (see Fig. 7 ). The pooled regression model was statistically significant, F (1, 1021) = 10180, p < .001 with adjusted \(\:{R}^{2}\) at 0.91. The main effect of log-transformed blur was significant at t (1021) = 100.91, p < .001, indicating a strong positive relationship between blur strength and logMAR score change. The high adjusted \(\:{R}^{2}\) (0.91) suggests that the regression model (without screen PPI or monitor type included) should provide a reliable guide for future application of the blur simulation method.. To further affirm the use of the pooled model instead of the more complicated mixed-effects model, we employed bootstrapping techniques of 1000 samples referencing data from the 170 PPI Acer monitor to assess the prediction accuracy of both models (pooled and mixed-effects model) using Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and R-squared (R²). The observed difference between the two models is summarised in Table 3 , where the differences were minimal. For instance, RMSE and MAE for both models were almost identical, suggesting that the predictions from the pooled model would only make a 0.001 logMAR difference (RMSE [Mean]: 0.152–0.151 = 0.001; MAE [Mean]: 0.116–0.117 = 0.001) compared to the mixed-effects model. Although the R² value demonstrated a higher divergence (0.08), the minimal incremental improvement for the mixed-effects model does not justify the added complexity. Table 3 Mean Square Error (RMSE), Mean Absolute Error (MAE), and R-squared (R²) between Pooled and Mixed-effects models from 1000 bootstrapping samples. Model Metric Mean 2.5% CI 97%CI Pooled RMSE 0.152 0.152 0.154 MAE 0.116 0.111 0.122 R² 0.918 0.918 0.918 Mixed-effects RMSE 0.151 0.151 0.152 MAE 0.117 0.115 0.119 R² 0.926 0.925 0.926 Sensitivity Analysis of Pooled vs. Mixed-effects Model To further examine the sufficiency of the pooled model, we also investigated the prediction difference of simulated logMAR given a specific σ between models. One example scenario is portrayed in Fig. 8 , where we used the pooled model for inverse prediction to estimate the appropriate σ to simulate moderate vision impairment (logMAR = 0.5–1.0); the targeted logMAR was 0.75, the middle-point of the classification to maximise the acceptable range of error (± 0.25 logMAR; Fig. 8 A). In Fig. 8 B, the logarithmic-like red, green, and orange lines with lower opacity represent the prediction difference of logMAR between the pooled model and the mixed-effects model for 94, 170 and 218 PPI screens. Furthermore, the orange vertical line represents the predicted σ value from the pooled model to simulate 0.75 logMAR, being σ = 7.3. When applying σ = 7.3 to the mixed-effects model, we can see that compared to the pooled model, 170 PPI screens predicted 0.11 logMAR less than the pooled model, meaning that applying σ = 7.3 to Gaussian blur on the 170 PPI screen would result in 0.64 logMAR instead of 0.75 logMAR predicted by the pooled model. Since the acceptable range of error for moderate classification of vision impairment is ± 0.25 logMAR (orange dotted line in Fig. 8 B), the simulation would be deemed appropriate as the prediction error between the two models does not exceed the acceptable range (i.e., the orange curved line does not touch or exceeds the orange dotted line at σ = 7.3). Moreover, even at a narrower range of acceptable errors, such as mild (± 0.10 logMAR) and severe (± 0.15 logMAR) vision impairment simulation, the error between the model does not exceed the boundary at σ = 2.4 (green vertical line) and σ = 19.8 (red vertical line). Regardless, there is a logarithmic increase in the prediction error of logMAR as σ increases, meaning the accuracy of the simulation may be lowered when the user intended to simulate a higher severity of acuity loss, which should be noted if precise simulation is required for future studies that aim beyond the simulation of vision impairment classifications from the International Classification of Diseases. Discussion To summarise, the experiment aimed to examine whether screen-based acuity tests are a valid approach to vision testing, which in turn allows for the generation of regression models that can be used to estimate the appropriate level of Gaussian blur for specific levels of simulated acuity loss for future in-person and remote simulation studies. Our findings suggested good agreement between screen-based acuity tests and physical ETDRS charts at different distances and vision characteristics (habitual vision, moderate and severe acuity loss via simulation spectacles). Affirmation of the validity of screen-based acuity tests then allowed us to create a simple regression model that can be used to infer the appropriate level of Gaussian blur to simulate a given level of impact from acuity loss. This will facilitate remote simulation studies. The results indicate that this is feasible, given the significant log-linear relationship between the standard deviation (σ) of the Gaussian function and the measured logMAR under the influence of Gaussian blur. Specifically, a slight increase in σ before σ = 10 resulted in a steep rise in logMAR. According to the classification of vision impairment from the International Classification of Diseases ( 24 ), it shows that the simulation of mild (logMAR = 0.30–0.50) and moderate (logMAR = 0.50–1.00) vision impairment falls within the steep incline which could make simulation of mild and moderate acuity loss more challenging than severe due to the need for precise control over σ. The resulting equation for estimating the required 𝜎 for a given desired level of logMAR change (denoted as 𝑥) is: $$\:\sigma\:=\:{e}^{\frac{x+0.04533908\:}{0.4001204}}$$ Sensitivity analyses were conducted that supported using the pooled model over the mixed-effects model. While we initially anticipated that PPI might drive the differences between screens, our findings suggest this is unlikely, as the highest and lowest PPI screens showed nearly identical effects. Such observation indicates that PPI was not a monotonic factor in mapping blur to logMAR change. Instead, it is plausible that other monitor features, such as those that impact contrast (e.g., the type of backlighting), are responsible for the differences we observed. For example, the 170 PPI Acer monitor (marketed as a gaming laptop) uses advanced mini-LED backlight technology that provides a higher contrast ratio and deeper black visuals ( 25 ) compared to the other two screens, which use more traditional LED backlighting methods. The advanced technology should create sharper outline of letters in the eye test which in turn results in a lower logMAR score. In practical terms, the effect of Gaussian blur was much larger than that of screen PPI or monitor type, suggesting that the monitor’s properties can be ignored in future simulation studies without seriously impacting design decisions. Nonetheless, those who prioritise precision (e.g., for small manipulations of acuity) may find it necessary to minimise variation in screen types among participants. This is especially pertinent for gaming laptops and monitors, which may have features designed to mitigate on-screen blur. Regardless, future research should validate the blur method across a broader range of screen types to investigate how factors like contrast ratio and the technologies that influence it might impact the application of Gaussian blur to images for simulations of acuity loss. Applications Employing our simplified model to simulate vision impairment using Gaussian blur could significantly enhance product accessibility within the technology industry. Despite efforts to make technology and essential services accessible to individuals with vision impairments, research indicates that many products remain inaccessible due to inadequate research on user needs and poor implementation of user-centred design ( 3 , 4 ). Many developers admit to a lack of accessibility education, resulting in inaccessible software and websites, leading to an estimated annual revenue loss of approximately 16 billion USD, with only 3.7% of the world’s top one million websites being fully accessible ( 5 ). Malkawi and colleagues ( 26 ) report that assistive technologies for people with vision impairment are often developed based on researchers’ personal knowledge without involving target groups during design phases. Beyond the lack of target group involvement, scientists are often insufficiently involved in technological design to carry out experiments that require knowledge and skillsets beyond developers’ educational background and expertise ( 27 ). In summary, many developers lack accessibility education and may be hesitant or unable to involve target groups in the design phase. They may also lack the resources to hire experts for specialised accessibility research on vision impairment. However, many of these challenges could be addressed by offering a low-labour, cost-effective research method, such as simulation research informed by our regression model, for controlled testing that informs the impact of vision impairment on the accessibility of assistive technologies during design phases for refinement before a more well developed product is then tested by those with vision impairment. Alternatively, this approach could incentivise remote simulation research, serving as a viable alternative to resource-intensive in-person experiments due to the required time, specialised equipment, and effort in getting participants to the lab. In prior research, Gaussian blur has been applied to more complex algorithms to simulate vision impairment, primarily in visual field simulation ( 6 ) and gaze-contingent paradigms ( 28 ). However, our findings point to simulation using simple Gaussian blur as an accessible alternative with several benefits. Firstly, implementing a more sophisticated simulation method may require a strong understanding of specific coding environments, such as MATLAB ( 6 ). These may not be familiar to all developers or researchers and require extensive training. Secondly, even if developers or researchers can operate the complex simulation environment, integrating sophisticated algorithms into websites or browser-based experimental platforms for convenient data collection could be challenging and require significant resources. Packaging these algorithms into software for installation to a research participant’s device could limit the sample pool to individuals with sufficient technical skills or confidence, thereby reducing participation among key groups (e.g., older adults). This means that it could pose threats to cross-sectional studies sensitive to age-related differences. In contrast, simple Gaussian blur filters are accessible and widely available in different coding environments (e.g., OpenCV [Python], Python Imaging Library [Python], Canvas API [JavaScript], imgaussfilt() [MATLAB]) as pre-installed and additional packages, including web development, and researchers who use web-based survey engines for experimentation can now easily apply simple Gaussian blur over images, text and videos to simulate acuity loss. For those interested in using Gaussian blur to simulate acuity loss, we have developed a web-based companion tool ( https://wailamleung.shinyapps.io/g-balsac/ ). This tool provides guidance on setting Gaussian blur parameters using one of the four available models (pooled model plus individual models based on data from each monitor we tested) to achieve a desired level of induced acuity loss (σ estimation) based on test subject(s) baseline visual acuity. It also supports inverse estimation, allowing users to convert σ values into logMAR for easy interpretation. Furthermore, the current findings could benefit future research. Take Moharrer et al ( 7 ) as an example, their research investigated the roles of motion perception and visual acuity in driving hazard perception where participants were equipped with simulation spectacles to simulate acuity loss (i.e., 20/120) and instructed to watch driving videos and presses a key as soon as hazard are recognised. Such an experiment can be streamlined further by applying blur onto the screen, avoiding the need for simulation spectacles and providing more accessible simulation conditions, which reduces labour and costs associated with acquiring spectacles. Limitations Two notable factors may threaten the validity of simulations, especially in the context of remote studies. First, the brightness of screens, which research will have no direct control over when experiments are conducted remotely, could cause deviation in the perception of blur among participants. The current experiment controlled for brightness across the three screens with different PPI at 330 lux and approximately the maximum brightness for each screen, meaning that the relationship between σ and acuity loss may differ at screen brightness below or above 330 lux and the extent of its impact on acuity loss simulation is yet to be investigated. However, some studies suggest that brightness is not a concern when measuring acuity. Take Cheng et al as an example ( 29 ); they found that there was a slight difference in visual acuity at different screen brightness levels at 50% and 100%, but the difference is equivalent to less than one letter on the visual acuity chart (-0.009 logMAR). Similarly, Nkeremuzor et al ( 30 ) demonstrated that not with electronic displays but with fluorescent and tungsten bulb-illuminated charts, where brightness ranges from 13.28 cd/m² to 14.64 cd/m² and 14.68 cd/m² to 15.76 cd/m², respectively, acuity measurements within and between fluorescent and tungsten bulbs were not statistically significant. These findings suggest that variations in screen brightness are unlikely to have a major impact on the accuracy of acuity measurement. Nonetheless, the addition of simple Gaussian blur may lead to an unforeseen interaction that says otherwise. Therefore, additional steps to control screen luminance may be wise. For example, participants could use smartphone applications to measure lux levels, which have been shown to demonstrate decent measurement agreement with traditional lux meters with appropriate settings ( 31 ), to ensure their electronic display can reach 330 lux and be asked to set screen brightness accordingly. Regardless, future research could investigate the effect of screen luminance on screen-based vision tests, including under the context of Gaussian blur to validate the present methodology further. Second, sitting distance or leaning habits are another uncontrollable factor in a remote setting. The current experiment controlled for sitting distance at a 1-metre distance and actively prevented leaning forward or backwards from the screen during data collection. However, without supervision, this may not always be the case. When moving closer to the screen, the blurred pixels may appear more pronounced as you see blurred pixels at a larger size. Meanwhile, blurring might become less noticeable when farther from the screen. Under the circumstances of clear instructions for participants to sit 1 metre away from their electronic display, considering the substantial prevalence of forward head posture (associated with leaning towards screens) ( 32 ), we can speculate an increased blurriness of stimuli in remote settings without supervision. However, the extent of sitting distance or head posture’s impact on perceived blur on electronic display remains unknown, which should be further investigated. Finally, using simple Gaussian blur as a simulation method of acuity loss by no means fully captures the complexities of vision impairment. Vision impairment often involves not only acuity loss but also varying levels of contrast sensitivity, visual field loss, and other factors that can interact in ways not explored in the present research. Using simple Gaussian blur is just a systematic means of reducing visual information to provide an approximate insight into how individuals with one kind of vision impairment perceive on-screen content. Therefore, whatever the merits of simulating vision impairment, it should not replace testing with individuals with vision impairments. Conclusion The current experiment demonstrated that simple Gaussian blur applied to electronic displays is a viable method for simulating visual acuity loss, offering a cost-effective and accessible alternative to resource-intensive simulation methods. The findings indicated a strong log-linear relationship between the degree of Gaussian blur and acuity loss, allowing precise control over simulating specific levels of acuity loss. Using the experimental data, we developed a pooled regression model that can guide the design of simulation experiments for acuity loss. The experiment provided proof-of-concept, suggesting that simulation studies using Gaussian blur applied to on-screen images have substantial potential for applied and translational research into visual impairment and for technical work to enhance the accessibility of technology and services. Regardless, future validations of this method should examine the impact of the factors that are not easily controlled in unsupervised settings (e.g., screen brightness, sitting distance) to further refine the value of the simulation method by developing necessary protocols. Declarations This study was reviewed and approved by Research Ethics Office in King’s College London with the approval number: MRSP-23/24-40842, dated 04/01/2024. Participation was voluntary, and written informed consent was obtained from all individual participants included in this study. Funding Information This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Acknowledgement N/A References International Agency for the Prevention of Blindness [Internet]. International Agency for the Prevention of Blindness (2021) Available from: https://www.iapb.org Remillard ET, Koon LM, Mitzner TL, Rogers WA (2024) Everyday Challenges for Individuals Aging with Vision Impairment: Technology Implications. Gerontologist 64(6):169 Franco M, Gaggi O, Merzougui SE, Palazzi CE (2023) Accessible Wayfinding for the Visually Impaired through Sustainable Smartphone Based Sensing. In: 2023 IEEE 20th Consumer Communications & Networking Conference (CCNC. IEEE; pp. 1–6 Ortiz-Escobar LM, Chavarria MA, Schönenberger K, Hurst S, Stein MA, Mugeere A et al (2023) Assessing the implementation of user-centred design standards on assistive technology for persons with visual impairments: a systematic review. Front Rehabilitation Sci 4:1238158 Parthasarathy PD, Joshi S (2024) Exploring the Need of Accessibility Education in the Software Industry: Insights from a Survey of Software Professionals in India. In: Proceedings of the 46th International Conference on Software Engineering: Software Engineering Education and Training [Internet]. pp. 212–20. Available from: https://doi.org/10.1145/3639474.3640079 Abraham CH, Sakyi-Badu G, Boadi‐Kusi SB, Morny E, Darko‐Takyi C, Ocansey S et al (2024) Simulation of visual impairment in persons with normal vision for scientific research. Ophthalmic Physiol Opt 44(2):442–456 Moharrer M, Tang X, Luo G (2020) With motion perception, good visual acuity may not be necessary for driving hazard detection. Translational Vis Sci Technol 9(13):18–18 Peters B, Bedrick S, Dudy S, Eddy B, Higger M, Kinsella M et al (2020) SSVEP BCI and eye tracking use by individuals with late-stage ALS and visual impairments. Front Hum Neurosci 14:595890 Reddingius P, Crabb DP, Jones P (2024) Shopping with sight loss: Using a virtual reality shopping task to systematically quantify the impact of a scotoma size and eccentricity on everyday quality of life. Investig Ophthalmol Vis Sci 65(7):1849–1849 Mann DL, Ravensbergen HJC (2018) International Paralympic Committee (IPC) and International Blind Sports Federation (IBSA) joint position stand on the sport-specific classification of athletes with vision impairment. Sports Med 48(9):2011–2023 Runswick OR, Rawlinson A, Allen PM, Sharpe BT, Pocock C, Datson N et al (2023) The effects of simulated vision impairment on performance in football. J Sports Sci 41(14):1410–1422 Heinrich SP, Strübin I (2019) Use of diffusing filters for artificially reducing visual acuity when testing equipment and procedures. Doc Ophthalmol 140(1):83–93 Hecht H, Hörichs J, Sheldon S (2015) The effects of simulated vision impairments on the cone of gaze. Atten Percept Psychophys 77:2399–2408 Kim SY, Moon BY, Cho HG, Yu DS (2022) Visual Acuity Outcomes in Diseases Associated with Reduced Visual Acuity: An Analysis of the National Health Insurance Service Database in Korea. Int J Environ Res Public Health 19(14):8689 Al-Namaeh M (2021) Common causes of visual impairment in the elderly. Medical Hypothesis. Discovery Innov Ophthalmol 10(4):191 Ehrlich JR, Agarwal A, Young C, Lee J, Bloom DE (2022) The prevalence of vision impairment and blindness among older adults in India: findings from the Longitudinal Ageing Study in India. Nat aging 2(11):1000–1007 Vinding T (1990) Visual impairment of age-related macular degeneration: An epidemiological study of 1000 aged individuals. Acta Ophthalmol 68(2):162–167 Schneck ME, Haegerstrom-Portnoy G, Lott LA, Brabyn JA, Gildengorin G (2004) Low contrast vision function predicts subsequent acuity loss in an aged population: the SKI study. Vision Res 44(20):2317–2325 Virtanen P, Gommers R, Oliphant TE, Haberland M, Reddy T, Cournapeau D et al (2020) SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat Methods 17(3):261–272 Bailey IL, Jackson AJ (2016) Changes in the clinical measurement of visual acuity. Journal of Physics: Conference Series. ;772:012046 Goodman-Deane J, Waller S, Collins AC, Clarkson J (2013) Simulating vision loss: what levels of impairment are actually represented? In: Contemporary Ergonomics and Human Factors 2013: Proceedings of the international conference on Ergonomics & Human Factors 2013. Cambridge, UK: CRC Press; p. 347 Bastawrous A, Rono HK, Livingstone IA, Weiss HA, Jordan S, Kuper H et al (2015) Development and validation of a smartphone-based visual acuity test (peek acuity) for clinical practice and community-based fieldwork. JAMA Ophthalmol 133(8):930–937 Lovie-Kitchin JE (1988) Validity and reliability of visual acuity measurements. Ophthalmic Physiol Opt 8(4):363–370 Organization WH (2019) Available from: https://www.who.int/publications/i/item/9789241516570 Ridley J (2023) Acer Predator Helios 16 [Internet]. PC Gamer; Available from: https://www.pcgamer.com/acer-predator-helios-16-review-benchmarks/ Malkawi A, Kamaruddin A, Halin AA, Admodisastro N (2023) Listening to the Voice of People with Vision Impairment. Int J Adv Comput Sci Appl. ;14(4) Shashkina V (2021) Assistive technology for the visually impaired: use cases and the particulars of development [Internet]. Available from: https://itrexgroup.com/blog/assistive-technology-for-visually-impared-examples-development-tips/#header Jones PR, Somoskeöy T, Chow-Wing-Bom H (2020) Seeing other perspectives: evaluating the use of virtual and augmented reality to simulate visual impairments (OpenVisSim). npj Digit Med 3:32 Cheng L, Peng S, Hao H (2024) Effect of different screen brightness and devices on online visual acuity test. Graefes Arch Clin Exp Ophthalmol 262:641–649 Nkeremuzor EC, Ikoro NC, Esenwah EC, Nzotta C, Ugwuoke GI, Nwaigwe OM et al (2020) Determination of Luminance Levels of Visual Acuity Charts in Selected Eye Clinics in Owerri and Their Effects on Vision. Asian J Res Rep Ophthalmol 3(2):13–23 Rodriguez RG, Paviglianiti V, Dumit C, Pattini A (2024) Dos & Don’ts in Measuring Illuminance With Smartphones. Ergonomics in Design [Internet]. ;10648046241263631. Available from: https://doi.org/10.1177/10648046241263631 Mohapatra S, Ganesh A, Zion N (2024) Assessment of Forward Head Posture in Information Technology Employees with Neck Pain: A Cross-Sectional Study. Journal of Orthopedics and Spine Trauma [Internet]. ; Available from: https://doi.org/10.18502/jost.v10i1.14962 Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7957115","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":535533272,"identity":"a652bf8f-9724-470c-b2a8-54cd83d4b0b8","order_by":0,"name":"Wai Lam Leung","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABAElEQVRIiWNgGAWjYFACxgaGB0CKD8ypSIAyGBLwawFJs4E5ZxKgDLxaoNJglYxtRGgxOH4YaEvNNgY29h6zh1/npcmzMTA//MDYloZby5lEoJZjtxnYeM6YG8tuyzFsY2AzlmBsy8GpxewASAsbUItEjpm05LaKBKDDzIAurMCt5fxDoJZ/QC3yb4Ba5oC0sH/Dr+UG0JbENpAtPGaSHxtygFp4QLbgdpj9DaAtiX23edh40sqkGY6lGbYx8xRLJJzD7X3J/vQHDB++3ZbjZz+8TfJHTbI8P3v7xg8fypJxagEC9h9AggfEYoaQDAQjEgEYfxCrchSMglEwCkYUAAD9GEn1YlC1zwAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0003-3982-0949","institution":"Institute of Psychiatry, Psychology \u0026 Neuroscience","correspondingAuthor":true,"prefix":"","firstName":"Wai","middleName":"Lam","lastName":"Leung","suffix":""},{"id":535533273,"identity":"d4c67948-d820-4551-b577-0a04a0f09cb9","order_by":1,"name":"Oliver Runswick","email":"","orcid":"https://orcid.org/0000-0002-0291-9059","institution":"Institute of Psychiatry, Psychology \u0026 Neuroscience","correspondingAuthor":false,"prefix":"","firstName":"Oliver","middleName":"","lastName":"Runswick","suffix":""},{"id":535533274,"identity":"c213c308-7e3c-4f59-8b30-a6649b7047c7","order_by":2,"name":"Claire Heard","email":"","orcid":"","institution":"Institute of Psychiatry, Psychology \u0026 Neuroscience","correspondingAuthor":false,"prefix":"","firstName":"Claire","middleName":"","lastName":"Heard","suffix":""},{"id":535533275,"identity":"8bfc9d87-d368-42a0-8f11-3daff3d1fca8","order_by":3,"name":"Tim Rakow","email":"","orcid":"","institution":"Institute of Psychiatry, Psychology \u0026 Neuroscience","correspondingAuthor":false,"prefix":"","firstName":"Tim","middleName":"","lastName":"Rakow","suffix":""}],"badges":[],"createdAt":"2025-10-27 12:05:24","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-7957115/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7957115/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":94762423,"identity":"129d13fa-6c63-4b94-bf27-ad2ba6d51ff1","added_by":"auto","created_at":"2025-10-30 12:11:17","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":938335,"visible":true,"origin":"","legend":"","description":"","filename":"TVSTManuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/5c1f4578db8d6d4374468d63.docx"},{"id":94824735,"identity":"8198958d-a016-4ff6-83e5-c5748b885c9b","added_by":"auto","created_at":"2025-10-31 06:49:16","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":342,"visible":true,"origin":"","legend":"","description":"","filename":"rs7957115.json","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/828b999056aab7b1417ed35b.json"},{"id":94762424,"identity":"bf4ae5e0-27c2-47ef-9ee8-395d6b33d727","added_by":"auto","created_at":"2025-10-30 12:11:17","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":106351,"visible":true,"origin":"","legend":"","description":"","filename":"rs79571150enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/1993fdafdfe65f10de62e144.xml"},{"id":94823553,"identity":"8fc5fd06-0efb-4daa-b9e3-c6f651979fdc","added_by":"auto","created_at":"2025-10-31 06:47:36","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":39910,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/2797908e1f7fb69ca7dc7a82.png"},{"id":94823813,"identity":"1e1f130a-6fda-438c-a734-d3c9949a80a2","added_by":"auto","created_at":"2025-10-31 06:48:03","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":52229,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/6fbdc53e476f14548c1fb055.png"},{"id":94824727,"identity":"5a0ff7d8-bf5f-4186-a4f7-7ce62adfec36","added_by":"auto","created_at":"2025-10-31 06:49:15","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":188848,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/48cd8326d9a74949a9ca6465.png"},{"id":94762432,"identity":"9659c2c5-ce69-46f3-ad10-5410df2ec627","added_by":"auto","created_at":"2025-10-30 12:11:18","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":184277,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/73e572d750fd13fb3104c50f.png"},{"id":94762431,"identity":"4aacb939-000f-4f59-9f66-309b8ef3faec","added_by":"auto","created_at":"2025-10-30 12:11:18","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":184092,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/f1c59820a02aff8e7d1d5205.png"},{"id":94762429,"identity":"c5644f7c-9eb0-4f3a-b3bd-1329efc3c990","added_by":"auto","created_at":"2025-10-30 12:11:18","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":64573,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/daf75a6c2ff1fb7ba8b5ef02.png"},{"id":94824497,"identity":"2f4ec5fb-2de2-4a21-a74f-e32dd55b2edc","added_by":"auto","created_at":"2025-10-31 06:49:02","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":47453,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/09a07710173992a57657e4f8.png"},{"id":94762437,"identity":"9358fb53-3081-41ba-8a90-9b4d0b61d83c","added_by":"auto","created_at":"2025-10-30 12:11:18","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":85505,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/aa66f3bb899027ba853a3561.png"},{"id":94824746,"identity":"2a020ef0-1931-4930-b876-6f085680e0f0","added_by":"auto","created_at":"2025-10-31 06:49:16","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":17198,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/7156fa4835f02a3a5fb72570.png"},{"id":94824730,"identity":"1a64db8d-7636-42d7-beb6-83d6fc532e13","added_by":"auto","created_at":"2025-10-31 06:49:15","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":15111,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/308f519e19e2b855d4fb73c6.png"},{"id":94762441,"identity":"55d7fb64-6bb5-482e-957b-4ef5a007bbde","added_by":"auto","created_at":"2025-10-30 12:11:18","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":50696,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/c2db7b728df05cdf3f7d450c.png"},{"id":94762436,"identity":"64014b7c-49b2-4dd3-a8b7-9c2bf4800e58","added_by":"auto","created_at":"2025-10-30 12:11:18","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":49074,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/2afa8b0919490b4c4075c7ee.png"},{"id":94824623,"identity":"a278390e-cdea-4ef4-8459-39e9eb703eca","added_by":"auto","created_at":"2025-10-31 06:49:10","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":49328,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/243f0aa0ec3b11d51ccd119b.png"},{"id":94762445,"identity":"e0ae01b5-7949-41d6-91fe-ab0b29a7e58b","added_by":"auto","created_at":"2025-10-30 12:11:18","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":22157,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/fe4c0e14b14cbb8884816fcb.png"},{"id":94824973,"identity":"e007cc32-bb5a-4bab-86ca-10979db4141f","added_by":"auto","created_at":"2025-10-31 06:49:38","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":17558,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/65aa8d9f2b11334198ca0a6e.png"},{"id":94762446,"identity":"3ca37ed9-0aec-4fce-ba0a-74e3eeba376f","added_by":"auto","created_at":"2025-10-30 12:11:18","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":29917,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/fae7b659570dc1c9a57fbffe.png"},{"id":94823769,"identity":"3521fbf2-3199-47f5-ab46-28c75a1eb5e3","added_by":"auto","created_at":"2025-10-31 06:47:57","extension":"xml","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":103987,"visible":true,"origin":"","legend":"","description":"","filename":"rs79571150structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/ede2514ba6d628e0bc9ef35a.xml"},{"id":94762447,"identity":"09048365-7bdb-4bda-aa0e-f0946ceffb90","added_by":"auto","created_at":"2025-10-30 12:11:18","extension":"html","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":115834,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/f9b4462cdd4582a9f7fa9e40.html"},{"id":94762419,"identity":"1eaea539-6453-454f-a4f3-189873e620d2","added_by":"auto","created_at":"2025-10-30 12:11:17","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":39910,"visible":true,"origin":"","legend":"\u003cp\u003eValidation phase structure\u003c/p\u003e","description":"","filename":"figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/f612664c822cbc2e009bb9df.png"},{"id":94762420,"identity":"87272e96-7707-48ad-ae87-b6108ac8f1b8","added_by":"auto","created_at":"2025-10-30 12:11:17","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":52229,"visible":true,"origin":"","legend":"\u003cp\u003eMapping phase structure\u003c/p\u003e","description":"","filename":"figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/ec036d42db6b32f649ad445b.png"},{"id":94762422,"identity":"395fa189-cca2-486f-9b01-40687fd779a2","added_by":"auto","created_at":"2025-10-30 12:11:17","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":188848,"visible":true,"origin":"","legend":"\u003cp\u003eBland-Altman Agreement Plot for alignment in logMAR between Simulation Software on the 218 PPI iMac monitor and ETDRS Charts (3 and 4 Metres) and between control, medium and severe simulation spectacles. Central dashed lines represent the mean difference, with upper and lower bounds representing limits of agreement (± 1.96 × the standard deviation of the differences).\u003c/p\u003e","description":"","filename":"figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/7483780843be8bbc38eeb4ac.png"},{"id":94824514,"identity":"d54410f8-b128-4e12-9fe8-125e15a94a06","added_by":"auto","created_at":"2025-10-31 06:49:03","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":184277,"visible":true,"origin":"","legend":"\u003cp\u003eBland-Altman Agreement Plot for alignment in logMAR between Simulation Software on the 170 PPI Acer monitor and ETDRS Charts (3 and 4 Metres) and between control, medium and severe simulation spectacles. Central dashed lines represent the mean difference, with upper and lower bounds representing limits of agreement (± 1.96 × the standard deviation of the differences).\u003c/p\u003e","description":"","filename":"figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/ea023d991488c068cd559030.png"},{"id":94762435,"identity":"c70efbff-929f-4230-b077-b5de08f0bdf9","added_by":"auto","created_at":"2025-10-30 12:11:18","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":184092,"visible":true,"origin":"","legend":"\u003cp\u003eBland-Altman Agreement Plot for alignment in logMAR between Simulation Software on the 94 PPI Samsung monitor and ETDRS Charts (3 and 4 Metres) and between control, medium and severe simulation spectacles. Central dashed lines represent the mean difference, with upper and lower bounds representing limits of agreement (± 1.96 × the standard deviation of the differences).\u003c/p\u003e","description":"","filename":"figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/6c7782068e83716773a3a7b9.png"},{"id":94762428,"identity":"234f5f83-a78c-4b6e-acf2-44f1e8b943bd","added_by":"auto","created_at":"2025-10-30 12:11:18","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":64573,"visible":true,"origin":"","legend":"\u003cp\u003eMixed-effects log-linear regression model for prediction of logMAR score change from monitor PPI and Gaussian blur strength.\u003c/p\u003e","description":"","filename":"figure6.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/d6e9c850659862b209b38644.png"},{"id":94762427,"identity":"6da44b5f-edc0-4e08-8aba-2795e68bf08d","added_by":"auto","created_at":"2025-10-30 12:11:18","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":47453,"visible":true,"origin":"","legend":"\u003cp\u003ePooled regression model of logMAR score change by Gaussian blur strength.\u003c/p\u003e","description":"","filename":"figure7.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/8c0be9019351b56223cb325b.png"},{"id":94762439,"identity":"314d57a5-7511-42bb-9f0e-432437c0c2b7","added_by":"auto","created_at":"2025-10-30 12:11:18","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":85505,"visible":true,"origin":"","legend":"\u003cp\u003e(A) WHO classification of vision impairment and the targeted classification for simulation (0.75 logMAR) and the acceptable range of error; (B) graphical presentation of logMAR difference (error) between pooled model vs. mixed-effects model.\u003c/p\u003e","description":"","filename":"figure8.png","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/90a5bb9d1e0e0724be0f5830.png"},{"id":94827255,"identity":"dc387827-f326-4a04-bdf4-b8be85aaf30a","added_by":"auto","created_at":"2025-10-31 06:56:24","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1500294,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7957115/v1/03ef4a26-66d4-4dde-8df2-0124a1e747ce.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eScreen-based Visual Impairment Simulation: An Exploratory Investigation of Acuity Loss Simulation Using Gaussian Blur Filters\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eInternational Agency for the Prevention of Blindness demonstrated that there are 1.1\u0026nbsp;billion people globally with vision loss, and it is projected to increase to 1.7\u0026nbsp;billion by 2050 (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). Various daily activities can be difficult for these individuals, including managing finances online, online shopping, navigation using apps, video calls, reading digital documents (\u003cspan additionalcitationids=\"CR3\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e), and even assistive technologies designed to aid accessibility because many developers lack knowledge of accessibility (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). Consequently, researchers have sought to understand the nature and consequences of vision impairment and to inform, develop and test mitigations for its adverse consequences.\u003c/p\u003e\u003cp\u003eOne cost-effective methodology used to understand the impacts of vision impairment and the accessibility of products (e.g., websites, navigation systems, accessibility tools) is simulating the loss of visual function in sighted individuals (\u003cspan additionalcitationids=\"CR7\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e). Simulation methods cannot replicate the lived experiences of individuals with vision impairment because sighted individuals can foveate the artificially degraded images and use their normally functioning visual pathways. However, simulations can facilitate experimental control by employing within-subject designs with different simulated degrees of impairment and increase recruitment by allowing less restrictive inclusion criteria. To this end, simulation methods have been used to understand the impact of vision loss in tasks ranging from everyday living to playing sports (\u003cspan additionalcitationids=\"CR10\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eA common approach to simulating vision impairment is via simulation spectacles that use different filters to induce varying degrees of acuity loss, a reduction in sharpness and clarity of vision (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e) valuable for research, simulation spectacles have some limitations. Remote testing without the researcher present is impractical, participants are still able to move their eye behind the simulation spectacle without the impairment remaining in the same place in the visual field, and eye tracking is generally not possible because the eye is not visible. Modern alternatives include the use of virtual reality simulations with gaze-contingent simulation of impairment (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e), but these approaches again require in-person testing and significant efforts in stimuli creation. For these reasons and others, we explored and tested an alternative approach to simulating vision loss, which involves applying blur to screen images. Such simulation via an electronic display could address the abovementioned limitations: reduce physical labour load for data collection, facilitate larger sample sizes by allowing remote testing with many individuals participating concurrently and could allow for gaze tracking when this is valuable for the research question. Additionally, compared to simulation spectacles, electronic displays are better for producing consistent and accurate simulations of visual defects (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe current experiment tests this image-blur method for simulating (different degrees of) visual acuity loss. Loss of visual acuity is the most common form of visual impairment; it is associated with several common pathologies (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e) and ageing (\u003cspan additionalcitationids=\"CR16\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e); and is highly correlated with other impairments such as contrast sensitivity (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). The World Health Organisation defines vision impairment through measures of acuity where mild, moderate and severe impairment are defined as 0.3 to 0.5, 0.5 to 1.0, and 1.0 to 1.3 logMAR (logarithm of the Minimum Angle of Resolution, a scale used to measure visual acuity, where lower values indicate better vision), respectively. We selected Gaussian blur as the sole method for simulating acuity loss because this image-processing technique is widely available and should be relatively straightforward for researchers to implement.\u003c/p\u003e\u003cp\u003eGaussian blur is an image processing technique that reduces image noise and detail by averaging the colour or intensity values (for greyscale images) of each pixel with its surrounding pixels using the Gaussian function, which smooths out the images to create a blurry and softened effect. Gaussian blur has previously been applied via more complex algorithms to simulate impairment in visual field and acuity (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). The degree (width) of blurring is adjusted by changing the value of the standard deviation (σ) of the Gaussian function, often referred to as the radius or sigma, where the larger the standard deviation, the greater the blur. In the following equation that defines image blur, x and y are the coordinates of a point in the 2D plane, and e is Euler\u0026rsquo;s number:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:G\\left(x,y\\right)=\\:\\frac{1}{2\\pi\\:{\\sigma\\:}^{2}}{e}^{-\\frac{{x}^{2}+{y}^{2}}{2{\\sigma\\:}^{2}}}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eThe primary aim of our experiment was to determine the appropriate σ for the Gaussian function to simulate specific levels of acuity loss. Therefore, this experiment aimed to understand the relationship between Gaussian blur across different values of σ and the resulting functional impairment as measured by vision tests. By developing a regression model that maps the σ value to an estimated level of visual impairment, we provide guidance for using simple Gaussian blur in well-controlled research on visual impairment. Because variations in electronic displays could impact the precision of the simulation, we also explored the impact of these display variations, focusing primarily on differences in screen pixel density. Our goal was to explore whether a single regression model could be used to identify simulation settings to be used across different types of screens or whether substantially different models are required for different displays.\u003c/p\u003e\u003cp\u003eTo achieve our primary objective, we first must ensure that visual acuity can be measured accurately using computer screens. To this end, participants completed visual acuity tests using the ETDRS (Early Treatment Diabetic Retinopathy Study) on a physical chart and a newly developed vision test emulated the ETDRS letters on three types of computer screens (94/170/218 PPI [pixel per inch]). To further ensure the reliability of capturing visual acuity using screens, test results from screen-based vision tests were compared to physical testing at two distances, three and four metres, and under three different vision conditions. Participants completed those screen-based and physical vision tests without wearing simulation spectacles and with simulation spectacles. These spectacles simulated moderate and severe acuity loss to ensure the screen-based vision test can still measure visual acuity in low vision individuals to the same extent as physical charts\u003c/p\u003e\u003cp\u003eFollowing this initial validation phase, we then applied different levels of Gaussian blur to the screens during the screen-based vision test to map the degree of acuity loss per increment of blur. Regression models were developed from these data to map out the relationship between blur strength (σ) and visual acuity that can be used to estimate the value of σ for specific levels of simulated acuity loss. This mapping process was conducted independently on three computer screens to examine whether differences in PPI impact the relationship between Gaussian blur applied on screen and visual acuity, in order to determine whether the simulation technique can be used generally across a range of screens with vary PPI or if each screen requires its own calibrated relationship between blur and visual acuity.\u003c/p\u003e\u003cp\u003eOverall, the current experiment aims to investigate whether screen-based acuity tests are a valid approach to vision testing, which in turn allows for the generation of regression models that can be used to estimate the appropriate level of Gaussian blur for specific levels of simulated acuity loss for future in-person and remote simulation studies.\u003c/p\u003e"},{"header":"Method","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eParticipants\u003c/h2\u003e\u003cp\u003eA total of 37 volunteers participated (8 male, 28 female, and 1 \u0026lsquo;prefer not to say\u0026rsquo;) with ages ranging from 18 to 23 years (\u003cem\u003eM\u003c/em\u003e\u0026thinsp;=\u0026thinsp;18.8, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.0). Participants were first-year psychology undergraduates, recruited via an online participant management platform. They participated in the study for course credit. Informed written consent was obtained from all participants before the study, ensuring they were fully aware of the procedures and their rights. The study received ethical approval from the local ethics committee (Ethics Registration Number: MRSP-23/24-40842).\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eMaterial\u003c/h3\u003e\n\u003cp\u003e\u003cem\u003eETDRS charts.\u003c/em\u003e Two early treatment diabetic retinopathy study (ETDRS) charts by \u003cem\u003ePrecision Vision\u003c/em\u003e were used, one designed for testing at 4 metres and one at 3 metres. The testing range of both charts is from the logMAR score of 1.0 to -0.4.\u003c/p\u003e\u003cp\u003e\u003cem\u003eSimulation software.\u003c/em\u003e Simulation software was developed to apply Gaussian blur from the SciPy python package (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e) onto tumbling E eye tests that measure logMAR scores. The letter size adheres to a conventional 4-metre (testing distance) ETDRS chart, where the letter for logMAR\u0026thinsp;=\u0026thinsp;1.0 is of the size of 5.82 x 5.82 cm. Compared to the \u003cem\u003eETDRS charts\u003c/em\u003e, four rows of letters above the logMAR\u0026thinsp;=\u0026thinsp;1.0 row and eight rows below the logMAR = -0.4 are added using the standard letter size increment/decrement of 0.1 log units (\u0026asymp;\u0026thinsp;1.26x) (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e). LogMAR score calculations are automated and exported into a CSV file.\u003c/p\u003e\u003cp\u003e\u003cem\u003eSimulation spectacles.\u003c/em\u003e Two pairs of simulation spectacles, adapted from \u0026ldquo;Cambridge simulation glasses\u0026rdquo; (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e) created by Runswick and colleagues using Hampshire frost filters were acquired for the experiment (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e). The level four (logMAR mean\u0026thinsp;=\u0026thinsp;1.21, SD\u0026thinsp;=\u0026thinsp;\u0026plusmn;\u0026thinsp;0.23 [Range: 0.88\u0026ndash;1.60]) and six (logMAR mean\u0026thinsp;=\u0026thinsp;1.84, SD\u0026thinsp;=\u0026thinsp;\u0026plusmn;\u0026thinsp;0.18 [Range: 1.60\u0026ndash;2.50]) spectacles are used and labelled as medium and severe acuity loss spectacles, respectively.\u003c/p\u003e\u003cp\u003e\u003cem\u003eMonitors\u003c/em\u003e. Three different monitors were used, each varying in pixels per inch (PPI), a measure of pixel density in digital displays and a higher PPI indicates more pixels per inch. The monitors used were: 1) iMac with Retina 5k display at 218 PPI, 2) Acer Predator Helios 16 at approximately 170 PPI, and 3) Samsung CF396 24\u0026rdquo; monitor at 94 PPI. All screens are adjusted to 330 lux units to adhere to the same emitted lux by the ETDRS charts, measured by a lux meter hovering directly above the screens.\u003c/p\u003e\n\u003ch3\u003eProcedure\u003c/h3\u003e\n\u003cp\u003eThe experiment was segmented into two parts, the first part being the \u003cem\u003evalidation phase\u003c/em\u003e, where participants completed tests on the \u003cem\u003eETDRS charts\u003c/em\u003e and \u003cem\u003esimulation software\u003c/em\u003e with habitual vision, moderate, and severe vision induced by \u003cem\u003esimulation spectacles\u003c/em\u003e to validate the accuracy of doing vision tests on multiple monitors compared to conventional charts at varying levels of acuity loss. In the second part, the \u003cem\u003emapping phase\u003c/em\u003e, participants completed a series of vision tests on the \u003cem\u003esimulation software\u003c/em\u003e with Gaussian blurs with incremental strength (i.e., standard deviation; σ) applied on the screen to map out the relationship between blur strength and acuity loss. The mapping phase is of mixed design, with the independent variable of \u003cem\u003emonitors\u003c/em\u003e (218/170/94 PPI) being the between-subject factor and Gaussian blur strength (σ\u0026thinsp;=\u0026thinsp;0 to 10 with the increment of 0.5 and σ\u0026thinsp;=\u0026thinsp;20 to 90 with the increment of 10) being the within-subject factor, and the dependent variable being the logMAR measured for each Gaussian blur strength level.\u003c/p\u003e\u003cp\u003e\u003cem\u003eValidation phase.\u003c/em\u003e Participants were instructed to wear or not wear \u003cem\u003esimulation spectacles\u003c/em\u003e and complete binocular vision tests on the \u003cem\u003eETDRS charts\u003c/em\u003e (4 and 3 metres) and \u003cem\u003esimulation software\u003c/em\u003e until all vision testing methods (\u003cem\u003eETDRS charts\u003c/em\u003e at 3 metres, 4 metres, or \u003cem\u003esimulation software\u003c/em\u003e) were completed on all \u003cem\u003esimulation spectacles\u003c/em\u003e levels (within-subjects: no, medium, and high severity). Overall, each participant completed 3 simulation spectacles x 3 testing methods\u0026thinsp;=\u0026thinsp;9 vision tests (see Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The order in which those combinations of \u003cem\u003esimulation spectacles\u003c/em\u003e and testing methods were completed was randomised.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eFor the \u003cem\u003eETDRS charts\u003c/em\u003e trials, participants were instructed to read aloud from the top row to the bottom. The test terminates when errors are made on three or more letters in a given row, and then the logMAR score is calculated. If participants cannot read the top row of the charts, their testing distance is halved, and a 0.3 logMAR score is added to their total score. For the \u003cem\u003esimulation software\u003c/em\u003e trials, participants were tested at a 1-metre distance using a tumbling E vision test on their randomly allocated monitor. Since the selected letter size (5.82 x 5.82 cm) is according to the 4 metres \u003cem\u003eETDRS chart\u003c/em\u003e, using the testing-distance principle described above, a 0.6 logMAR score was added to the participants\u0026rsquo; logMAR score due to halving twice from 4 metres. Participants used four directional arrow keys on a keyboard to indicate which direction the tumbling E is facing, based on the direction of the opening of the letter E (e.g., press up-arrow if the E\u0026rsquo;s opening is facing upward). Each test is terminated when three or more incorrect responses are given in a row.\u003c/p\u003e\u003cp\u003e\u003cem\u003eMapping Phase.\u003c/em\u003e Participants completed a series of tumbling E vision tests on the \u003cem\u003esimulation software\u003c/em\u003e with Gaussian blur applied on the screen. The settings, scoring, and instructions were identical to those used in the \u003cem\u003evalidation phase\u003c/em\u003e. In a pilot study (\u003cem\u003eN\u003c/em\u003e\u0026thinsp;=\u0026thinsp;10), a logarithmic trend was observed in the relationship between the change in logMAR score and Gaussian blur strength (σ). Specifically, increases in σ at higher levels of blur resulted in minimal changes to the logMAR score, whereas increments in σ at lower levels of blur produced a steep increase in the logMAR score. Therefore, to avoid fatigue effects from testing an excessive number of levels, the tested blur strength was from σ\u0026thinsp;=\u0026thinsp;0 to 10 with an increment of 0.5, and from σ\u0026thinsp;=\u0026thinsp;20 to 90 with an increment of 10. This resulted in 29 separate vision tests, each with a different level of Gaussian blur applied; their presentation order was randomised (see Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eThe data for all 37 participants were analysed using Bland-Altman analysis, which involves plotting differences in logMAR scores between the \u003cem\u003esimulation software\u003c/em\u003e against \u003cem\u003eETDRS charts\u003c/em\u003e at 3 metres and (separately) at 4 metres (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e \u0026amp; Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). This analysis assesses the agreement between the simulation software and the ETDRS charts in measuring visual acuity to identify systematic biases or variability across different conditions and distances. The 95% CI for the mean difference in logMAR shows that the systematic biases in the control and severe simulation spectacles condition are relatively small, with narrow confidence intervals across \u003cem\u003emonitors\u003c/em\u003e and \u003cem\u003eETDRS charts\u003c/em\u003e (see Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). However, there is a noticeable positive bias for medium simulation spectacles across \u003cem\u003emonitors\u003c/em\u003e and \u003cem\u003eETDRS charts\u003c/em\u003e, particularly for the \u003cem\u003eETDRS chart\u003c/em\u003e at 3 metres with the highest mean difference at 0.09 [95% CI: 0.05, 0.13] logMAR. Nevertheless, the limits of agreement indicate acceptable variability, aligning with previous research comparing ETDRS charts to other validated vision tests (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e). Importantly, this finding falls below the 0.2 logMAR threshold typically considered a meaningful change (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e). Moreover, this trend is consistent across \u003cem\u003emonitors\u003c/em\u003e, suggesting good agreement across \u003cem\u003emonitors\u003c/em\u003e, though with a slightly larger variability of measured logMAR for the medium simulation spectacles across all three \u003cem\u003emonitors\u003c/em\u003e.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eBland-Altman Agreement: Mean Difference in logMAR Between Simulation Software on 218, 170 and 94 PPI screens and ETDRS Charts (3 and 4 Metres)\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003ePixel Per Inch (PPI)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eSimulation Spectacles\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003eMean difference [95% CI]\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eETDRS (3-metres)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eETDRS (4-metres)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003e218 PPI\u003c/p\u003e\u003cp\u003e(n\u0026thinsp;=\u0026thinsp;13)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eControl\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-0.058 [-0.10, -0.01]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.015 [-0.15, 0.18]\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMedium\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.090 [0.05, 0.13]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.064 [0.004, 0.12]\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSevere\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.002 [-0.05, 0.05]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.005 [-0.04, 0.05]\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003e170 PPI\u003c/p\u003e\u003cp\u003e(n\u0026thinsp;=\u0026thinsp;12)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eControl\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-0.019 [-0.06, 0.02]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.017 [-0.07, 0.04]\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMedium\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.088 [0.04, 0.13]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.084 [0.04, 0.12]\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSevere\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.053 [0.01, 0.09]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.047 [0.006, 0.08]\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003e94 PPI\u003c/p\u003e\u003cp\u003e(n\u0026thinsp;=\u0026thinsp;12)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eControl\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-0.026 [-0.07, 0.01]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.03 [-0.07, 0.009]\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMedium\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.033 [-0.02, 0.09]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.018 [-0.05, 0.08]\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSevere\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.012 [-0.03, 0.05]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.007 [-0.03, 0.02]\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe logMAR score change was calculated by subtracting the logMAR score obtained at σ\u0026thinsp;=\u0026thinsp;0 from logMAR scores obtained from σ\u0026thinsp;\u0026gt;\u0026thinsp;0. A mixed-effects log-linear regression model was fit by REML, where the dependent variable was the logMAR score change, and the independent variables were \u003cem\u003emonitor\u0026rsquo;s\u003c/em\u003e PPI and Gaussian blur strength (σ) with random effects of different participants accounted for. Note that the PPI variable was treated as categorical in all analyses since the difference between screens may involve more than just a difference in pixel density (e.g., contrast). A total of 1023 data points were analysed after 13 were removed due to misinput during the experiment by participants (i.e., stuck keys on the keyboard and accidental key holds that led to completion of trials despite no input from participants). The mixed-effects model\u0026rsquo;s formula is as follows:\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\:logMAR\\:Change\\sim\\:PPI\\:+\\text{log}\\left(Gaussian\\:Blur\\:\\sigma\\:\\right)+(PPI\\:\\times\\:\\text{log}\\left(Gaussian\\:Blur\\:\\sigma\\:\\right))+(1\\left|Participant\\right)$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eThe conditional R\u0026sup2; indicated that fixed and random effects explained 95.19% of the variance, while the marginal R\u0026sup2; showed that the fixed effects alone explained 90.71%. A type III ANOVA with Satterthwaite\u0026rsquo;s method found significant effects for the main effect of monitor PPI, \u003cem\u003eF\u003c/em\u003e(2, 44.88)\u0026thinsp;=\u0026thinsp;3.91, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.02, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\omega\\:}^{2}\\)\u003c/span\u003e\u003c/span\u003e = 0.15, 95% CI [0.00, 1.00], the main effect of blur, \u003cem\u003eF\u003c/em\u003e(1, 983.22) = 19590.73, \u003cem\u003ep\u003c/em\u003e \u0026lt; .001, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\omega\\:}^{2}\\)\u003c/span\u003e\u003c/span\u003e = 0.95, 95% CI [0.95, 1.00], and the interaction term, \u003cem\u003eF\u003c/em\u003e(2, 983.22) = 12.13, \u003cem\u003ep\u003c/em\u003e \u0026lt; .001, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\omega\\:}^{2}\\)\u003c/span\u003e\u003c/span\u003e = 0.02, 95% CI [0.01, 1.00] (see Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e\u003cp\u003ePairwise comparison between monitors with Bonferroni correction found that the screen with 94 PPI Samsung and 218 PPI iMac monitors had significantly higher increase of logMAR compared to the 170 PPI Acer monitor, \u003cem\u003et\u003c/em\u003e(35.5)\u0026thinsp;=\u0026thinsp;4.299, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003eSE\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.03, 95% CI [0.07, 0.27], and, \u003cem\u003et\u003c/em\u003e(35.5)\u0026thinsp;=\u0026thinsp;4.63, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003eSE\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.03, 95% CI [0.27, 0.08], respectively. However, the 94 PPI Samsung and 218 PPI iMac monitors did not significantly differ from each other, \u003cem\u003et\u003c/em\u003e(35.6) = -0.25, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.00, \u003cem\u003eSE\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.03, 95% CI [-0.10, 0.08].\u003c/p\u003e\u003cp\u003eAs the interaction term was significant in the omnibus test, it suggests that blur may perceived differently based on display type; therefore, a comparison of slopes was conducted to examine the extent of such effect. The analysis found that the slope of log(Gaussian Blur σ) for the 94 PPI Samsung monitor was \u003cem\u003eb\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.416, \u003cem\u003eSE\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.005, 95% CI [0.40,0.42], \u003cem\u003edf\u003c/em\u003e\u0026thinsp;=\u0026thinsp;984, for the 170 PPI Acer monitor was \u003cem\u003eb\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.381, \u003cem\u003eSE\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.005, 95% CI [0.37, 0.39], \u003cem\u003edf\u003c/em\u003e\u0026thinsp;=\u0026thinsp;983, and for the 218 PPI iMac monitor was \u003cem\u003eb\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.404, \u003cem\u003eSE\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.004, 95% CI [0.39, 0.41], \u003cem\u003edf\u003c/em\u003e\u0026thinsp;=\u0026thinsp;983. Such findings suggests that the log slopes of σ differ across screen levels, but the differences are quite small in absolute terms. Specifically, the difference between 94 PPI Samsung and 170 PPI Acer monitors is only 0.035, and between the 170 PPI Acer and 218 PPI iMac monitors is 0.023. Given that the slopes are within a narrow range and the confidence intervals are quite narrow, it suggests that the relationship between log(Gaussian Blur σ) and logMAR change is relatively consistent across screen levels. Practically speaking, when Gaussian Blur\u0026rsquo;s σ is at 2, 10, and 20, the logMAR change for each screen according to the slope analysis corresponds fairly well to the classification of mild, moderate, and severe vision impairment, respectively (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eComparison of Predicted logMAR Change Across Different levels and Screen Types\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e\u003cp\u003eσ\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u003cb\u003e94 PPI Samsung\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cb\u003e170 PPI Acer\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003e218 PPI iMac\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eσ\u0026thinsp;=\u0026thinsp;2\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.288\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.264\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.280\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eσ\u0026thinsp;=\u0026thinsp;10\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.958\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.877\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.930\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eσ\u0026thinsp;=\u0026thinsp;20\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1.246\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.141\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.210\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eCoupled with the small effect size of the interaction (ω\u0026sup2; = 0.02), these findings suggest that screen PPI or monitor type is unlikely to be of practical concern for monitors of comparable specification to those that we tested. However, it is important to note that as σ increases, the magnitude of differences in logMAR change also increases. Nonetheless, this effect remains of limited practical concern for simulations ranging across mild to severe acuity loss, because the differences do not exceed the 0.2 logMAR threshold for meaningful changes in visual acuity (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e). Therefore, we opted for a pooled model that disregards the fixed effect of screen PPI or monitor type to simplify future implementation of this method (see Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe pooled regression model was statistically significant, \u003cem\u003eF\u003c/em\u003e(1, 1021)\u0026thinsp;=\u0026thinsp;10180, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001 with adjusted \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}^{2}\\)\u003c/span\u003e\u003c/span\u003e at 0.91. The main effect of log-transformed blur was significant at \u003cem\u003et\u003c/em\u003e(1021)\u0026thinsp;=\u0026thinsp;100.91, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, indicating a strong positive relationship between blur strength and logMAR score change. The high adjusted \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}^{2}\\)\u003c/span\u003e\u003c/span\u003e (0.91) suggests that the regression model (without screen PPI or monitor type included) should provide a reliable guide for future application of the blur simulation method..\u003c/p\u003e\u003cp\u003eTo further affirm the use of the pooled model instead of the more complicated mixed-effects model, we employed bootstrapping techniques of 1000 samples referencing data from the 170 PPI Acer monitor to assess the prediction accuracy of both models (pooled and mixed-effects model) using Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and R-squared (R\u0026sup2;). The observed difference between the two models is summarised in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, where the differences were minimal. For instance, RMSE and MAE for both models were almost identical, suggesting that the predictions from the pooled model would only make a 0.001 logMAR difference (RMSE [Mean]: 0.152\u0026ndash;0.151\u0026thinsp;=\u0026thinsp;0.001; MAE [Mean]: 0.116\u0026ndash;0.117\u0026thinsp;=\u0026thinsp;0.001) compared to the mixed-effects model. Although the R\u0026sup2; value demonstrated a higher divergence (0.08), the minimal incremental improvement for the mixed-effects model does not justify the added complexity.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eMean Square Error (RMSE), Mean Absolute Error (MAE), and R-squared (R\u0026sup2;) between Pooled and Mixed-effects models from 1000 bootstrapping samples.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eModel\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMetric\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eMean\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2.5% CI\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003e97%CI\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePooled\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRMSE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.152\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.152\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.154\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMAE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.116\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.111\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.122\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eR\u0026sup2;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.918\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.918\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.918\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eMixed-effects\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRMSE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.151\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.151\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.152\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMAE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.117\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.115\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.119\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eR\u0026sup2;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.926\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.925\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.926\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\n\u003ch3\u003eSensitivity Analysis of Pooled vs. Mixed-effects Model\u003c/h3\u003e\n\u003cp\u003eTo further examine the sufficiency of the pooled model, we also investigated the prediction difference of simulated logMAR given a specific σ between models. One example scenario is portrayed in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e, where we used the pooled model for inverse prediction to estimate the appropriate σ to simulate moderate vision impairment (logMAR\u0026thinsp;=\u0026thinsp;0.5\u0026ndash;1.0); the targeted logMAR was 0.75, the middle-point of the classification to maximise the acceptable range of error (\u0026plusmn;\u0026thinsp;0.25 logMAR; Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA). In Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eB, the logarithmic-like red, green, and orange lines with lower opacity represent the prediction difference of logMAR between the pooled model and the mixed-effects model for 94, 170 and 218 PPI screens. Furthermore, the orange vertical line represents the predicted σ value from the pooled model to simulate 0.75 logMAR, being σ\u0026thinsp;=\u0026thinsp;7.3. When applying σ\u0026thinsp;=\u0026thinsp;7.3 to the mixed-effects model, we can see that compared to the pooled model, 170 PPI screens predicted 0.11 logMAR less than the pooled model, meaning that applying σ\u0026thinsp;=\u0026thinsp;7.3 to Gaussian blur on the 170 PPI screen would result in 0.64 logMAR instead of 0.75 logMAR predicted by the pooled model. Since the acceptable range of error for moderate classification of vision impairment is \u0026plusmn;\u0026thinsp;0.25 logMAR (orange dotted line in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eB), the simulation would be deemed appropriate as the prediction error between the two models does not exceed the acceptable range (i.e., the orange curved line does not touch or exceeds the orange dotted line at σ\u0026thinsp;=\u0026thinsp;7.3). Moreover, even at a narrower range of acceptable errors, such as mild (\u0026plusmn;\u0026thinsp;0.10 logMAR) and severe (\u0026plusmn;\u0026thinsp;0.15 logMAR) vision impairment simulation, the error between the model does not exceed the boundary at σ\u0026thinsp;=\u0026thinsp;2.4 (green vertical line) and σ\u0026thinsp;=\u0026thinsp;19.8 (red vertical line). Regardless, there is a logarithmic increase in the prediction error of logMAR as σ increases, meaning the accuracy of the simulation may be lowered when the user intended to simulate a higher severity of acuity loss, which should be noted if precise simulation is required for future studies that aim beyond the simulation of vision impairment classifications from the International Classification of Diseases.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eTo summarise, the experiment aimed to examine whether screen-based acuity tests are a valid approach to vision testing, which in turn allows for the generation of regression models that can be used to estimate the appropriate level of Gaussian blur for specific levels of simulated acuity loss for future in-person and remote simulation studies. Our findings suggested good agreement between screen-based acuity tests and physical ETDRS charts at different distances and vision characteristics (habitual vision, moderate and severe acuity loss via simulation spectacles). Affirmation of the validity of screen-based acuity tests then allowed us to create a simple regression model that can be used to infer the appropriate level of Gaussian blur to simulate a given level of impact from acuity loss. This will facilitate remote simulation studies. The results indicate that this is feasible, given the significant log-linear relationship between the standard deviation (σ) of the Gaussian function and the measured logMAR under the influence of Gaussian blur. Specifically, a slight increase in σ before σ\u0026thinsp;=\u0026thinsp;10 resulted in a steep rise in logMAR. According to the classification of vision impairment from the International Classification of Diseases (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e), it shows that the simulation of mild (logMAR\u0026thinsp;=\u0026thinsp;0.30\u0026ndash;0.50) and moderate (logMAR\u0026thinsp;=\u0026thinsp;0.50\u0026ndash;1.00) vision impairment falls within the steep incline which could make simulation of mild and moderate acuity loss more challenging than severe due to the need for precise control over σ. The resulting equation for estimating the required \u0026#120590; for a given desired level of logMAR change (denoted as \u0026#119909;) is:\u003cdiv id=\"Equc\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e\n$$\\:\\sigma\\:=\\:{e}^{\\frac{x+0.04533908\\:}{0.4001204}}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eSensitivity analyses were conducted that supported using the pooled model over the mixed-effects model. While we initially anticipated that PPI might drive the differences between screens, our findings suggest this is unlikely, as the highest and lowest PPI screens showed nearly identical effects. Such observation indicates that PPI was not a monotonic factor in mapping blur to logMAR change. Instead, it is plausible that other monitor features, such as those that impact contrast (e.g., the type of backlighting), are responsible for the differences we observed. For example, the 170 PPI Acer monitor (marketed as a gaming laptop) uses advanced mini-LED backlight technology that provides a higher contrast ratio and deeper black visuals (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e) compared to the other two screens, which use more traditional LED backlighting methods. The advanced technology should create sharper outline of letters in the eye test which in turn results in a lower logMAR score.\u003c/p\u003e\u003cp\u003eIn practical terms, the effect of Gaussian blur was much larger than that of screen PPI or monitor type, suggesting that the monitor\u0026rsquo;s properties can be ignored in future simulation studies without seriously impacting design decisions. Nonetheless, those who prioritise precision (e.g., for small manipulations of acuity) may find it necessary to minimise variation in screen types among participants. This is especially pertinent for gaming laptops and monitors, which may have features designed to mitigate on-screen blur. Regardless, future research should validate the blur method across a broader range of screen types to investigate how factors like contrast ratio and the technologies that influence it might impact the application of Gaussian blur to images for simulations of acuity loss.\u003c/p\u003e\n\u003ch3\u003eApplications\u003c/h3\u003e\n\u003cp\u003eEmploying our simplified model to simulate vision impairment using Gaussian blur could significantly enhance product accessibility within the technology industry. Despite efforts to make technology and essential services accessible to individuals with vision impairments, research indicates that many products remain inaccessible due to inadequate research on user needs and poor implementation of user-centred design (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e). Many developers admit to a lack of accessibility education, resulting in inaccessible software and websites, leading to an estimated annual revenue loss of approximately 16\u0026nbsp;billion USD, with only 3.7% of the world\u0026rsquo;s top one million websites being fully accessible (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). Malkawi and colleagues (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e) report that assistive technologies for people with vision impairment are often developed based on researchers\u0026rsquo; personal knowledge without involving target groups during design phases. Beyond the lack of target group involvement, scientists are often insufficiently involved in technological design to carry out experiments that require knowledge and skillsets beyond developers\u0026rsquo; educational background and expertise (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e). In summary, many developers lack accessibility education and may be hesitant or unable to involve target groups in the design phase. They may also lack the resources to hire experts for specialised accessibility research on vision impairment. However, many of these challenges could be addressed by offering a low-labour, cost-effective research method, such as simulation research informed by our regression model, for controlled testing that informs the impact of vision impairment on the accessibility of assistive technologies during design phases for refinement before a more well developed product is then tested by those with vision impairment. Alternatively, this approach could incentivise remote simulation research, serving as a viable alternative to resource-intensive in-person experiments due to the required time, specialised equipment, and effort in getting participants to the lab.\u003c/p\u003e\u003cp\u003eIn prior research, Gaussian blur has been applied to more complex algorithms to simulate vision impairment, primarily in visual field simulation (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e) and gaze-contingent paradigms (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e). However, our findings point to simulation using simple Gaussian blur as an accessible alternative with several benefits. Firstly, implementing a more sophisticated simulation method may require a strong understanding of specific coding environments, such as MATLAB (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). These may not be familiar to all developers or researchers and require extensive training. Secondly, even if developers or researchers can operate the complex simulation environment, integrating sophisticated algorithms into websites or browser-based experimental platforms for convenient data collection could be challenging and require significant resources. Packaging these algorithms into software for installation to a research participant\u0026rsquo;s device could limit the sample pool to individuals with sufficient technical skills or confidence, thereby reducing participation among key groups (e.g., older adults). This means that it could pose threats to cross-sectional studies sensitive to age-related differences. In contrast, simple Gaussian blur filters are accessible and widely available in different coding environments (e.g., OpenCV [Python], Python Imaging Library [Python], Canvas API [JavaScript], imgaussfilt() [MATLAB]) as pre-installed and additional packages, including web development, and researchers who use web-based survey engines for experimentation can now easily apply simple Gaussian blur over images, text and videos to simulate acuity loss.\u003c/p\u003e\u003cp\u003eFor those interested in using Gaussian blur to simulate acuity loss, we have developed a web-based companion tool (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://wailamleung.shinyapps.io/g-balsac/\u003c/span\u003e\u003cspan address=\"https://wailamleung.shinyapps.io/g-balsac/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). This tool provides guidance on setting Gaussian blur parameters using one of the four available models (pooled model plus individual models based on data from each monitor we tested) to achieve a desired level of induced acuity loss (σ estimation) based on test subject(s) baseline visual acuity. It also supports inverse estimation, allowing users to convert σ values into logMAR for easy interpretation.\u003c/p\u003e\u003cp\u003eFurthermore, the current findings could benefit future research. Take Moharrer et al (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e) as an example, their research investigated the roles of motion perception and visual acuity in driving hazard perception where participants were equipped with simulation spectacles to simulate acuity loss (i.e., 20/120) and instructed to watch driving videos and presses a key as soon as hazard are recognised. Such an experiment can be streamlined further by applying blur onto the screen, avoiding the need for simulation spectacles and providing more accessible simulation conditions, which reduces labour and costs associated with acquiring spectacles.\u003c/p\u003e\n\u003ch3\u003eLimitations\u003c/h3\u003e\n\u003cp\u003eTwo notable factors may threaten the validity of simulations, especially in the context of remote studies. First, the brightness of screens, which research will have no direct control over when experiments are conducted remotely, could cause deviation in the perception of blur among participants. The current experiment controlled for brightness across the three screens with different PPI at 330 lux and approximately the maximum brightness for each screen, meaning that the relationship between σ and acuity loss may differ at screen brightness below or above 330 lux and the extent of its impact on acuity loss simulation is yet to be investigated. However, some studies suggest that brightness is not a concern when measuring acuity. Take Cheng et al as an example (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e); they found that there was a slight difference in visual acuity at different screen brightness levels at 50% and 100%, but the difference is equivalent to less than one letter on the visual acuity chart (-0.009 logMAR). Similarly, Nkeremuzor et al (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e) demonstrated that not with electronic displays but with fluorescent and tungsten bulb-illuminated charts, where brightness ranges from 13.28 cd/m\u0026sup2; to 14.64 cd/m\u0026sup2; and 14.68 cd/m\u0026sup2; to 15.76 cd/m\u0026sup2;, respectively, acuity measurements within and between fluorescent and tungsten bulbs were not statistically significant. These findings suggest that variations in screen brightness are unlikely to have a major impact on the accuracy of acuity measurement. Nonetheless, the addition of simple Gaussian blur may lead to an unforeseen interaction that says otherwise. Therefore, additional steps to control screen luminance may be wise. For example, participants could use smartphone applications to measure lux levels, which have been shown to demonstrate decent measurement agreement with traditional lux meters with appropriate settings (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e), to ensure their electronic display can reach 330 lux and be asked to set screen brightness accordingly. Regardless, future research could investigate the effect of screen luminance on screen-based vision tests, including under the context of Gaussian blur to validate the present methodology further.\u003c/p\u003e\u003cp\u003eSecond, sitting distance or leaning habits are another uncontrollable factor in a remote setting. The current experiment controlled for sitting distance at a 1-metre distance and actively prevented leaning forward or backwards from the screen during data collection. However, without supervision, this may not always be the case. When moving closer to the screen, the blurred pixels may appear more pronounced as you see blurred pixels at a larger size. Meanwhile, blurring might become less noticeable when farther from the screen. Under the circumstances of clear instructions for participants to sit 1 metre away from their electronic display, considering the substantial prevalence of forward head posture (associated with leaning towards screens) (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e), we can speculate an increased blurriness of stimuli in remote settings without supervision. However, the extent of sitting distance or head posture\u0026rsquo;s impact on perceived blur on electronic display remains unknown, which should be further investigated.\u003c/p\u003e\u003cp\u003eFinally, using simple Gaussian blur as a simulation method of acuity loss by no means fully captures the complexities of vision impairment. Vision impairment often involves not only acuity loss but also varying levels of contrast sensitivity, visual field loss, and other factors that can interact in ways not explored in the present research. Using simple Gaussian blur is just a systematic means of reducing visual information to provide an approximate insight into how individuals with one kind of vision impairment perceive on-screen content. Therefore, whatever the merits of simulating vision impairment, it should not replace testing with individuals with vision impairments.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThe current experiment demonstrated that simple Gaussian blur applied to electronic displays is a viable method for simulating visual acuity loss, offering a cost-effective and accessible alternative to resource-intensive simulation methods. The findings indicated a strong log-linear relationship between the degree of Gaussian blur and acuity loss, allowing precise control over simulating specific levels of acuity loss. Using the experimental data, we developed a pooled regression model that can guide the design of simulation experiments for acuity loss. The experiment provided proof-of-concept, suggesting that simulation studies using Gaussian blur applied to on-screen images have substantial potential for applied and translational research into visual impairment and for technical work to enhance the accessibility of technology and services. Regardless, future validations of this method should examine the impact of the factors that are not easily controlled in unsupervised settings (e.g., screen brightness, sitting distance) to further refine the value of the simulation method by developing necessary protocols.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eThis study was reviewed and approved by Research Ethics Office in King\u0026rsquo;s College London with the approval number: MRSP-23/24-40842, dated 04/01/2024. Participation was voluntary, and written informed consent was obtained from all individual participants included in this study.\u003c/p\u003e\u003ch2\u003eFunding Information\u003c/h2\u003e\u003cp\u003eThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eN/A\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eInternational Agency for the Prevention of Blindness [Internet]. International Agency for the Prevention of Blindness (2021) Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.iapb.org\u003c/span\u003e\u003cspan address=\"https://www.iapb.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRemillard ET, Koon LM, Mitzner TL, Rogers WA (2024) Everyday Challenges for Individuals Aging with Vision Impairment: Technology Implications. Gerontologist 64(6):169\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFranco M, Gaggi O, Merzougui SE, Palazzi CE (2023) Accessible Wayfinding for the Visually Impaired through Sustainable Smartphone Based Sensing. In: 2023 IEEE 20th Consumer Communications \u0026amp; Networking Conference (CCNC. IEEE; pp. 1\u0026ndash;6\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOrtiz-Escobar LM, Chavarria MA, Sch\u0026ouml;nenberger K, Hurst S, Stein MA, Mugeere A et al (2023) Assessing the implementation of user-centred design standards on assistive technology for persons with visual impairments: a systematic review. Front Rehabilitation Sci 4:1238158\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eParthasarathy PD, Joshi S (2024) Exploring the Need of Accessibility Education in the Software Industry: Insights from a Survey of Software Professionals in India. In: Proceedings of the 46th International Conference on Software Engineering: Software Engineering Education and Training [Internet]. pp. 212\u0026ndash;20. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1145/3639474.3640079\u003c/span\u003e\u003cspan address=\"10.1145/3639474.3640079\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAbraham CH, Sakyi-Badu G, Boadi‐Kusi SB, Morny E, Darko‐Takyi C, Ocansey S et al (2024) Simulation of visual impairment in persons with normal vision for scientific research. Ophthalmic Physiol Opt 44(2):442\u0026ndash;456\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMoharrer M, Tang X, Luo G (2020) With motion perception, good visual acuity may not be necessary for driving hazard detection. Translational Vis Sci Technol 9(13):18\u0026ndash;18\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePeters B, Bedrick S, Dudy S, Eddy B, Higger M, Kinsella M et al (2020) SSVEP BCI and eye tracking use by individuals with late-stage ALS and visual impairments. Front Hum Neurosci 14:595890\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eReddingius P, Crabb DP, Jones P (2024) Shopping with sight loss: Using a virtual reality shopping task to systematically quantify the impact of a scotoma size and eccentricity on everyday quality of life. Investig Ophthalmol Vis Sci 65(7):1849\u0026ndash;1849\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMann DL, Ravensbergen HJC (2018) International Paralympic Committee (IPC) and International Blind Sports Federation (IBSA) joint position stand on the sport-specific classification of athletes with vision impairment. Sports Med 48(9):2011\u0026ndash;2023\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRunswick OR, Rawlinson A, Allen PM, Sharpe BT, Pocock C, Datson N et al (2023) The effects of simulated vision impairment on performance in football. J Sports Sci 41(14):1410\u0026ndash;1422\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHeinrich SP, Str\u0026uuml;bin I (2019) Use of diffusing filters for artificially reducing visual acuity when testing equipment and procedures. Doc Ophthalmol 140(1):83\u0026ndash;93\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHecht H, H\u0026ouml;richs J, Sheldon S (2015) The effects of simulated vision impairments on the cone of gaze. Atten Percept Psychophys 77:2399\u0026ndash;2408\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKim SY, Moon BY, Cho HG, Yu DS (2022) Visual Acuity Outcomes in Diseases Associated with Reduced Visual Acuity: An Analysis of the National Health Insurance Service Database in Korea. Int J Environ Res Public Health 19(14):8689\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAl-Namaeh M (2021) Common causes of visual impairment in the elderly. Medical Hypothesis. Discovery Innov Ophthalmol 10(4):191\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEhrlich JR, Agarwal A, Young C, Lee J, Bloom DE (2022) The prevalence of vision impairment and blindness among older adults in India: findings from the Longitudinal Ageing Study in India. Nat aging 2(11):1000\u0026ndash;1007\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVinding T (1990) Visual impairment of age-related macular degeneration: An epidemiological study of 1000 aged individuals. Acta Ophthalmol 68(2):162\u0026ndash;167\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSchneck ME, Haegerstrom-Portnoy G, Lott LA, Brabyn JA, Gildengorin G (2004) Low contrast vision function predicts subsequent acuity loss in an aged population: the SKI study. Vision Res 44(20):2317\u0026ndash;2325\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVirtanen P, Gommers R, Oliphant TE, Haberland M, Reddy T, Cournapeau D et al (2020) SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat Methods 17(3):261\u0026ndash;272\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBailey IL, Jackson AJ (2016) Changes in the clinical measurement of visual acuity. Journal of Physics: Conference Series. ;772:012046\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGoodman-Deane J, Waller S, Collins AC, Clarkson J (2013) Simulating vision loss: what levels of impairment are actually represented? In: Contemporary Ergonomics and Human Factors 2013: Proceedings of the international conference on Ergonomics \u0026amp; Human Factors 2013. Cambridge, UK: CRC Press; p. 347\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBastawrous A, Rono HK, Livingstone IA, Weiss HA, Jordan S, Kuper H et al (2015) Development and validation of a smartphone-based visual acuity test (peek acuity) for clinical practice and community-based fieldwork. JAMA Ophthalmol 133(8):930\u0026ndash;937\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLovie-Kitchin JE (1988) Validity and reliability of visual acuity measurements. Ophthalmic Physiol Opt 8(4):363\u0026ndash;370\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOrganization WH (2019) Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.who.int/publications/i/item/9789241516570\u003c/span\u003e\u003cspan address=\"https://www.who.int/publications/i/item/9789241516570\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRidley J (2023) Acer Predator Helios 16 [Internet]. PC Gamer; Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.pcgamer.com/acer-predator-helios-16-review-benchmarks/\u003c/span\u003e\u003cspan address=\"https://www.pcgamer.com/acer-predator-helios-16-review-benchmarks/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMalkawi A, Kamaruddin A, Halin AA, Admodisastro N (2023) Listening to the Voice of People with Vision Impairment. Int J Adv Comput Sci Appl. ;14(4)\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShashkina V (2021) Assistive technology for the visually impaired: use cases and the particulars of development [Internet]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://itrexgroup.com/blog/assistive-technology-for-visually-impared-examples-development-tips/#header\u003c/span\u003e\u003cspan address=\"https://itrexgroup.com/blog/assistive-technology-for-visually-impared-examples-development-tips/#header\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJones PR, Somoske\u0026ouml;y T, Chow-Wing-Bom H (2020) Seeing other perspectives: evaluating the use of virtual and augmented reality to simulate visual impairments (OpenVisSim). npj Digit Med 3:32\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCheng L, Peng S, Hao H (2024) Effect of different screen brightness and devices on online visual acuity test. Graefes Arch Clin Exp Ophthalmol 262:641\u0026ndash;649\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNkeremuzor EC, Ikoro NC, Esenwah EC, Nzotta C, Ugwuoke GI, Nwaigwe OM et al (2020) Determination of Luminance Levels of Visual Acuity Charts in Selected Eye Clinics in Owerri and Their Effects on Vision. Asian J Res Rep Ophthalmol 3(2):13\u0026ndash;23\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRodriguez RG, Paviglianiti V, Dumit C, Pattini A (2024) Dos \u0026amp; Don\u0026rsquo;ts in Measuring Illuminance With Smartphones. Ergonomics in Design [Internet]. ;10648046241263631. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/10648046241263631\u003c/span\u003e\u003cspan address=\"10.1177/10648046241263631\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMohapatra S, Ganesh A, Zion N (2024) Assessment of Forward Head Posture in Information Technology Employees with Neck Pain: A Cross-Sectional Study. Journal of Orthopedics and Spine Trauma [Internet]. ; Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.18502/jost.v10i1.14962\u003c/span\u003e\u003cspan address=\"10.18502/jost.v10i1.14962\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"King's College London","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"visual acuity, simulation, low vision, accessibility, psychophysics","lastPublishedDoi":"10.21203/rs.3.rs-7957115/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7957115/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003ePurpose\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAccessible and scalable tools for simulating visual impairment are essential for studying accessibility and human-computer interaction under visual constraints. This study assessed whether Gaussian blur applied to electronic displays can serve as a valid and practical alternative to traditional simulation spectacles for inducing visual acuity loss.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethod\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThirty-seven participants completed and compared standard visual-acuity assessments using ETDRS charts and a screen-based tumbling-E test displayed on monitors with three pixel densities (94, 170, and 218 PPI). Tests were conducted under normal vision, moderate impairment, and severe impairment induced by simulation spectacles. Participants then performed 29 screen-based tumbling-E tests incorporating varying Gaussian-blur standard deviations (σ) to discover the relationship between σ and measured visual acuity loss.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGaussian blur strength exhibited a strong log-linear relationship with measured visual-acuity loss, with minimal influence from monitor pixel density. A pooled regression model explained over 90% of the variance in visual acuity and accurately predicted impairment levels across display types at different σ.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eScreen-based Gaussian blur offers a valid and reproducible method for simulating visual acuity loss. The derived regression model enables estimation of the blur level (σ) corresponding to specific degrees of impairment, facilitating consistent and controllable simulation across studies.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTranslational Relevance\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis approach supports remote experimentation and low-cost accessibility testing, providing researchers and designers with a scalable tool to investigate perceptual and attentional processes under degraded visual conditions.\u003c/p\u003e","manuscriptTitle":"Screen-based Visual Impairment Simulation: An Exploratory Investigation of Acuity Loss Simulation Using Gaussian Blur Filters","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-30 12:11:12","doi":"10.21203/rs.3.rs-7957115/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"9fcf8fd9-c1aa-4a56-9c3a-2e1f8f8cadde","owner":[],"postedDate":"October 30th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":56933712,"name":"Psychology"}],"tags":[],"updatedAt":"2025-10-30T12:11:13+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-30 12:11:12","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7957115","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7957115","identity":"rs-7957115","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}