Age-Related Changes in Mesopic Reading Vision Across Adulthood

Abstract

Purpose: Reading is indispensable for daily activities such as reading books, menus, and food labels, occurring under a wide range of luminance conditions from mesopic (dim light) to photopic (daylight). Despite its significance, there has been limited attention on age-related changes in mesopic reading vision. The current study aims to investigate how mesopic reading vision changes across adulthood.

Methods: Using the MNREAD iPad app, we assessed both mesopic (2 cd/m²) and photopic (220 cd/m²) reading vision in 157 normally-sighted individuals aged from 18 to 84, grouped into seven age groups. Reading vision was evaluated using four MNREAD parameters: maximum reading speed (MRS), critical print size (CPS), reading acuity (RA), and reading accessibility index (ACC).

Results: There was a significant age-related decline in reading vision under both mesopic and photopic conditions, with a more pronounced decline observed in mesopic conditions. The decline was linear from age 20 to 80: MRS decreased by 30 words-per-minute in mesopic conditions and 29 words-per-minute in photopic conditions; ACC declined by 0.18 (mesopic) and 0.12 (photopic); CPS declined by 0.3 logMAR (mesopic) and 0.16 logMAR (photopic); RA declined by 0.24 logMAR (mesopic) and 0.18 logMAR (photopic).

Conclusions: Our results show a monotonic decline in reading vision from ages 20 to 80 under both mesopic and photopic conditions, with a more pronounced decline in mesopic light. Given the significance of reading vision as a clinical measure, assessing reading vision under mesopic conditions may offer a more comprehensive evaluation of functional vision in everyday life.

In daily life, the human visual system operates in diverse luminance environments, spanning from dimly lit conditions (mesopic) to bright daylight (photopic). Mesopic vision, mediated by both cone and rod photoreceptors, typically operates under luminance levels of 0.01 to 3 cd/m². In contrast, photopic vision, mediated by cone photoreceptors, operates under luminance levels of 10 to 10⁸ cd/m². Despite the wide range of luminance levels encountered in everyday life, most clinical vision assessments are typically conducted under ideal, photopic light levels,¹ which may not fully represent the challenges encountered in real-world settings. A study, for example, showed that median home lighting was three to four times lower than clinical lighting environments, with 85% of participants having home lighting conditions below the recommended levels.² Importantly, previous studies using surveys and structured focus groups have shown that older adults are more likely to report difficulties with daily activities at night and under dim lighting conditions than young adults.^3–5 Owsley et al.⁶ demonstrated a significant association between older adults’ mesopic contrast sensitivity deficits and crash involvement, regardless of the time of day, whereas photopic contrast sensitivity showed no such association. Reading difficulties are a major issue reported by both older adults and visually impaired individuals.^4,5,7 Furthermore, despite having normal photopic visual acuity, our recent work^8,9 showed that older healthy controls, along with AMD and glaucoma patients, presented greater reading vision deficits under mesopic viewing conditions (2 cd/m²) compared to photopic conditions (220 cd/m²). 

These findings underscore the importance of assessing visual functions under low luminance conditions, particularly for reading. Reading is essential to daily life, playing an important role in numerous activities such as reading books and newspapers, interpreting traffic and street signs, navigating interfaces on mobile or desktop devices, and reading menus or medicine labels. Consequently, reading is a fundamental aspect of vision-related quality of life.⁴ 

Normal aging alters the optical,^10–13 retinal,^14,15 and cortical^16,17 properties of the human visual system. These alterations include changes in the optical density of the crystalline lens and pupil size (pupillary miosis),^10–12 as well as neural loss¹⁴ in photoreceptors,¹⁵ retinal ganglion cells,^18–20 and cortical neurons.^16,17 For example, Curcio and Drucker²¹ reported a 25% reduction in the density of retinal ganglion cells subserving the central 11° of the visual field in healthy older adults compared to young adults. The volume of the human lateral geniculate nucleus, which receives information directly from the ascending retinal ganglion cells via the optic tract, also decreases by approximately 15% between the ages of 20 and 70, indicating age-related changes in the lateral geniculate nucleus.¹⁷ Furthermore, Curcio et al.¹⁵ showed that although cone density remained stable with age, rod density decreased by 30% from midlife, particularly in the zone 0.5 to 3 mm from the fovea. The selective vulnerability of rods in the central retina, as opposed to cones, suggests age-related declines in mesopic and scotopic sensitivity. Consistent with the aforementioned age-related structural and neural changes, a decline in various visual functions has been well documented,⁵ including reading speed,^22,23 visual acuity,^24–26 contrast sensitivity,^{10,18,27–30} visual processing speed,^31–33 dark adaptation,^34–36 crowding,^22,37 visual span²² (i.e., the number of letters that can be recognizable at one glance), and the functional field of view.³⁷ 

Using a wide age range (8–81 years), Calabrèse et al.²³ demonstrated that MNREAD reading vision, as indicated by both maximum reading speed and the print size requirement for achieving maximum reading speed, declines with age. Specifically, maximum reading speed decreases by approximately 16% from age 20 to 80, whereas the print size requirement increases by about 48% over the same age range. Understanding this age-related decline in reading vision is important because it reflects complex perceptual processes influenced by various factors such as visual crowding (i.e., the ability to recognize a target in clutter), visual acuity, contrast sensitivity, visual span, and oculomotor control, which are not captured by simple visual acuity measures (i.e., the ability to recognize isolated, high-contrast letters).^38–41 It is important to note that MNREAD reading vision parameters (detailed below) are assessed in a reading context. For instance, reading acuity represents the smallest print size a person can read with minimal significant errors, whereas critical print size indicates the smallest print size required to achieve maximum reading speed. This approach qualitatively differs from traditional visual acuity tests, leading to outcomes specific to reading performance.^42,43 Consistent with this view, Liu et al.²² demonstrated that age-related decline in reading speed is closely linked to increased crowding and a shrinkage of visual span (i.e., the number of letters recognizable at a glance).²² They found significant correlations between reading speed and crowding zone (r = −0.39, P = 0.01), as well as between reading speed and visual span size (r = 0.47, P < 0.01). However, no statistically significant correlation was observed between reading speed and visual acuity or contrast sensitivity. Although informative, these reading studies were conducted only under photopic conditions, leaving a gap in our understanding of age-related changes in reading vision under dim or mesopic light conditions. 

In this study, we are particularly interested in investigating whether the rate of decline in MNREAD reading vision across adulthood is more pronounced under mesopic conditions than under photopic conditions and whether there is a specific age range at which this decline becomes more significant. Our previous work has shown that eye movements, such as saccade latencies and microsaccade rates during reading, differ under low luminance conditions compared to photopic conditions.⁴⁴ These findings collectively suggest that the rate of age-related decline in reading vision may differ under mesopic conditions due to the additional optical, neural and oculomotor inefficiencies introduced in such low-light environments. 

To address these questions, we investigated the effect of age on reading vision under mesopic and photopic viewing conditions in a population of normally sighted individuals aged 18 to 84 years old. Reading vision was assessed using the MNREAD iPad app,⁴⁵ which is the digital version of the well-established MNREAD chart.⁴⁶ The MNREAD chart is a widely used continuous-text reading acuity test⁴⁶ used for evaluating eye disorders, treatments, and vision rehabilitation. The MNREAD test consists of short (10 standard-word sentences) and simple sentences, effectively minimizing the influence of higher-level linguistic and cognitive factors while focusing on the impact of visual factors, such as print size,⁸ on reading performance. The MNREAD test evaluates reading performance through four parameters: (i) maximum reading speed (MRS), measured in words per minute (wpm); (ii) critical print size (CPS), the smallest print size at which a participant can read at maximum speed; (iii) reading acuity (RA), the smallest print size readable without substantial errors; and (iv) reading accessibility index (ACC), indicating access to text across everyday print sizes. 

In short, the current study aims to evaluate how age impacts MNREAD reading vision under mesopic and photopic conditions across adulthood. The outcome of the current study will highlight the challenges posed by age-related declines in reading vision, particularly under low luminance conditions. Given the significance of reading performance as an outcome measure in evaluating treatments,⁴⁷ surgical procedures,^48,49 and rehabilitation techniques,^50,51 the current study provides valuable insights into strategies for assessing and improving visual function across different lighting environments. 

Methods

Participants

All participants had normal retinal health (i.e., no AMD, glaucoma, or diabetic retinopathy in either eye) and normal binocular vision with no known history of neurological or ocular diseases, except for cataract surgery. Four participants over the age of 60 reported having undergone cataract surgery. Participants had normal or corrected-to-normal vision, defined as having a corrected visual acuity of better than or equal to 0.15 logMAR (equivalent to 20/28 Snellen notation) and a contrast sensitivity of better than or equal to 1.65 log units. 

Additionally, all participants were native or fluent English speakers without known cognitive or neurologic impairments, confirmed by the Mini Mental Status Exam (MMSE; ≥ 25 MMSE score). Exclusion criteria included the following: (i) Diagnosis of age-related macular degeneration, other retinal conditions, optic nerve conditions, or corneal disease; (ii) Diagnosis of diabetic eye disease; (iii) Diagnosis of Alzheimer's disease, Parkinson's disease, brain injury, or other neurological or psychiatric conditions as revealed by the medical record or self-report; (iv) Known dyslexia; (v) Non-native English speakers. 

The experimental protocols followed the tenets of the Declaration of Helsinki and were approved by the Institutional Review Board of Northeastern University. Written informed consent was obtained from all subjects before the experiment and after explanation of the nature and possible consequences of the study. 

Apparatus

The MNREAD test was administered with the MNREAD iPad app (1.2 version; © 2017 University of Minnesota; https://apps.apple.com/us/app/mnread/id1196638274) running on an iPad Air 2 tablet with Retina display (2048 × 1536-pixel resolution at 264 ppi; Apple, Inc., Cupertino, CA, USA). The iPad was mounted vertically on a stand in landscape mode. For photopic conditions, iPad screen brightness was set to a value of 75 (equivalent to 220 cd/m²). Testing under mesopic screen luminance (2 cd/m²) was carried out by setting the iPad screen brightness to a value of 0 (equivalent to 3.9 cd/m²) in combination with a neutral density filter (Kodak ND 0.30; Kodak, Rochester, NY, USA) applied on the screen. The filter had a factor of 2 luminance reduction and the luminance attenuation of the filter was confirmed with photometric readings from a luminance meter (Minolta LS-110 Luminance Meter; Konica Minolta, Inc., Tokyo, Japan). All the text was black on a uniform white background with a luminance contrast of 99%. The luminance contrast of the text remained unchanged for both mesopic and photopic conditions. 

Study Design and Task Procedure

The study design and task procedure are identical to our previous work.⁹ However, for completeness, here we provide a detailed summary. 

For each participant, the MNREAD test was administered under mesopic (2 cd/m²) and photopic conditions (220 cd/m²). The MNREAD test sequence was photopic-mesopic-mesopic-photopic to minimize fatigue and training effects. Different MNREAD charts were used for each test to prevent participants from reading the same text twice. The average of the two measurements for each viewing condition (mesopic vs. photopic) was reported. 

Participants read binocularly at a distance of 40 to 80 cm, adjusted to ensure a proper plateau and drop-off point on the MNREAD curve (Fig. 2A) from which we could extract reliable MNREAD parameters. Near refractive error was corrected using either habitual reading glasses or trial lens refraction. Participants with bifocal or progressive lenses used the appropriate portion of their lens. 

The reading task was performed in a dimly lit room (approximately 0.3 cd/m²), with an eight-minute adaptation period before the main task. We opted to conduct the task in a dimly lit room to ensure that the iPad screen was the sole source of luminance. This approach allowed us to control external lighting variables and objectively assess reading vision under both photopic and mesopic conditions without the influence of ambient light. Sentences were presented one at a time on an iPad screen at a fixed viewing distance. The MNREAD chart uses 10-word sentences to determine reading speed across print sizes that decrease logarithmically in steps of 0.1 log units. 

The iPad app follows the same format as the standard print chart, using three-line sentences with a reduced range of print sizes (14 sentences vs. 19 in the printed version). Print sizes ranged from 6.3 M to 0.32 M (Sloan M notation), corresponding to angular print sizes from 1.2 to −0.1 logMAR at 40 cm. Sentences were initiated by the experimenter and displayed instantly. Participants read aloud from largest to smallest print size, with errors recorded after each sentence. Testing ended when print size was unreadable. Reading speed was calculated excluding missed or incorrectly read words, with more than 10 errors equating to zero speed. Errors were also used to estimate reading acuity (acuity = 1.4 − (sentences × 0.1) + (errors × 0.01)). 

The app displayed the MNREAD curve of log reading speed vs. print size and provided four parameters: MRS, CPS, RA, and ACC. ACC (0-1 scale) is the mean reading speed across the 10 largest print sizes (0.4 to 1.3 logMAR at 40 cm), normalized by 200 wpm. This normalization value is predefined and utilized by the MNREAD iPad App for computing the reading accessibility index, reflecting the average reading speed for normally sighted young adults, as established in previous research.⁵² An ACC value of 1 indicates normal reading performance, whereas a value of 0 signifies an inability to read any sentences in the range. 

For each participant, we measured contrast sensitivity and distance visual acuity both monocularly and binocularly under photopic light-levels. Distance visual acuity was assessed at 2 meters with ETDRS charts,⁵³ and contrast sensitivity at 1 meter using Pelli-Robson charts, scored by the letter-by-letter method (0.05 log units per correct letter).⁵⁴ Binocular near visual acuity was measured with trial lenses or habitual near prescription using a near ETDRS chart. Chart luminance for ETDRS and Pelli-Robson tests ranged from 120 to 145 cd/m². We ensured the use of near-binocular visual acuity and contrast sensitivity measurements, because our primary reading test was designed for near viewing. 

Data Analysis

In this study, “reading vision” refers to reading performance indicated by the four MNREAD parameters: MRS, ACC, CPS, and RA. First, to determine whether there is any difference in reading vision between mesopic and photopic viewing conditions across all participants, independent of age, we conducted a paired samples t-test for each MNREAD parameter. We then analyzed the relationship between reading vision and age by grouping participant's reading vision data into seven age groups (in decades), ranging from the 20s to the 70s. To test for the effects of age group (seven age groups) and viewing condition (mesopic vs. photopic) on reading vision, as well as their interaction effects, we performed a two-way ANOVA on reading vision with age group as a between-subject factor and viewing condition as a within-subject factor. We conducted a separate ANOVA analysis for each MNREAD parameter. To further analyze the relationship between reading vision and age, we also conducted a simple linear regression analysis, using reading vision as the dependent variable and age group as the independent variable. This analysis enabled us to quantify the rate of change in reading vision across different age groups. Finally, statistical analyses and data visualization were performed using MATLAB R2022b (The MathWorks Inc., Natick, MA, USA). 

Results

A total of 157 subjects participated in this study, of which 85 were males. All participants had normal or corrected-to-normal vision (see the Methods section for details). Participants’ ages ranged from 18 to 84 years (mean age 44 ± 20) and were grouped into seven age groups, as shown in Table 1. 

Table 1.

 

Characteristics of Study Participants (Mean ± Standard Deviation)

 

Across participants, the mean binocular near visual acuity was −0.04 ± 0.09 logMAR and the mean binocular distance acuity was −0.07 ± 0.10 logMAR. The mean distance visual acuity was −0.01 ± 0.13 logMAR for the right eye and −0.01 ± 0.12 logMAR for the left eye. The mean binocular contrast sensitivity was 1.85 ± 0.14 log units. The mean contrast sensitivity was 1.63 ± 0.15 log units for the right eye and 1.64 ± 0.13 log units for the left eye. 

Next, we present the observed age-related changes in binocular near visual acuity, binocular distance visual acuity, and binocular contrast sensitivity measured under photopic conditions. Consistent with previous studies,^{25,26,28–30} there was a significant age-related decline in binocular near visual acuity (F_(1,5) = 22.4, P < 0.001), binocular distance visual acuity (F_(1,5) = 53.4, P < 0.001), and in contrast sensitivity (F_(1,5) = 27, P < 0.001). As depicted in Figure 1, binocular near visual acuity declined by 0.029 logMAR units each decade, binocular distance visual acuity declined by 0.032 logMAR units per decade, whereas binocular contrast sensitivity decreased by 0.031 log units per decade. These changes reflect an approximate decline of 0.2 logMAR in visual acuity from age 20 to age 80, the equivalent to a two-line decline on the ETDRS chart, and a 0.2 log unit reduction in contrast sensitivity (about a triplet and a third on the Pelli-Robson chart) over the same period. Notably, a decrease of 0.2 log units corresponds to a 58% reduction in both photopic visual acuity and contrast sensitivity. 

Figure 1.

 

Age-related decline in photopic visual acuity (VA) and contrast sensitivity (CS). Binocular near visual acuity (blue color) and binocular distance visual acuity (green color) are plotted on the right y-axis, whereas binocular contrast sensitivity (orange color) is plotted on the left y-axis, all against age group. For the analysis, participants’ ages were grouped into seven bins spanning one decade each. Each dot represents the average of each age group. The dashed lines depict the regression lines for visual acuity and contrast sensitivity across age groups. Error bars: ±1 SEM; the shaded region represents the 68% CI for the regression line.

It is important to note that even modest age-related declines in contrast sensitivity (0.2 log unit) can affect daily activities such as reading or driving in low-contrast environments.^55–58 Similarly, a 0.2 logMAR decline in visual acuity could push individuals who are borderline for driving standards below the threshold, impairing their ability to perform tasks that require clear distance vision, such as reading street signs and navigating roads. This could ultimately impact their mobility and independence.⁵⁹ These concerns are particularly relevant under mesopic conditions, where declines in both visual acuity and contrast sensitivity tend to be more pronounced.^6,60 

Significant Decrease in Reading Vision From Photopic to Mesopic Conditions, Even Among Individuals With Normal Vision, Independent of Age

Before analyzing how age, viewing condition, and their interaction affects reading vision, we first checked whether there are any significant differences in reading vision between mesopic and photopic conditions for all participants, independent of age. Figures 2B to 2E plots the average MNREAD parameter values for mesopic and photopic viewing conditions across all participants. Our pair-wise t-test showed that mesopic reading vision is significantly different from photopic reading vision for all MNREAD parameters (P < 0.05). More specifically, on average, participants had slower reading speed (by 2 wpm, 95% CI = [0.67, 3.49], P = 0.004), reduced reading accessibility (by 0.04, 95% CI = [0.03, 0.05], P < 0.001), larger critical print size (by 0.23 logMAR, 95% CI = [0.21, 0.25], P < 0.001), and worse reading acuity (by 0.15 logMAR, 95% CI = [0.13, 0.16], P < 0.001) under mesopic conditions compared to photopic conditions. It is noteworthy that an increase of 0.15 or 0.23 logMAR in print size under mesopic conditions corresponds to approximately a two-line difference on the visual acuity chart, which is considered clinically significant. The observed increase in critical print size under mesopic conditions suggests that participants required larger text to achieve their maximum reading speed in dim light. Importantly, the relatively small decrease in maximum reading speed under mesopic conditions (i.e., two words per minute) indicates that individuals can still maintain reading speeds comparable to those achieved under photopic conditions, as long as the print size is appropriately adjusted. Furthermore, although a decrease of two words per minute in reading speed may seem clinically insignificant for the normally sighted population, it can have a profound impact on individuals with visual impairments, who typically begin with a lower baseline reading speed. For these individuals, even minor reductions can significantly diminish reading efficiency and exacerbate existing challenges.⁴³ The results are summarized in Table 2. 

Figure 2.

 

(A) Examples of an MNREAD curve (i.e., reading speed as a function of print size) from which the four MNREAD parameters were extracted: MRS, ACC, CPS, and RA. Any impairment in MNREAD reading vision can be represented as a downward and/or rightward shift of the MNREAD curve (e.g., orange curve vs. blue curve). (B) to (E) plot the differences in MNREAD parameter values between mesopic (blue bar) and photopic (orange bar) conditions for normally sighted participants. The bar graph indicates the mean value of each MNREAD parameter collapsed across subjects. Gray circles represent the participants responses. (B) Maximum reading speed (wpm). (C) Reading accessibility index. (D) Critical print size (logMAR). (E) Reading acuity (logMAR). Error bars: ±1 SEM. ***P < 0.005.

Table 2.

 

Mean MNREAD Parameter Values for Two Viewing Conditions (n = 157)

 

Significant Age-Related Decline in Reading Vision for Both Mesopic and Photopic Conditions, But the Decline Is More Pronounced Under Mesopic Conditions

Next, we examined how age affects reading vision under mesopic and photopic conditions. We conducted a two-way ANOVA using age group (seven groups) as a between-subject factor and viewing condition (mesopic vs. photopic) as a within-subject factor. Our ANOVA analysis revealed that there was a main effect of age group on reading vision for all MNREAD parameters: maximum reading speed (F_(6,150) = 6.48, P < 0.001), critical print size (F_(6,150) = 8.53, P < 0.001), reading accessibility index (F_(6,150) = 6.37, P < 0.001), and reading acuity (F_(6,150) = 22.57, P < 0.001). These results suggest a significant difference in reading vision among the different age groups, averaged across viewing conditions. 

We also observed a significant main effect of viewing condition for all MNREAD parameters: maximum reading speed (F_(1,150) = 8.85, P < 0.003), critical print size (F_(1,150) = 383.54, P < 0.001), reading accessibility index (F_(1,150) = 106.37, P < 0.001), and reading acuity (F_(1,150) = 543.69, P < 0.001). Overall, these results suggest a significant difference in reading vision between the two viewing conditions, averaged across all age groups. 

We did not find any significant interaction effects between viewing condition and age group for maximum reading speed (F_(6,150) = 13.79, P < 0.333) and reading accessibility index (F_(6,150) = 1.71, P < 0.121). However, we found a significant interaction effect for the critical print size (F_(6,150) = 2.20, P < 0.05) and reading acuity (F_(6,150) = 9.51, P < 0.001). The estimated difference in reading acuity between mesopic and photopic conditions was 0.16 logMAR for older adults (60–80 years) and 0.07 logMAR for younger adults (<20 years), which correspond to approximately a two-line difference and a one-line difference on the visual acuity chart, respectively. A similar twofold difference between the mesopic and photopic viewing conditions was observed in critical print size: 0.28 logMAR for older adults (60–80 years) and 0.16 logMAR for younger adults (<20 years). These results suggest that the print size requirement for reading, as indicated by critical print size and reading acuity, is more significantly impacted by low luminance conditions in older adults compared to younger adults. However, maximum reading speed is consistently affected by low luminance conditions across all age groups. 

Next, to better quantify the age-related decline in reading vision, we conducted a simple regression analysis of each parameter value on age group for both mesopic and photopic conditions. In Figures 3A to 3D, we plotted each MNREAD parameter against individual ages in the left column and age groups in the right column. Table 3 summarizes the mean values for the MNREAD parameters for each age group under both viewing conditions. The regression analysis of reading vision on age group revealed a significant age-related decline across all MNREAD parameters for both mesopic and photopic conditions (P < 0.05). The dashed lines in each panel of the right column represent the binned reading vision data regressed on age group, with the best linear fit applied. Our model fits indicate that the rate of decline is both monotonic and linear across the adult lifespan, with no observable age range where this decline becomes more pronounced. 

Figure 3.

 

(A) The left panel represents maximum reading speed as a function of age for both mesopic (blue) and photopic (orange) conditions, with each dot representing data from one participant. Each solid line represents the moving average (using a window size of five years) of maximum reading speed across age for each viewing condition. The shaded region around the lines corresponds to ±1 SEM. The right panel illustrates the regression of maximum reading speed on age group. For the analysis, participants’ ages were grouped into seven bins representing one decade. Each dot corresponds to the group average. Error bars: ±1 SEM. The shaded region around the line represents the 95% CI for the regression line. (B) Reading accessibility index. (C) Critical print size (logMAR). (D) Reading acuity (logMAR).

Table 3.

 

MNREAD Parameter Values (Mean ± SEM) for Each Age Group Under Both Viewing Conditions

 

More specifically, our results show that maximum reading speed decreases by approximately 5 words per minute for each decade of age under both mesopic (F_(1,5) = 29.5, P < 0.001, r² = 0.86) and photopic conditions (F_(1,5) = 22.7, P < 0.001, r² = 0.82). From age 20 to age 80, the overall maximum reading speed declines by 30 wpm under mesopic conditions and by 29 wpm under photopic conditions. 

Similarly, we also observe that for each decade of age, critical print size increases by 0.05 logMAR (F_(1,5) = 98.8, P < 0.001, r² = 0.95) under mesopic conditions and by 0.03 logMAR (F_(1,5) = 58.3, P < 0.001, r² = 0.92) under photopic conditions. Across the entire age range from 20 to 80 years, this results in a total increase in critical print size of 0.3 logMAR (equivalent to a three-line difference on the visual acuity chart) under mesopic conditions, compared to an increase of 0.16 logMAR (approximately a two-line difference) under photopic conditions. 

We also find that for each decade of age, reading acuity declines by 0.04 logMAR (F_(1,5) = 77, P < 0.001, r² = 0.94) under mesopic conditions and by 0.03 logMAR (F_(1,5) = 297, P < 0.001, r² = 0.98) under photopic conditions. As a result, from age 20 to age 80, overall reading acuity declines by 0.24 logMAR under mesopic conditions and by 0.18 logMAR under photopic conditions, both corresponding to approximately a two-line decline on the visual acuity chart. 

Furthermore, age-related decline in reading accessibility index was also statistically significant: it decreases by 0.03 units per decade under mesopic (F_(1,5) = 16.8, P < 0.01, r² = 0.77) and by 0.02 units per decade for photopic conditions (F_(1,5) = 29.5, P < 0.001, r² = 0.82). Consequently, from age 20 to age 80, reading accessibility index decreases by 0.18 units under mesopic conditions and by 0.12 units under photopic conditions. 

Taken together, the results suggest that reading vision linearly declines with age leading to slower reading speed, reduced reading accessibility, larger critical print size requirements, and worsened reading acuity for both mesopic and photopic conditions. Importantly, this age-related decline in reading vision is more pronounced under mesopic conditions compared to photopic conditions. 

Discussion

People engage in activities under a broad spectrum of illumination conditions, ranging from photopic (e.g., daylight) to mesopic (e.g., dusk) light levels. For instance, most nighttime outdoors and street lighting conditions fall within mesopic luminance values.^61,62 Unsurprisingly, visual functions such as acuity and contrast sensitivity are often compromised under dim light conditions,⁶³ particularly in older adults and visually impaired individuals.^56,57,60,64 As a result, older adults, even those with normal vision, frequently report significant difficulties performing everyday tasks, including reading, in dim lighting conditions.^3,4,65 Our previous work^8,9 has shown that even older adults with healthy vision exhibit significantly compromised MNREAD reading vision under mesopic conditions when compared to photopic conditions, though the decrement is not as pronounced as in older adults with visual impairments such as glaucoma or age-related macular degeneration. 

The current study was undertaken to investigate the effects of age on reading vision under mesopic and photopic conditions in a cohort of normally sighted individuals aged 18 to 84 years, using the MNREAD iPad app,⁴⁵ the digital version of the MNREAD chart.⁴⁶ We were particularly keen to determine if deterioration of reading vision (if any) accelerates more under mesopic conditions compared to photopic conditions across adulthood even in individuals with normal vision. 

We observed a linear decline in reading vision with increasing age from 20 to 80 years under both mesopic and photopic conditions, with the decline being more pronounced in mesopic conditions. Specifically, from age 20 to 80, maximum reading speed decreased by 30 wpm under mesopic conditions and by 29 wpm under photopic conditions. Over the same age range, critical print size increased by 0.3 logMAR (equivalent to three lines on the visual acuity chart) under mesopic conditions, compared to 0.16 logMAR (two lines) under photopic conditions. Reading acuity also declined by 0.24 logMAR in mesopic conditions and by 0.18 logMAR in photopic conditions, both corresponding to a two-line decrease on the visual acuity chart. In addition, the reading accessibility index decreased by 0.18 units under mesopic conditions and by 0.12 units under photopic conditions. Importantly, the 0.14 logMAR increase in critical print size under mesopic conditions indicates that readers required larger text to maintain their maximum reading speed in dim light. However, the small reduction in reading speed (2 wpm) under mesopic conditions suggests that once the print size is adjusted, individuals can still read at speeds comparable to those achieved under photopic conditions. 

Using the chart version of the MNREAD test,⁴⁶ Calabrese et al.²³ investigated the impact of age on reading vision at photopic light levels (200 cd/m²) in individuals aged 8 to 81. They found that all four MNREAD parameters declined with age, and established baseline normative values for different age groups. The current work measured and compared the relationship between age and reading vision under photopic and mesopic viewing conditions. This allowed us to establish MNREAD baseline values not only for photopic viewing conditions but also for mesopic conditions, highlighting a greater age-related decline in reading vision under dim light. These findings may also be useful for assessing vision in clinical settings across the age span. The photopic condition baseline values in the present study differed slightly from those reported in Calabrese et al.²³ For example, they reported an average MRS of 202 wpm for young adults, whereas our study found it to be 178 wpm. This decrease of approximately 12% is likely to be attributed to the differences previously observed between the MNREAD chart and its digital version.⁴⁵ Additionally, in their study, the MRS decreased by 6 wpm each decade for age groups 40 to 80, whereas in our work, we found a decline of 5 wpm per decade. This small discrepancy may result from their use of a bilinear model, which aimed to capture a plateau in young adults’ responses (aged 16 to 40) and a decline in responses among older adults (40 to 80). However, their results were also well-fitted by a simple linear model²³ from age 16 to 80, resulting in a linear relationship with a slightly shallower slope, which aligns with our findings. Furthermore, the ACC index obtained in our study was generally lower than theirs, which may also be due to the differences between the chart and digital versions of the MNREAD test.⁴⁵ Given the advantages of the MNREAD iPad app over the MNREAD chart,⁴⁵ including increased intertester reliability and significantly shorter testing time, our work provides a valuable update on baseline data regarding age-related changes in reading vision under both photopic and mesopic viewing conditions. 

The observed age-related decline in reading vision under mesopic viewing conditions, as reported in our study, may be attributed to several optical and neural-processing factors. For example, at mesopic light levels, reading vision is mediated by both cone and rod photoreceptors. Reading is primarily carried out by central vision, where the fovea (i.e., the central 1.7° visual field) contains a high density of cone photoreceptors, offering high spatial resolution.⁶⁶ However, the parafoveal region (i.e., from 1° to 2.5° eccentric to the fovea), containing both cones and a high density of rod photoreceptors,⁶⁷ also plays a critical role in efficient reading behavior.^68,69 It aids in tasks such as previewing upcoming words, planning saccades, and facilitating subsequent foveal processing.^70–72 However, rods provide lower spatial resolution than cones, and at lower light intensity, cones are not as effective at detecting contrast because they are under photopic conditions.^73,74 The result of this suboptimal functioning of cones and rods is a reduction in visual acuity and contrast sensitivity, leading to a diminished ability to distinguish fine details and clear text, both of which are essential for effective reading. Additionally, in low-light conditions, reduced visual acuity and contrast sensitivity also result from a decreased signal-to-noise ratio due to increased neural noise,^75,76 which interferes with accurate signal transmission. Moderate pupil dilation under low light can further introduce optical aberrations, such as spherical and chromatic aberrations, which can degrade image sharpness.⁷⁷ As age progresses, other optical and neural factors contribute to the decline of reading vision. For example, age-related changes to the optics of the eye, such as increased lens density,^12,13 pupillary miosis,¹¹ and increased light scatter, and aberration,^10,78 affect how light is transmitted to the retina—effects that becomes particularly significant under mesopic viewing conditions. This results in an overall reduction of retinal image quality, particularly at higher spatial frequencies, where fine details and sharp edges are critical for reliable pattern recognition. 

Besides these optical changes, aging also affects neural mechanisms, including a decline in cone sensitivity⁷⁹ and reduced photon absorption efficiency, with older adult cones absorbing four times fewer photons than those of young adults.⁸⁰ In addition, significant age-related loss of retinal ganglion cells⁸¹ contributes to these effects. As a result, aging degrades visual acuity,^24–26 contrast sensitivity,^10,27,28 and other visual functions,⁵ directly impacting reading performance. Particularly relevant for mesopic vision, previous studies have shown that the density of rod photoreceptors^15,34 and the regeneration of rhodopsin⁸² decline with age, which might significantly contribute to the impairment in reading vision among older adults in low-light conditions. Taken together, the interaction of these optical and neural processes may explain the observed age-related decline in reading vision at both mesopic and photopic light levels, with a more pronounced decline under mesopic conditions. 

Among the limitations to our study is the use of short-standard sentences for reading. Given that sustained reading is closely linked to reading fatigue in older adults with visual impairment,⁸³ the challenges presented by mesopic conditions may be pronounced during long-passage reading, particularly for older adults. In addition, our reading measures required participants to read out loud, which may have led to an underestimation of overall reading speed and an underestimation of the decline in reading speed with age. We also acknowledge that some participants, particularly those aged 60 years and older, may have had unreported or undiagnosed eye diseases, including cataracts, which could have impacted our results. While the presence of media opacities likely reflects real-world conditions commonly observed in older adults, controlling for lens status in future studies would help clarify the specific contribution of lens changes to age-related declines in mesopic and photopic reading vision. Another limitation is the relatively small sample size, particularly among participants over 70 years old, which should be considered when interpreting the findings. Nonetheless, our study provides valuable initial data and contributes to the foundation of a larger normative database. In future studies, we aim to recruit a broader sample to enhance the generalizability of our results. Furthermore, we only assessed one luminance level within the mesopic range. Although this controlled condition allowed us to systematically examine reading performance in reduced light, it may not fully replicate the diverse lighting environments encountered in daily life, where ambient lighting, screen brightness, and task lighting vary significantly. Future research could examine a range of mesopic light levels to better reflect real-world lighting variability. Despite these limitations, our study is the first to demonstrate an age-related decline in mesopic reading vision across adulthood. This finding is particularly relevant for older individuals who report difficulty reading under dim light despite having normal photopic visual acuity. 

In summary, our study demonstrates that reading vision declines monotonically from ages 20 to 80 under both photopic and mesopic lighting conditions, with a more pronounced decline under mesopic conditions. Reading is indispensable to daily life and represents a fundamental aspect of vision-related quality of life.⁴ Therefore, understanding age-related changes in reading vision under mesopic conditions for normally sighted individuals is important for developing effective strategies to support visual function in low-light environments and informing the design of interventions and assistive technologies that cater to the needs of an aging population. Furthermore, given the importance of reading vision as a clinical outcome measure, our findings suggest that assessing reading vision under mesopic conditions may provide a more comprehensive evaluation of a person's functional vision in daily life. 

Acknowledgments

Supported by NIH/NEI Grant R01 EY027857, R01EY029595, P30EY03039, Research to Prevent Blindness (RPB)/Lions’ Clubs International Foundation (LCIF) Low Vision Research Award, RPB, Dorsett Davis Discovery Fund, Alfreda J. Schueler Trust, and EyeSight Foundation of Alabama. 

Disclosure: B. Peñaloza, None; T.-L. Goddin, None; D.S. Friedman, None; C. Owsley, None; M.Y. Kwon, None