Abstract
Purpose:
To evaluate spontaneous eye blink rate (SEBR) and percentage of incomplete blinks in different hard-copy and visual display terminal (VDT) reading conditions, compared with baseline conditions.
Methods:
A sample of 50 participants (29 females, age range, 18–74 years) were recruited for this study. All participants had good ocular health and reported no symptoms of dry eye (OSDI score < 15). Face video recordings were captured while participants observed in silence a landscape picture at 2 m (baseline) and during six different, 6-minute controlled reading experimental conditions. Texts were presented in electronic (tablet and computer display at 100% and 330% zoom levels) and hard-copy (text in book position in silence and aloud and text pasted on the computer display) formats. Video analysis was subsequently conducted to assess blink parameters.
Results:
All reading conditions resulted in a decrease in SEBR when compared with baseline conditions (all P < 0.001), with the least negative impact corresponding to reading in a 330% expanded display. The percentage of incomplete blinks was found to increase when reading was conducted on an electronic platform, in contrast to hard-copy text.
Conclusions:
The high cognitive demands associated with a reading task led to a reduction in SEBR, irrespective of type of reading platform. However, only electronic reading resulted in an increase in the percentage of incomplete blinks, which may account for the symptoms experienced by VDT users.
Spanish Abstract
It may be argued that reading text presented in hard-copy or electronic formats is one of the most common cognitive demanding near-vision tasks. Researchers from different disciplines agree that the choice of reading platform, however, is not trivial.
1–5 For example, while dry eye is a frequently reported symptom amongst visual display terminal (VDT) users,
1–4 reading in paper format has not been traditionally associated with complaints of dry eye.
Visual fatigue among VDT users was first documented by Hultgren and Knave in 1974,
6 with symptoms increasing toward the end of the day.
7 Several factors may account for the differences in visual fatigue among reading platforms. First, ocular exposure, which results in tear film evaporation,
8 is influenced by actual screen position, and is more relevant for desktop computers.
9 It may be noted that ergonomic recommendations for these devices suggest that the center of the display should be placed slightly lower than the horizontal line of sight. In contrast, laptop users have been observed to place their devices in a variety of positions,
10 usually opting for a lower position, compared with desktop computer users. As for tablets, e-books, and other handheld devices such as smartphones, they are usually viewed in inferior gaze, similar to traditional printed reading material, and at a shorter distance, although user preferences also may vary.
10,11 Although ergonomic recommendations aim at reducing postural related symptoms arising from prolonged computer use, as far as we know, they do not consider the relationship between display position and ocular surface exposure.
Second, subtle differences between hard-copy and electronic formats have been observed in spontaneous eye blink rate (SEBR) and blink amplitude (complete or incomplete). Changes in SEBR have been documented to be modulated by fine motor controls, speech centers, emotional states, cognitive demands, and attention,
8,9,12–14 although previous research has also revealed a possible influence of other, device-related factors on SEBR. Thus, Benedetto and colleagues
15 compared a liquid crystal display (LCD) tablet, an electronic ink reader (E-ink) and a paper book, describing an overall subjective preference for the paper book, with the LCD tablet presenting the worst results in terms of visual fatigue and reduction in SEBR, which the authors attributed to the higher level of luminance emitted by the LCD device. On the contrary, Chu et al.,
16 while also reporting higher levels of discomfort (in terms of blurred vision) when viewing a text on the computer screen, did not uncover any difference in SEBR between both conditions.
17 Interestingly, however, they described a higher percentage of incomplete blinks during computer use, which may have accounted for ocular discomfort. Other authors have also documented that incomplete blinking, rather than an actual decrease in SEBR, is the main contributory factor of dry eye symptoms, further supporting the role of blink amplitude in visual fatigue.
18
Lastly, reading involves horizontal saccade eye movements, followed by fixations. To maintain stable and continuous vision, saccades are accompanied by a certain degree of visual suppression,
19 the depth of which depends on the actual amplitude of the saccade.
20 Thus, for small amplitude saccades, such as those involved in reading, visual suppression is effective in stabilizing vision. However, visual suppression in saccades with amplitude larger than 33° is less effective, often requiring a coupled eye blink, with its corresponding suppression, to maintain visual stability.
21 It may be hypothesized that reading larger text presented in a panoramic display may result in an increase of the percentage of large amplitude horizontal saccades which, in turn, may introduce changes in SEBR.
The aim of the present study was to explore SEBR and blink amplitude from video recordings of a sample of nondry eye subjects while they were reading the same text in six different reading configurations including three hard-copy (A4 text pasted on a desktop display; A4 text in normal lower gaze reading position in silent reading; A4 text in normal lower gaze reading position reading aloud) and three electronic (text at a 100% zoom level presented on a panoramic desktop display [PC100]; text at a 330% zoom level presented on a panoramic desktop display [PC330]; text displayed on a tablet in normal lower gaze reading position and in silent reading) formats. Results were compared with those obtained while subjects were observing in silence a landscape picture on the median plane at 2 m. In addition, given the documented—albeit still controversial—influence of age and sex on SEBR, which some authors have associated with the higher prevalence of dry eye in the elderly population and in females,
22–26 the relationship between these predictor variables and blink parameters was investigated.
Seven different experimental configurations were tested (baseline and six reading conditions).
Table 1 presents a summary of the main characteristics of each experimental setting. During baseline (3 minutes), subjects were instructed to observe in silence a high-contrast landscape picture pasted on the wall at 2 m and eye level. All the other experimental settings (6 minutes) required subjects to read a text in various conditions, either in hard-copy or electronic format, in silence or aloud.
Table 1 Summary of the Experimental Configurations for Baseline and Reading Conditions
Table 1 Summary of the Experimental Configurations for Baseline and Reading Conditions
A collection of short easy reading stories by a famous Catalan author (Quim Monzó) was used as reading material and all texts were presented in the same typeface (Arial), font size (9), line spacing (1.15), and approximate number of words per page. The same reading stories were presented either in Catalan or Spanish according to the mother tongue of the participants. Electronic reading took place on a panoramic 24-inch, 16:9 liquid crystal display (TFT-LCD) set to a resolution of 1920 × 1080 pixels, 32-bit color configuration, contrast ratio 700:1, and 75 Hz refresh rate, or on a 9.7-inch, 4:3 display tablet (Energy i10 Quad SuperHD, Energy Sistem Soyntec S.A., Alicante, Spain) at a resolution of 2048 × 1536 pixels. The same display was employed to present the text at 330% magnification. This magnification value was set by adjusting the zoom level slider of the word processor software until the lines of the text fitted completely the whole width of the display. It must be noted that whereas hard-copy A4 size (297 × 210 mm) is very similar to the electronic page size when displayed at 100% scale (PC100), the actual screen size of the tablet is slightly smaller (239 × 179 mm).
The level of luminance emitted by each display (computer and tablet) was measured with a light meter (Gossen Mavolux 5032; Gossen Foto- und Lichtmesstechnik GmbH, Nürnberg, Germany) with the luminance attachment and adjusted to allow comparison among themselves and with the hard-copy text format (
Table 1). Small differences in luminance were allowed to guarantee correct visualization of the text.
Room temperature and humidity were maintained at 20°C (±2°C) and 40% (±10%), respectively, by adjusting and as displayed in the air conditioning settings. Background illumination was between 750 and 800 lx, and provided by diffuse lighting to avoid unwanted screen reflections.
Following a complete visual and ocular examination according to the inclusion and exclusion criteria, each participant completed the sequence of experimental conditions in a different random order to account for the potential effect of fatigue on the results. Block randomization was employed to assign a different order of experimental conditions to each participant. Baseline and reading sessions took place in the same day between 10 AM and 2 PM and all measurements were completed in approximately 40 minutes.
Subjects were instructed to scroll down using the middle wheel of a mouse or to flip pages by lightly tapping the edge of the screen (tablet) or by physically turning the page (hard-copy text). Subjects marked the last word they were able to read in each session and continued reading from the same word in the following session. Subjects were allowed time to familiarize themselves with the corresponding reading device before each reading session. All participants were instructed, if necessary, to use their spectacles instead of their contact lenses on the day of the study.
The aim of the present study was to assess several blink-related parameters while participants read texts in hard-copy and electronic format under controlled conditions, compared with a baseline situation of silent observation of a target at 2 m. Previous efforts have been directed at investigating differences in SEBR and percentage of incomplete eye blinks between texts presented in hard-copy format and on various types of displays,
4,8,9,13–18 although, to the best of our knowledge, these studies were in general limited to a single device.
Spontaneous blink rate in baseline conditions (median of 15.5 blinks per minute; interquartile range of 16 blinks per minute) was similar or slightly higher than the results reported in previous studies. For example, Doughty
32 calculated an average of 14.5 ± 3.3 blinks per minute in primary gaze, based on the findings of 22 previous studies. It must be noted that, in contrast with other authors, our criteria did not exclude participants with SEBR higher than 21 blinks per minute in baseline conditions, previously labeled as “frequent blinkers.”
33,34 This might have resulted in a higher median SEBR and in a larger intersubject variability than previously described, as well as in an increased probability of type II error. In addition, measurements in baseline conditions were restricted to 2 minutes, instead of the recommended 5 minutes,
9 as it was observed that some of the participants failed to remain interested in the fixation target after 3 minutes. This may be acknowledged as a limitation of the present research.
In agreement with other studies, all reading conditions, both in hard-copy and electronic format, led to a similar reduction in SEBR when compared with baseline conditions.
17,35 These findings may be explained by the attentional and cognitive demands associated with the reading task, which have been found to influence the central “pacemaker” mechanism governing SEBR. Tablet and computer at 100% zoom level resulted in a similar compromise in SEBR, with median values of 6 and 6.5 blinks per minute, respectively. Interestingly, however, this reduction was not as manifest when participants read text presented on an expanded display at 330%. The inclusion of an experimental condition requiring participants to read text in expanded format aimed at testing the hypothesis that SEBR would improve in a situation involving an increase in the percentage of large amplitude saccades (larger than 33°), many of which have been found to be associated with blinks (to reinforce visual suppression).
21 Nevertheless, it may be noted that, as far as we know, the association of blinking and large amplitude saccades has not been previously documented in the context of large format reading material or platforms.
Previous researchers have reported that the ocular discomfort experienced by computer users may be associated with an increase in the percentage of incomplete blinks, rather than with an actual reduction in SEBR.
17,18 Therefore, the advantage offered by the expanded display in terms of SEBR may not result in an actual reduction in ocular discomfort, particularly when considering the percentage of incomplete blinks observed with this reading condition. In effect, in contrast with hard-copy reading, reading in electronic format led to an increase in the percentage of incomplete blinks. In addition, this difference was particularly significant when comparing the expanded 330% display (13.5% of incomplete blinks) and the tablet (14.5% of incomplete blinks) with all hard-copy texts (from 0 to 5% of incomplete blinks), as well as with electronic reading at 100% (9% of incomplete blinks). These findings are in agreement with a recent investigation,
17 although the difference in incomplete blinks between electronic and hard-copy texts has not been explored in detail in the literature. Nevertheless, it may be argued that the contribution of incomplete blinking may be critical to explain the dry eye symptoms reported by VDT users and not by readers in traditional formats,
8,18,36 although it may be noted that one of our participants exhibited 100% incomplete blinks in both tablet and reading aloud conditions. The specific influence of VDT on incomplete blinks remains unexplained and may not only be attributed to differences in the actual position of the reading source. Indeed, incomplete blinks were found when participants read on a VDT display at 100%, but were less frequent when the same participants read a hard-copy text pasted on the switched off display under approximately the same, previously defined conditions (illumination, luminance, distance, font type and size, etc.).
Regarding age, OSDI scores and SEBR our findings were inconsistent. Thus, although a weak statistically significant correlation was found between SEBR and age (
ρ = 0.310;
P = 0.021), with elder participants presenting larger SEBR scores, there was a lack of correlation between OSDI scores and age (
ρ = 0.079;
P = 0.585). The influence of age in SEBR has been described in the literature,
22,23 and has been attributed to the associated increase in dry eye in the elderly population. The age distribution of the present study sample was skewed toward younger participants without dry eye symptoms (OSDI < 15). Therefore, even though it may be speculated that with a wider age range, a stronger correlation between SEBR and age might have been found, further research is needed to clarify this issue.
On the other hand, females were found to blink more frequently than men in all but one of the experimental conditions, and to report larger OSDI scores (median values of 12.7 for females and 8.3 for males, although
P > 0.05). Controversial evidence exists in the literature about blink parameter differences between males and females,
24–26 with some authors describing a possible link between SEBR sex differences and hormone-related factors such as use of estroprogestinics or phase of the ovarian cycle, both of which may affect blinking on the one hand and tear production on the other. The present findings give support to the existence of sex differences, although further research with two groups of OSDI-matched males and females is required to investigate whether these differences have an ocular surface or central origin.
It must be noted that we based our estimation of the required sample size on the meta-analysis conducted by Doughty,
9 in which reading resulted in a SEBR of 7.9 ± 3.3 blinks per minute. With an
α level of 0.05 and an 80% power to detect a difference between reading conditions of at least one standard deviation, a sample size of 16 was deemed necessary. In addition, recent work by Chu et al.
17 on blink amplitude in computer screen versus hard copy disclosed a change in the percentage of incomplete blinks from 7.02% (SD: 7.96%) to 4.33% (SD: 6.27%), respectively. The required sample size to replicate these findings (i.e., to detect changes in the percentage of incomplete blinks between different electronic and hard copy conditions) was found to be 80. Therefore, it may not be ruled out that, regarding blink amplitude, the present statistical analysis based on a sample of 50 participants was underpowered to detect differences.
In conclusion, the present findings—in which significant differences were disclosed in several blink parameters among electronic and hard-copy reading conditions—highlight the need to explore the percentage of incomplete blinks in addition to the commonly assessed SEBR. In effect, statistically significant differences were found between hard-copy and electronic reading in the percentage of incomplete blinks, a finding that warrants further research in this direction. In a constantly changing society in which handheld devices are not only ubiquitous but usually accessed under less than optimal viewing conditions (distance, font size, luminance), the evaluation of blink parameters in real-life situations and across platforms may be particularly relevant to assist in the design of strategies aimed at improving the ocular comfort and health of VDT users.
Genís Cardona and Elisabet Pérez-Cabré thank the Spanish Ministerio de Economía y Competitividad and Fondos FEDER for financial support (project number DPI2013-43220-R).
Disclosure: M. Argilés, None; G. Cardona, None; E. Pérez-Cabré, None; M. Rodríguez, None