The influence of downward gaze and accommodation on ocular aberrations over time

Abstract

A Shack–Hartmann wavefront sensor was modified to allow measurement of ocular aberrations in downward gaze with binocular fixation. Subjects first performed a control task prior to four different trials involving a distance task or a near task (accommodative demand of 2.5 D) performed in primary and downward gaze (25°). Immediately after beginning and then again 5 and 10 min after the commencement of each trial, ocular aberrations were measured. To observe the recovery in ocular aberrations following each test condition, subjects again viewed a distance target in primary gaze and aberration measurements were taken at 0, 5, and 10 min. During the distance task, small but significant changes in refractive power and higher order aberrations were observed in downward gaze compared to primary gaze. The changes in ocular aberrations that occurred in downward gaze recovered almost immediately after shifting gaze from downward back to primary gaze. During the accommodation tasks, there was a significant influence of gaze for changes in primary spherical aberration C(4, 0) [p = 0.004] and secondary spherical aberration C(6, 0) [p = 0.02]. There was also a significant gaze by time interaction (p = 0.04) for changes in C(6, 0). These findings show that ocular aberrations change from primary to downward gaze, particularly during accommodation.

Introduction

Many visual tasks such as reading and computer work involve accommodation and downward gaze. However, clinical wavefront sensors are typically designed to only measure the eye's wavefront aberrations in primary gaze while viewing a target with optical vergence at the far point of the eye or beyond. A number of studies (Mathur et al., 2009; Prado et al., 2009; Radhakrishnan & Charman, 2007b) have found minimal significant change in the optics of the eye with changes in horizontal gaze; however, during downward gaze, there is evidence of an increase in the total optical power of the eye (Ripple, 1952; Takeda, Neveu, & Stark, 1992). Downward gaze may yield different outcomes due to the differences in biomechanical forces (e.g., extraocular muscles or eyelid forces) and gravitational effects acting on the components of the eye. 

Accommodation is known to cause substantial changes in some aspects of ocular aberrations. These changes typically include a systematic decrease in spherical aberration with increasing accommodation (Atchison, Collins, Wildsoet, Christensen, & Waterworth, 1995; Cheng et al., 2004; He, Burns, & Marcos, 2000; Smirnov, 1962) and changes in astigmatism (Mutti, Enlow, & Mitchell, 2001; Radhakrishnan & Charman, 2007a; Tsukamoto et al., 2000) and third-order comatic aberrations (Atchison et al., 1995; Cheng et al., 2004). These changes in the optical characteristics of the eye with accommodation are most likely due to changes in the crystalline lens since the cornea shows little evidence of change with accommodation, when the cyclotorsion of the eye is taken into consideration (Buehren, Collins, Loughridge, Carney, & Iskander, 2003; Read, Buehren, & Collins, 2007). 

In downward gaze, the narrower vertical palpebral aperture induces subtle but distinctive changes in astigmatism and higher order aberrations (HOAs) of the eye (Buehren, Collins, & Carney, 2003; Collins, Buehren, Bece, & Voetz, 2006). These optical effects that appear to originate from corneal changes associated with eyelid forces in downward gaze have been evaluated immediately following short-term (Shaw, Collins, Davis, & Carney, 2008) and longer term (Buehren, Collins & Carney, 2003) downward gaze tasks but never during the downward gaze task. 

Therefore, in order to better understand the optics of the eye during the natural conditions of typical near tasks, we aimed to investigate changes in ocular aberrations associated with accommodation in downward gaze. We measured ocular aberrations in both primary and downward gaze (25 degrees), in binocular viewing conditions, with natural pupils at two levels of accommodation (0.2 D and 2.5 D) using a modified Shack–Hartmann aberrometer system over a period of time (10-min task followed by 10-min recovery). We chose to make the measurements both during and after the near, downward gaze task to investigate the time course of any changes that may occur in the optics of the eye and to understand the recovery characteristics of these changes. Again, we reasoned that many visual tasks involve sustained periods of viewing, and therefore, the appropriate methodology should involve periodic measurements of the eye's optics over time. 

Methods

Subjects

Twenty-six young adult subjects (12 emmetropes and 14 myopes) aged between 19 and 30 years (mean age of 25 years) were recruited for this study. All subjects were free of any significant ocular diseases and had no history of eye surgery. Any subject who habitually wore soft contact lenses (n = 6) was asked to abstain from lens wear for 2 days prior and during their involvement with the study. Approval was obtained from the University Human Research Ethics Committee prior to the commencement of the study. Subjects were treated in accordance with the Declaration of Helsinki. All subjects had best corrected visual acuity of logMAR 0.00 or better in both eyes. The mean spherical equivalent of the myopic subjects was −3.26 ± 1.60 DS (mean ± SD) and that of the emmetropic subjects was −0.29 ± 0.36 DS. Myopes who had evidence of at least 0.5-DS progression of their myopia within the past 2 years were recruited. Refractive error progression data were obtained from the subject's eye care practitioner if the details were not known by the subject. None of the subjects had anisometropia greater than 1.00 DS or astigmatism greater than 1.50 DC. 

Experimental design

A modified Shack–Hartmann wavefront sensor was used to measure ocular aberrations in primary gaze and downward gaze (25 degrees). Figure 1 illustrates the schematic diagram of this modified system. The working principle, measurement technique, and validation of this modified system are explained in detail elsewhere (Ghosh, Collins, Read, Davis, & Iskander, 2011). The system allowed binocular fixation on a free-space, high-contrast Maltese cross target with various levels of accommodation in different angles of gaze. Multiple measurements of wavefront aberrations (total of 4 measures × 25 frames = 100 frames in total) from the subject's left eye were taken during four different test conditions: (i) no accommodation (0.2 D) in primary gaze, (ii) no accommodation (0.2 D) in downward gaze, (iii) 2.5-D accommodation in primary gaze, and (iv) 2.5-D accommodation in downward gaze, using the modified wavefront sensor. The order of testing the conditions was randomized (Latin square design) to avoid the potential for systematic bias in the results. 

Figure 1

 

Schematic diagram of a modified Shack–Hartmann wavefront sensor. M1 is the first surface mirror that reflects measurement beam toward the relay lens. L1 and L2 are the identical achromatic doublets (100-mm focal length). M2 is a hot mirror that reflects measurement beam toward the subject's eye. Downward gaze was achieved by tilting mirror M2 and shifting the vertical location of the Maltese cross target.

Data collection procedures

Correction of refractive error during testing

Subjects viewed the fixation targets through their full distance refractive error correction (spherocylinder) during each of the four experimental testing conditions. The vertex distance of the trial lenses was considered to determine the appropriate refractive error correction. The refractive error correction was placed in the form of trial lenses mounted in the optical path to the fixation target (Maltese cross) and TV monitor but outside the measurement beam path of the wavefront sensor to avoid wavefront magnification issues (Figure 1). 

Wavefront measurement procedure

For each test condition, the subject was asked to focus on a Maltese cross target with binocular fixation, while wavefront measurements were taken from the left eye. The subject's head position was adjusted (vertically and horizontally) until the center of the Maltese cross target coincided with the center of the wavefront sensor fixation target. Therefore, the center of the wavefront sensor target and the center of the Maltese cross were coaxial with the visual axis of the subject's tested left eye (OS). This alignment ensured that we were measuring the on-axis aberrations of the subject's left eye. Then, the examiner was centering the wavefront sensor on the subject's pupil center in which case the axis measured was assumed to be the line of sight (Applegate, Thibos, Twa, & Sarver, 2009). The subject was asked to keep the Maltese cross target “as clear as possible” throughout the measurement protocol. The subject was asked to blink and then to view the fixation target, while 25 frames of wavefront measures were acquired with the modified wavefront sensor (approximately 2.5 s of recording). The subject was asked to blink again, and the measurement of 25 frames was repeated (total of 4 measures × 25 frames = 100 frames in total). The measurement was repeated if the subject blinked or reported “losing focus”. The subject was instructed to keep their eyelids in the natural position during wavefront measurements (i.e., “not to open their eyes wide”). This was done to ensure that the natural influence of the eyelids on the eye was maintained during the wavefront measurements. The intention of this measurement procedure was to capture the optical characteristics of the eye in their natural state during primary and downward gaze, both with and without accommodation. To minimize the potential influence of diurnal changes in corneal optics (Read & Collins, 2009), we ensured that subjects had woken at least 2 hours before the measurement session. 

In order to standardize the ocular parameters prior to the measurements in each experimental condition, a control task involving constant binocular viewing of a distant target (watching TV at 5 m) in primary gaze for 10 min was performed by each subject. At the end of the 10-min control task, ocular aberrations of the left eye were measured in primary gaze with binocular distance fixation using the modified wavefront sensor. During wavefront measurements, the fixation was controlled by pausing the DVD being viewed on the TV and asking the subject to fixate on a high-contrast Maltese cross target displayed on the TV screen. The four different experimental conditions (accommodation and gaze combinations) are listed below. 

No accommodation (0.2 D) in primary gaze condition: In this condition, the subject watched TV binocularly at a 5-m distance in primary gaze for 10-min duration. After 0 min (i.e., immediately after commencing the task), 5 min, and 10 min from the starting time (watching TV), we took the measurements of ocular aberrations using the modified wavefront sensor with binocular vision and with fixation on the free space target (Maltese cross at 5 m). To observe the recovery of the optical characteristics of the eye following each test condition, the subject again looked in primary gaze and watched TV at 5 m (control task) for a duration of 10 min. Measurements were carried out over this recovery period at time 0 min (i.e., immediately after commencing the recovery task of primary gaze and no accommodation), 5 min, and 10 min after commencing the recovery period. The subject remained in the headrest throughout each 30-min testing session that included the control task (10 min), test condition (10 min), and recovery condition (10 min). For this particular condition, the control task, the test condition, and the recovery period all involved the same visual conditions (i.e., primary gaze, with no accommodation viewing a TV at 5 m). 

No accommodation (0.2 D) in downward gaze condition: In this test session, subjects again performed the control task (10-min distance viewing of TV in primary gaze) before beginning the next test condition. During this condition, ocular aberrations were measured for downward gaze (25 degrees), while the subject watched TV in downward gaze at a 5-m distance via a mirror (Ghosh et al., 2011). The same 30-min measurement protocol was followed during the control, test, and recovery periods. 

2.5-D accommodation in primary gaze: In this condition, the subject's task involved near work with a 2.5-D accommodation stimulus level (watching a movie on the screen of a pocket-size device) in primary gaze. In all of the accommodation conditions, the target distance was adjusted so that the accommodative stimulus was constant for all subjects. This adjustment was based on the spectacle lens effectivity power for each subject (Mutti, Jones, Moeschberger, & Zadnik, 2000). 

2.5-D accommodation in downward gaze: In this condition, the subject watched a movie on the screen of a pocket-size device in 25-degree downward gaze at a 2.5-D accommodation level. An identical protocol was followed as the previous condition (2.5-D accommodation in primary gaze) except that the measurements of ocular aberrations using the modified wavefront sensor were taken at 25-degree downward gaze rather than primary gaze. 

A break of 10 min after the completion of each testing condition was provided to all subjects, and they were instructed not to perform any sustained near or downward gaze tasks during this time. 

Stimulus conditions

The TV monitor used at a 5-m viewing distance was a high-definition plasma screen with horizontal and vertical dimensions of 92 cm and 55 cm (visual angle of 11 × 6 degrees). The screen luminance was approximately 20 cd/m². The monitor of a pocket-size device used for near viewing distances was a liquid crystal display (LCD) screen with horizontal and vertical dimensions of 5.4 cm and 7.5 cm (visual angle of 11 × 8 degrees), similar to the visual angle subtended by the TV monitor for far viewing. The screen luminance for the LCD was also approximately 20 cd/m². The room illuminance for all testing conditions was set to mesopic levels of 1 lux to maximize the pupil size during the wavefront measurements. 

Data analysis

The presence of the relay lens system in our modified wavefront sensor inverted the wavefront in primary gaze and there was an additional rotation of the wavefront due to the hot mirror rotation in downward gaze. Therefore, we had to rotate the wavefront (Lundstrom & Unsbo, 2007) to its original position using custom written software. In downward gaze, the eye rotation is also associated with cyclotorsion, which will introduce an artifact in the changes of wavefront aberrations. Therefore, in order to correct for cyclotorsion in downward gaze, the iris images from the wavefront sensor in primary and downward gaze were analyzed for each subject using custom written software (Ghosh et al., 2011). 

Orthogonal refractive components (M, J0, and J45) were determined from the wavefront aberrations (Iskander, Davis, Collins, & Franklin, 2007) using custom written software (Matlab-based) along with conventional Zernike polynomials up to the 8th radial order using the nomenclature recommended by the Optical Society of America (American National Standards Institute (ANSI), 2004). For each frame of the 100 wavefront measurements, a Matlab-based algorithm was used to detect artifacts in the Zernike polynomial fit coefficients and then to filter these data before averaging (Iskander, Collins, Morelande, & Zhu, 2004). Sources of these artifacts could include blinking, poor stability of the subject's head, and poor tear film quality. 

The wavefront was fitted with Zernike polynomials for a fixed 5.0-mm pupil diameter for the no accommodation conditions (primary and downward gaze) and for a 3.0-mm pupil diameter for the 2.5-D accommodation conditions (primary and downward gaze). The 5.0-mm and 3.0-mm fixed entrance pupils were selected because they were smaller than the minimum diameter of natural pupil sizes for all subjects during the no accommodation condition and 2.5-D accommodation conditions, respectively. Note that the no accommodation condition corresponded to the 5-m distance, so the actual level of accommodation stimulus during the distance task was 0.2 D. Therefore, the difference between the accommodation (2.5 D) and no accommodation (0.2 D) conditions was 2.3 D. 

Statistical analyses were performed using SPSS (version 17.0) software. A repeated measures ANOVA was performed to assess the significance of wavefront changes in the various conditions (within-subjects factors) including primary gaze versus downward gaze and the effects of measurement time within the task and recovery periods. The between-subjects factor was refractive error group (myopes and emmetropes). 

Results

Primary gaze vs. downward gaze without accommodation

Analysis of the ocular wavefront showed a number of significant changes in the refractive components and the Zernike coefficients in downward gaze with no accommodation, compared to the primary gaze with no accommodation condition. The group mean changes of refractive components from baseline (for 5.0-mm pupil diameter) in primary gaze and downward gaze over 10-min duration are presented in Figure 2 and Table 1. The mean spherical equivalent (M) showed a small (−0.10 D at 0-min duration, −0.12 D at 5-min duration, and −0.11 D at 10-min duration) but significant myopic shift in downward gaze from baseline (repeated measures ANOVA, p = 0.001). Primary astigmatism (J0) shifted in the against-the-rule (ATR) direction in downward gaze (repeated measures ANOVA, p = 0.001). There were also significant gaze by time (p = 0.04), gaze by refractive error (p = 0.012), and gaze by time by refractive error (p = 0.029) interactions for the changes in J0. The myopic group showed a greater ATR astigmatic shift (mean change of 0.10 ± 0.005 D from baseline) compared with the emmetropic group (mean change of 0.01 ± 0.007 D from baseline) after the 10-min task in downward gaze (pairwise comparison, p = 0.003). It is apparent that the refractive components (M, J0) that changed significantly in downward gaze recovered almost immediately after shifting viewing direction from downward gaze back to primary gaze (Figure 2). 

Figure 2

 

The group mean changes (±SE) of refractive components in primary gaze and in downward gaze with respect to baseline with the no accommodation condition over the 10-min task for a fixed 5.0-mm pupil diameter. The baseline value was taken after 10 min of viewing a 5-m target [i.e., no accommodation (0.2 D)] in primary gaze. Values at time 0 min of the task and recovery were taken immediately after the task/recovery periods commenced. In this plot, M and J0 show a significant difference between primary gaze and downward gaze during the no accommodation task (ANOVA, p < 0.05). J0 shows a significant gaze by time interaction (ANOVA, p < 0.05).

Table 1

 

Group mean (±SE) changes of refractive components from baseline in primary gaze and downward gaze in the no accommodation condition over 10-min duration for 5.0-mm fixed pupil diameter.

Refractive components	Mean (±SE) changes of refractive error from baseline (D)				P values
		0 min	5 min	10 min	Gaze	Gaze * time	Gaze * refractive error	Gaze * time * refractive error
M	Primary gaze	0.026 ± 0.012	0.001 ± 0.022	0.024 ± 0.024	0.001	0.610	0.652	0.155
	Downward gaze	0.099 ± 0.032	−0.144 ± 0.032	−0.110 ± 0.024
J0	Primary gaze	0.006 ± 0.010	−0.002 ± 0.008	−0.010 ± 0.009	0.001	0.040	0.012	0.029
	Downward gaze	0.031 ± 0.013	0.072 ± 0.020	0.061 ± 0.018†
J45	Primary gaze	0.005 ± 0.005	0.006 ± 0.008	0.006 ± 0.008	0.769	0.926	0.146	0.150
	Downward gaze	0.013 ± 0.012	−0.002 ± 0.010	0.002 ± 0.013

 

Notes: Bold numbers indicate the level of significance (p < 0.05) revealed by repeated measures ANOVA.

 

^†Indicates a statistically significant variation between refractive error groups (Bonferroni-corrected pairwise comparison, p < 0.05).

Of the HOAs, the terms that changed significantly in downward gaze compared to primary gaze were vertical trefoil C(3, −3) [p < 0.001], vertical coma C(3, −1) [p = 0.031], secondary astigmatism C(4, 2) [p = 0.012], tetrafoil C(4, 4) [p = 0.009], secondary coma C(5, −1) [p = 0.002], pentafoil C(5, −5) [p = 0.003], and secondary spherical aberration C(6, 0) [p = 0.001] (Figure 3). Primary spherical aberration C(4, 0) shifted in the positive direction in downward gaze (mean difference = 0.004 ± 0.005 μm at 0 min, mean difference = 0.010 ± 0.007 μm at 5 min, and mean difference = 0.007 ± 0.004 μm at 10 min). The interaction with angle of gaze approached significance for the changes in C(4, 0) for the no accommodation condition (p = 0.08). 

Figure 3

 

The group mean changes (±SE) of higher order coefficients (HOAs) in primary gaze and in downward gaze with respect to baseline with the no accommodation condition over the 10-min task for fixed 5.0-mm pupil diameter. The baseline value was taken after 10 min of viewing a 5-m target [i.e., no accommodation (0.2 D)] in primary gaze. Values at time 0 min of the task and recovery were taken immediately after the task/recovery periods commenced. Each of the plotted coefficients shows a significant difference between primary gaze and downward gaze during the no accommodation task (ANOVA, p < 0.05). Coefficients C(3, −3), C(3, −1), C(4, −2), and C(4, 4) also show a significant gaze by time interaction (ANOVA, p < 0.05). None of the remaining coefficients up to 8th order show significant changes with repeated measures ANOVA (p > 0.05).

There was a trend for the vertical coma C(3, −1) to shift in the positive direction, whereas the vertical trefoil C(3, −3) shifted in the negative direction (Figure 3). We found a significant correlation (Pearson's R ² = 0.701, p < 0.001) between the changes of these two coefficients [C(3, −1) and C(3, −3)] in downward gaze. There was also a gaze by time interaction in the changes of higher order coefficients such as C(3, −3) [p = 0.048], C(3, −1) [p = 0.044], C(4, −2) [p = 0.032], and C(4, 4) [p = 0.044]. 

Changes in secondary astigmatism along 45° [C(4, −2)] were significantly greater in myopes than emmetropes after 5 min (mean difference = −0.014 ± 0.007 μm; pairwise comparison, p = 0.03) and 10 min (mean difference = −0.016 ± 0.07; pairwise comparison, p = 0.02) of the task (no accommodation) in downward gaze compared to the baseline measurement. Similar to the refractive components, the changes in HOAs during downward gaze also recovered immediately after shifting the viewing gaze from downward back to primary. 

The group mean refractive power maps were determined over a 5.0-mm pupil diameter for the baseline condition and then 10 min after the 25-degree downward gaze task (Figure 4). The difference between the group mean refractive power maps at baseline and after 10 min of downward gaze shows a band-like horizontal distortion just above the central zone of the pupil and parallel to the typical position of the upper eyelid (Figure 4). It also shows a small comatic change at the center of the map that was parallel to the visual axis. 

Figure 4

 

(A) Group mean baseline refractive power map and (B) group mean refractive power map after 10-min task in downward gaze for 5.0-mm pupil diameter. No accommodation in both conditions was represented in (A) and (B). (Bottom center) Refractive power difference map of post-task in downward gaze minus baseline refractive power (B − A) for 5.0-mm pupil diameter.

Primary gaze vs. downward gaze with accommodation

Mean refractive power (M) shifted in the myopic direction, and primary astigmatism (J0) shifted in the direction of more with-the-rule (WTR) astigmatism (Figure 5) with the 2.5-D accommodation stimulus for both the primary and downward gaze compared with the control condition (no accommodation and primary gaze) for a 3.0-mm fixed pupil (both p < 0.001). After 10 min of the near task with the 2.5-D accommodation stimulus, mean refractive power (M) had a slightly greater myopic shift in downward gaze compared to primary gaze (mean difference = −0.05 D; Figure 5). However, unlike the no accommodation condition, there were no significant differences in the changes of M and J0 with accommodation between primary and downward gaze (p > 0.05) and there were no significant interactions between the changes in M, J0, and J45 components and refractive error group or the gaze by time factor. 

Figure 5

 

The group mean changes (±SE) of (A) mean refractive power (M) and (B) primary astigmatism (J0) in primary gaze and in downward gaze with respect to baseline with 2.5-D accommodation condition over the 10-min task for fixed 3.0-mm pupil diameter. The baseline value was taken after 10 min of viewing a 5-m target [i.e., no accommodation (0.2 D)] in primary gaze. The effective accommodation stimulus was (2.5 D − 0.2 D) = 2.3 D. Values at time 0 min of the task and recovery were taken immediately after the task/recovery periods commenced. This plot shows no significant difference between primary gaze and downward gaze during 2.3-D accommodation task (ANOVA, p > 0.05).

Considering the individual Zernike coefficients of the HOAs, primary spherical aberration C(4, 0) changed significantly with accommodation (p = 0.005). The interaction with angle of gaze approached significance for the changes in horizontal coma C(3, 1) with accommodation (p = 0.09) and there was a significant gaze by time interaction for the change in horizontal coma C(3, 1) (p = 0.04; Figure 6A). 

Figure 6

 

The group mean changes (±SE) of (A) vertical coma C(3, −1) and horizontal coma C(3, 1) and (B) primary spherical aberration C(4, 0) and secondary spherical aberration C(6, 0) in primary gaze and in downward gaze with respect to baseline with 2.5-D accommodation condition over the 10-min task for fixed 3.0-mm pupil diameter. The baseline value was taken after 10 min of viewing a 5-m target [i.e. no accommodation (0.2 D)] in primary gaze. The effective accommodation stimulus was (2.5 D − 0.2 D) = 2.3 D. Values at time 0 min of the task and recovery were taken immediately after the task/recovery periods commenced. In this plot, C(4, 0) shows a significant difference between primary gaze and downward gaze during 2.3-D accommodation task (ANOVA, p < 0.05). C(3, 1) shows a significant gaze by time interaction (ANOVA, p < 0.05).

Primary spherical aberration C(4, 0) was found to become more negative with accommodation in downward gaze compared to primary gaze accommodation (p = 0.004; Figure 6B). Secondary spherical aberration C(6, 0) became more positive with accommodation in downward gaze compared to primary gaze accommodation (p = 0.02; Figure 6B). For both primary and secondary spherical aberrations, the difference in magnitude of the coefficients between primary and downward gaze became statistically significant after 10 min of the near task [C(4, 0) pairwise comparison, p = 0.004; C(6, 0) pairwise comparison, p = 0.04] but not at 0 or 5 min after commencing the near task (pairwise comparisons, p > 0.05 for both times and coefficients). 

The level of change in secondary spherical aberration between primary gaze and downward gaze systematically increased over the 10-min time course of the near task [p = 0.04 (gaze by time)]. Moderate but significant negative correlations were found between changes in primary and secondary spherical aberrations for both primary gaze (R ² = 0.55, p < 0.001 at 0 min; R ² = 0.60, p = 0.004 at 5 min; and R ² = 0.30, p < 0.001 at 10 min of task) and downward gaze (R ² = 0.29, p = 0.006 at 0 min; R ² = 0.52, p < 0.001 at 5 min; and R ² = 0.54, p < 0.001 at 10 min of task). 

Group mean refractive power analysis (Figure 7) showed that the changes in primary and secondary spherical aberrations resulted in more negative power in the periphery of the pupil with downward gaze compared to primary gaze, and this effect increased with time during the 2.5-D accommodation task. This indicates a gradual negative shift in longitudinal spherical aberration (LSA) over time in downward gaze (Figure 7). There were no significant interactions between gaze and refractive group for the changes in higher order aberrations with accommodation. Refractive power is observed to return to baseline levels immediately following cessation of the tasks. 

Figure 7

 

Group mean refractive power maps illustrating changes in spherical aberration [primary SA C(4, 0) + secondary SA C(6, 0)] with respect to baseline during accommodation (2.5-D stimulus) in primary gaze and downward gaze over the 10-min course of the near task for a 3.0-mm fixed pupil. Longitudinal spherical aberration (LSA) is given in terms of the dioptric difference between the marginal ray focus and the paraxial ray focus. Defocus and astigmatism were excluded for the refractive power analysis.

Discussion

Changes in wavefront aberrations with no accommodation

We found small but significant changes in the optics of the eye in downward gaze compared to primary gaze, when both gaze conditions involved distance viewing (i.e., minimal accommodation stimulus). Downward gaze led to a shift in astigmatism in the direction of more ATR astigmatism [positive shift in primary astigmatism (J0)] and this change in optics during downward gaze is consistent with the results of previous studies (Buehren, Collins, & Carney, 2003; Collins et al., 2006; Shaw et al., 2008) that have shown substantial changes in the topographical and optical characteristics of the cornea associated with eyelid pressure on the corneal surface following reading in downward gaze. We found that the degree of these changes in astigmatism were greater as a function of time and it has also previously been shown that the magnitude of lid-induced corneal change increases with the length of time spent reading (Collins, Kloevekorn-Norgall, Buehren, Voetz, & Lingelbach, 2005). 

The mean refractive power of the eye (M) shifted slightly in the myopic direction in downward gaze in the no accommodation condition, consistent with previous findings (Ripple, 1952; Takeda et al., 1992). While the change in mean spherical refractive power was small (about −0.12 D) in magnitude, it was consistent and statistically significant. A study with monkeys observed a downward sag of the crystalline lens during centrally stimulated accommodation, presumably due to the action of gravity (Glasser & Kaufman, 1999). Another recent study reported that anterior chamber depth in human eyes may alter under the influence of gravity due to lens movement (Kasthurirangan, Markwell, Atchison, & Pope, 2011). Therefore, the myopic shift in refractive power in downward gaze that we observed may be the result of a small forward movement of the crystalline lens under the action of gravity. From previously reported modeling, we expect that the crystalline lens would be required to move about 100 μm anteriorly to increase the eye's power by −0.12 D in downward gaze (Atchison, 2005; Collins, Davis, & Wood, 1995). In contrast, Shaw et al. (2008) found a small hyperopic shift in corneal refractive power after a short period of reading in 20-degree downward gaze. A possible reason for the discrepancy between changes in corneal power and total eye power could be that the pattern of change in optics of the cornea and crystalline lens is different in downward gaze. Earlier studies have confirmed that the magnitudes and signs of the corneal aberrations are different from the aberrations of the internal optics of the eye and there is a degree of compensation of optical aberrations between the corneal and internal optics that reduces the total amount of aberrations in younger eyes (Artal, Guirao, Berrio, & Williams, 2001; Tabernero, Benito, Alcon, & Artal, 2007). It would, therefore, be useful to simultaneously measure the changes in corneal and total optics in downward gaze to determine the relative contributions of the corneal and internal optics to the total optical changes occurring in downward gaze. 

There were numerous changes in ocular HOAs in downward gaze in terms of both magnitude and sign of the terms. The changes in vertical trefoil and vertical coma were consistent with a previous report of changes in the total wavefront following reading (Buehren, Collins, & Carney, 2003). The group mean refractive difference map (Figure 4) was also similar to the corneal refractive power changes observed by Shaw et al. (2008) who found a horizontal band-like distortion in corneal topography parallel to the upper eyelid, and the similar changes that we have observed are also likely to be associated with biomechanical forces due to eyelid pressure on the corneal surface during downward gaze. 

Other than coma and trefoil, there were also significant changes in secondary astigmatism, tetrafoil, secondary coma, pentafoil, and secondary spherical aberration in downward gaze. As we observed greater changes in refractive components and higher order aberrations of the eye during downward gaze than the previous studies of corneal or ocular optics following downward gaze (Buehren, Collins, & Carney, 2003; Shaw et al., 2008), these changes may be related to the changes in the internal optics of the eye such as crystalline lens tilt or changes of the crystalline lens shape during downward gaze due to the variation in zonular tension or extraocular muscle force on the eye during downward gaze. 

We observed an immediate recovery of wavefront changes of the eye (to baseline values) following the 10-min visual task in downward gaze. A previous study (Buehren, Collins, & Carney, 2003) found that a recovery period of 120 min was required for the wavefront to approach pre-reading corneal shape after a 60-min reading task in downward gaze. Because the time period of the visual task in the present study extended only 10 min, the changes recovered rapidly. However, it is possible that longer duration tasks may exacerbate the magnitude of certain wavefront changes that occur in downward gaze, and the recovery period may be longer, depending on the underlying causes of the wavefront changes. 

Changes in wavefront aberrations with accommodation

Spherical aberration (SA) changed consistently in the negative direction with accommodation in both primary and downward gaze. This result is in agreement with previous findings that show a systematic shift of spherical aberration from positive to negative sign with accommodation (Atchison et al., 1995; Cheng et al., 2004; Collins, Wildsoet, & Atchison, 1995; Hazel, Cox, & Strang, 2003; He et al., 2000; Ivanoff, 1956; Koomen, Tousey, & Scolnik, 1949; Ninomiya et al., 2002). It has also been mathematically predicted that the primary spherical aberration will shift in the negative direction and that the secondary spherical aberration will shift in the positive direction during accommodation due to the hyperbolic shape changes of the crystalline lens (Lopez-Gil & Fernandez-Sanchez, 2010). Our data show a similar pattern to these predictions. During accommodation, changes occur in the shape (particularly the anterior surface), thickness, and refractive index of the crystalline lens (Brown, 1973; Dubbelman, Van der Heijde, & Weeber, 2005; Jones, Atchison, & Pope, 2007; Kasthurirangan, Markwell, Atchison, & Pope, 2008). The changes of spherical aberration associated with accommodation are probably related to changes in the asphericity or refractive index distribution of the crystalline lens as it alters with accommodation. 

We found that spherical aberration varied significantly over time with accommodation, being more stable during accommodation in primary gaze but showing a greater change in downward gaze over time (Figure 7). The reason for this difference between primary and downward gaze could be that gravitational effects during downward gaze combined with accommodation may exacerbate the normal changes in shape and refractive index distribution of the crystalline lens. The greater change noticed in the mean refractive power (M) of the eye with no accommodation in downward gaze compared with primary gaze adds some support for this hypothesis. We can only speculate on the possible reasons for the variations in spherical aberration and total power of the eye over time in downward gaze, but it seems conceivable that the gravitational force may influence the interaction between the pupil margin and the lens surface with accommodation in downward gaze if the lens moves slightly anteriorly. It is known that interactions between the iris and lens surface contribute to substantial changes in lens power during accommodation in some bird species (Levy & Sivak, 1980), and it has recently been hypothesized that interactions between the iris and lens surface underlie subtle changes in ocular aberrations associated with pupil constriction (Montes-Mico et al., 2010). Therefore, gravitational forces during downward gaze may increase the pressure applied by the pupil margin to the anterior lens surface with accommodation that could alter the shape of the crystalline lens (in a rotationally symmetrical pattern) and thereby influence primary and secondary spherical aberrations. This hypothesis may have a link with the other observation of a gradual recovery to baseline in spherical aberration over time in primary gaze but not in downward gaze (Figure 7). It is possible that the deformed crystalline lens surface could take some time to regain the original shape after a sustained near task in downward gaze. Further research examining the biometric changes of the eye during downward gaze is required to better understand the origins of these changes. 

We found that third-order horizontal coma changed in the negative direction with accommodation in primary and downward gaze (Figure 6A). Rosales, Wendt, Marcos, and Glasser (2008) found a significant tilt of the crystalline lens of rhesus monkeys around the horizontal axis during accommodation. Therefore, it seems feasible that accommodation may also cause crystalline lens tilt in human eyes, with resulting changes in astigmatism and coma during accommodation. It is also apparent from Figure 6A that the changes in vertical coma and horizontal coma with accommodation are greater in downward gaze compared to primary gaze after 10 min of the near task. A study by Glasser and Kaufman (1999) found that gravity may influence the movement of the crystalline lens during accommodation in monkeys. It is therefore possible that gravitational effects on the tilt and centration of the human crystalline lens could also lead to a greater change in coma in downward gaze during accommodation. 

Analysis of the effects of refractive error group showed some small but significant differences in the changes occurring in primary and secondary astigmatism. However, there were no obvious systematic differences in HOA between refractive error groups during accommodation or downward gaze or as a function of time. Although this is true for fixed pupil diameters (5 mm and 3 mm), it could still be argued that there might be a potential variation in the level of monochromatic aberrations between refractive error groups during a near task for natural pupil diameters. We chose to conduct analyses in this study with fixed pupil diameters to allow direct comparison of the wavefront aberration components between the conditions (gaze angle, accommodation, and time), since pupil size is likely to change with both accommodation and time. However, it could be argued that there is further information to be gained by also comparing between refractive error groups with natural pupil diameters, since this reflects the true state of the optics of the eye during near tasks. 

Conclusion

We have highlighted that significant changes occur in the optical properties of the eye during a near task under the influence of both gaze angle and accommodation. The spherical refractive power of the eye shifts slightly in the myopic direction in downward gaze with no accommodation, while astigmatism shifts in the direction of ATR. Higher order aberrations of the eye also change significantly in downward gaze over time. There are several factors that are likely to contribute to these changes including eyelid pressure on the cornea and gravitational effects on the crystalline lens. Accommodation also causes significant changes in astigmatism and higher order aberrations [in particular, spherical aberration (SA) and coma], which probably occur due to the changes in the crystalline lens. Changes over time in spherical aberration and coma with accommodation appear to be significantly greater in downward gaze compared to primary gaze. These variations in ocular aberrations over time highlight the dynamic nature of visual optics during a near task. 

Acknowledgments

The authors thank Payel Chatterjee for assistance with data collection and analysis. Aspects of this work have been presented at the Association for Research in Vision and Ophthalmology (ARVO) Meeting in 2011. 

Commercial relationships: none. 

Corresponding author: Atanu Ghosh. 

Email: [e-mail]. 

Address: School of Optometry, Queensland University of Technology, Room No. B557, B Wing, Level 5, O Block, Victoria Park Road, Kelvin Grove, Brisbane, Australia.