- Email: [email protected]
Tonic accommodation predicts closed-loop accommodation responses
Vision Research 129 (2016) 25–32
Contents lists available at ScienceDirect
Vision Research journal homepage: www.elsevier.com/locate/visres
Tonic accommodation predicts closed-loop accommodation responses Chunming Liu a,⇑, Stefanie A. Drew a,b, Eric Borsting c, Amy Escobar a,1, Lawrence Stark c, Christopher Chase a,⇑ a
Western University of Health Sciences, College of Optometry, Pomona, CA, United States California State University Northridge, Northridge, CA, United States c Southern California College of Optometry at Marshall B. Ketchum University, Fullerton, CA, United States b
a r t i c l e
i n f o
Article history: Received 14 July 2015 Received in revised form 24 August 2016 Accepted 26 August 2016 Available online 1 November 2016 Keywords: Tonic accommodation Accommodation response TA-adaptation Vergence
a b s t r a c t The purpose of this study is to examine the potential relationship between tonic accommodation (TA), near work induced TA-adaptation and the steady state closed-loop accommodation response (AR). Forty-two graduate students participated in the study. Various aspects of their accommodation system were objectively measured using an open-field infrared auto-refractor (Grand Seiko WAM-5500). Tonic accommodation was assessed in a completely dark environment. The association between TA and closed-loop AR was assessed using linear regression correlations and t-test comparisons. Initial mean baseline TA was 1.84 diopter (D) (SD ± 1.29 D) with a wide distribution range (0.43 D to 5.14 D). For monocular visual tasks, baseline TA was significantly correlated with the closed-loop AR. The slope of the best fit line indicated that closed-loop AR varied by approximately 0.3 D for every 1 D change in TA. This ratio was consistent across a variety of viewing distances and different near work tasks, including both static targets and continuous reading. Binocular reading conditions weakened the correlation between baseline TA and AR, although results remained statistically significant. The 10 min near reading task with a 3 D demand did not reveal significant near work induced TA-adaptation for either monocular or binocular conditions. Consistently, the TA-adaptation did not show any correlation with AR during reading. This study found a strong association between open-loop TA and closed-loop AR across a variety of viewing distances and different near work tasks. Difference between the correlations under monocular and binocular reading condition suggests a potential role for vergence compensation during binocular closed-loop AR. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Under natural viewing conditions, the accommodative system of the human eye mainly employs a retinal blur signal to drive an accurate accommodative response (AR) using a closed-loop negative feedback system under monocular condition (Toates, 1972). In the absence of an adequate visual stimulus, the feedback loop is opened and accommodation rests at an intermediate myopic posture that typically varies individually between 0.5 diopter (D) and 4.5 D (Gilmartin, Hogan, & Thompson, 1984; Heron, Smith, & Winn, 1981; Leibowitz & Owens, 1975, 1978; Maddock, Millodot, Leat, & Johnson, 1981). This posture has been described as tonic accommodation (TA), (McBrien & Millodot, 1987) but terms such as dark focus (Leibowitz & Owens, 1975) or dark
⇑ Corresponding authors at: Western University of Health Sciences, College of Optometry, 309 E. Second St., Pomona, CA 91766, United States. E-mail addresses: [email protected] (C. Liu), [email protected] (C. Chase). 1 Current address: Coastline Community College, Fountain Valley, CA, United States. http://dx.doi.org/10.1016/j.visres.2016.08.010 0042-6989/Ó 2016 Elsevier Ltd. All rights reserved.
accommodation (Rosenfield, Ciuffreda, & Gilmartin, 1992) also have been used when the AR is measured in complete darkness under open-loop conditions. TA measurements are influenced by a number of factors, including which instrumentation method are used, task and environmental conditions, and mental state of the observer (see Rosenfield, Ciuffreda, Hung, & Gilmartin, 1993 for an extensive review). To open the accommodation loop and measure TA, one preferred method is to record accommodation posture in the dark with an open-field, infrared auto-refractor during passive viewing. In general, studies that have used the dark focus procedure find mean TA values in the range of 0.74–1.15 D (Andre & Owens, 1999; Bullimore & Gilmartin, 1989; Bullimore, Gilmartin, & Hogan, 1986; Gray, Strang, Winfield, Gilmartin, & Winn, 1998; McBrien & Millodot, 1987; Strang, Gilmartin, Gray, Winfield, & Winn, 2000). However, only a few of these studies reported correcting for distance residual uncorrected refraction error. Based on the combined results from two studies that adjusted for residual refractive error, a young adult sample of 226 students had a mean
26
C. Liu et al. / Vision Research 129 (2016) 25–32
dark TA value of 0.78 D (McBrien & Millodot, 1987; Strang et al., 2000). Many different factors have been found to influence the accuracy of AR under normal or abnormal binocular conditions, such as task instructions (Winn, Gilmartin, Mortimer, & Edwards, 1991), residual uncorrected refractive error (Hasebe, Nonaka, & Ohtsuki, 2005), higher-order aberrations of the eye (Hazel, Cox, & Strange, 2003), and pupil diameter (via mechanism of depth-offocus) (Wang & Ciuffreda, 2006). Based on the currently wellaccepted dual-interaction model of steady-state accommodation (Hung & Semmlow, 1980), the impact of TA on the steady state AR at near is usually considered to be minimal and the accommodative controller gain (ACG) would be the primary contributor to AR, assuming a normal depth of focus (Hung & Semmlow, 1980; Rosenfield et al., 1993). However, this conclusion was based on very limited empirical evidence (4 subjects’ data in Hung’s original 1980 paper) and computer simulation thereafter (Hung, 1998). In contrast to the conventional view that TA makes a minimal contribution to the AR, unpublished data from our research group showed a significant correlation between TA level and clinical accommodative testing such as accommodative facility (Liu, Chase, Drew, & Castellanos, 2013), leading us to re-examine the relationship. Among previous literature, Miller (1980) had revealed the only direct evidence that individuals’ AR to target located at various viewing distances was significantly correlated to their dark focus (TA) baseline. Miller examined this correlation under both monocular and binocular viewing conditions for the closed-loop AR. Results indicated that the relationship of AR to TA was stronger for the monocular condition than for the binocular condition. This difference indicates a potential role of the vergence system on AR through the convergence accommodation (CA) pathway. Some limitations exist in this study, such as low sample size (n = 13) and the lack of continuous monitoring of the AR during the task. Other studies also have examined the relationship between near work-induced TA adaptation and AR accuracy. Schor, Kotulak, and Tsuetaki (1986) found that the amplitude of near work-induced TA adaptation was reciprocally related to the amplitude of accommodative lag. But such an effect only manifested when the TA was assessed by opening the loop with Maxwellian view or Ganzfield, not with darkness. Owens and Wolf-Kelly (1987) also observed that one-hour of reading at 20 cm caused a myopic shift in both TA and monocular closed-loop AR. This increase in TA after reading was associated with less AR error. Rosenfield and Gilmartin (1999) assessed TA adaptation by comparing pre- and post-near task dark accommodation level. The adaptor group showed more than +0.30 D post-task adaptation in the initial 10 s and exhibited significant reduction in monocular closed-loop AR error during the near task. These results suggest that TA adaptation may improve the accuracy of AR under closed-loop conditions. The present study examined the potential impact of TA and near work induced TA-adaptation on the steady state AR. Most research groups have shown a large individual variability in their TA dataset similar to that of the original report from Leibowitz and coworkers (Leibowitz & Owens, 1975, 1978). Such variability is an important characteristic of dynamic biological system such as accommodation and vergence. Instead of looking at the group mean as a representative value, our approach has been to focus on the individual variability in different accommodative parameters and their correlations with each other. Accommodation responses during sustained monocular reading and to static targets at different viewing distances were compared to measures of baseline TA as well as post-reading TA adaptation. Under the vergenceaccommodation dual-interaction model, the vergence system would have an obvious impact on closed-loop AR under binocular viewing conditions. To isolate the accommodative system and to
focus on the pure influence of TA on steady state AR, most of our experiments were conducted under monocular conditions. To further test if vergence system involvement weakens this relationship, as shown previously by Miller, 1980, we also recorded AR during a sustained reading task under binocular conditions. 2. Method 2.1. Participants profile Forty-two first year graduate students, age between 22 and 29 years-old (24.5 ± 1.8), were recruited from the Western University of Health Sciences student body over a one-year period of time by solicitation during a first-year orientation session. There were 27 female and 15 male subjects. Western University of Health Sciences is a private institution that consists of nine graduate level colleges of different health professions. All participants received and signed informed consent approved by the Institutional Review Board at Western University. The research followed the tenets of the Declaration of Helsinki. All subjects were required to have a best-corrected monocular visual acuity of 20/25 or better, uncorrected (or residual) refractive error of 61.25 D hyperopia, 60.50 D myopia, 61.00 D uncorrected astigmatism or anisometropia. If correction was necessary for the subjects during the study, they were required to have habitual contact lens correction that had been prescribed and worn for a least one month. The requirement for contact lens correction during objective AR recording is necessitated by our study protocol of prolonged (10 min) continuous AR recording during reading task. The spectacle reflection resulted in excessive loss of data during autorefraction recording. The residual uncorrected refractive error was determined by making three static auto-refraction recordings while participants viewed the 20/25 row of a Snellen chart at 6 m from the right-eye. Recordings were made under subjects’ habitual viewing condition, either uncorrected, or wearing their habitual contact lens correction. Major exclusion criteria for this study included history of treatment for binocular disorder, corneal refractive surgery, epilepsy or head trauma, multiple sclerosis, Graves’s thyroid disease, myasthenia gravis, diabetes or Parkinson’s disease. Participants were also ineligible if they were currently taking non-SSRI anti-anxiety drugs, anti-arrhythmic agents, anticholinergic or tri-cyclic antidepressants. Additionally, participants that were deaf or stuttering were excluded. 2.2. Objective accommodation response measure All AR were measured monocularly from the right eye using the Grand Seiko WAM-5500, an open-field infrared auto-refractor (AIT Industries, Bensenville, IL, USA), set in either static recording mode or dynamic continuous recording mode. The left eye was occluded during the experiment unless otherwise stated. Although we did not directly monitor pupil size, the instrument limited data to be recorded only when pupil diameters were 3 mm or larger during recording (Mallen, Wolffsohn, Gilmartin, & Tsujimura, 2001), thus minimizing the effect of changes in pupil diameter on the AR. Calibration studies have shown the WAM to be accurate in both static and dynamic recording modes (Win-Hall & Glasser, 2009; Win-Hall, Houser, & Glasser, 2010). Several different accommodation measures were made during an hour-long test session. First, measurement of monocular AR for static targets was recorded. The procedures had previously been described (Tosha, Borsting, Ridder, & Chase, 2009). A 2 cm target of high-contrast (Michelson = 79%) star symbol was presented at five viewing distances measured from the corneal plane in a
C. Liu et al. / Vision Research 129 (2016) 25–32
fixed order: 600, 50, 33, 25, and 20 cm and subjects were instructed to keep the target in clear focus for the entire viewing time of two minutes. The accommodative response was recorded continuously with a 5 Hz sampling rate and data from the 2-min recording were averaged. Two sets of Baseline TA and post-task TA were recorded for monocular and binocular reading task, and were labeled as TA1 and TA2 respectively (please see the Fig. 1 below for experimental task sequence). After the static recordings were made, the subjects were dark adapted for 5 min in a dark room with a dark cloth covering the head and shoulders. During this time, the subjects were allowed to close their eyes. Baseline TA1 (monocular) was recorded continuously for 2 min after dark adaptation while the subjects were instructed to look straight ahead. Consistent with other studies, the early portion of the TA baseline recording was discarded while subjects adjusted to the dark, and only the last 90 s of the recording were averaged for TA baseline. After Baseline TA1 recordings, participants read monocularly for 10 min, during which continuous AR recordings were made (Monocular Reading AR). A short story was presented on a computer screen set at 33.3 cm, with each paragraph containing six lines of text with a print size where the height of a lower case letter ‘‘x” subtending a visual angle of 2.5 min of arc (Snellen equivalent of 20/50). The subtended horizontal visual angle of each line was small enough so that most subjects would make up to 3 horizontal saccades for reading each line. This ensures that the horizontal saccadic eye movement during reading would not interfere with continuous AR recording. Subjects pressed a button when they reached the end of the paragraph to present the next paragraph. Immediately after the completion of the 10 min reading task, post-task TA1 was re-assessed continuously for 2 min by completely darkening the room and covering the subjects’ head and shoulders with a dark cloth. This maneuver usually took less than 5 s. Subjects were instructed to keep looking straight ahead to maintain stable fixation. The last 90 s of the recording were averaged for post-task TA1. After the post-task TA1 recording, while remaining in total darkness, participants were asked three yes/no questions about the story to ensure that attention was maintained during reading. The subjects were dark adapted for another 5 min and Baseline TA2 (binocular) was recorded continuously for 2 min. Afterward, 10 min of 3 diopter reading experiment, with the target located at 33.3 cm distance, was repeated under binocular viewing condition (Binocular Reading AR), and post-task TA2 was recorded with the same procedure. This sequence of events is shown in the following graph (Fig. 1). TA adaptation induced by either the monocular or binocular reading task was calculated by subtracting subjects’ baseline TA from their post-task TA respectively. A positive value for TA adaptation indicates an increase in TA level after 10 min of continuous reading at the 3 D distance, thus a myopic shift. Conversely, a negative value for TA adaptation indicates a decrease in TA level, thus a hyperopic shift. The average of the residual uncorrected refractive error was subtracted from TA and AR to adjust accommodation measurements in the data analysis.
27
2.3. Data analysis Due to AR recording errors, data was missing for seven participants in some of the test conditions. Data were analyzed using all participants, resulting in differences in the sample size for conditions where data were missing (see Table 1). Data were analyzed by linear regression correlations and t-test comparisons made using Prism 6.0 (GraphPad Software, San Diego, CA). 3. Results 3.1. Descriptive statistics Table 1 provides descriptive statistics for different aspects of accommodative function. The baseline TA1 frequency distribution is shown in Fig. 2. Similar to previous studies, (Leibowitz & Owens, 1978; McBrien & Millodot, 1987) TA scores had an asymmetrical positive distribution (Skewness = 0.60). The mean TA of 1.84D is larger than previous studies (McBrien & Millodot, 1987; Strang et al., 2000) that used similar recording methods and in which data were corrected for distance refractive error (t(41) = 5.34, p < 0.0001, 95% CI[1.44, 2.24]). 3.2. Relationship between baseline TA and AR to static target Regression analysis between Baseline TA (independent variable) and objective measurement of AR during different tasks (dependent variable) are presented in Fig. 3 and Table 2. Fig. 3 shows scatterplots for accommodation responses (AR) to static targets and monocular baseline TA (TA1). Linear regression analysis shows that Baseline TA1 accounted for significant amounts (32–41%) of variance in monocular AR to static targets at every viewing distance. Accommodation errors for those with a higher TA level tended to lead the target (above the solid stimulus demand line), but those with a lower TA tended to show accommodative lags (below the solid stimulus demand line). Linear regression equations for different static target distances are presented in Table 2. Slopes of the regression equation suggest that each diopter of Baseline TA1 is associated with about onethird of a diopter change to the AR. 3.3. Relationship between TA and AR during continuous reading task One might argue that accommodative responses to static targets may not accurately represent subjects’ responses during natural reading tasks due to the factors such as cognitive demand and saccadic eye movement changes. To test this hypothesis, we compared the AR for the 3D static target with the AR during reading. Regression analysis between Baseline TA (predictor variable) and AR during reading tasks (predicted variable) were performed and the results are presented in Fig. 4 and Table 2. Fig. 4 shows a side-by-side comparison of the scatterplot for Baseline TA and AR to continuous reading task at 3 D, both monocular (Fig. 4B) and binocular (Fig. 4C), with the scatterplot for AR to 3 D static targets
Fig. 1. Schematic representation of experimental task sequence.
28
C. Liu et al. / Vision Research 129 (2016) 25–32
Table 1 Descriptive statistics for different AR measures described in Fig. 1. Measure
N
Mean
Median
SD
Min
Max
Baseline TA1 (D) Post-task TA1 (D) Monocular Reading AR (D) Monocular TA Adaptation (D) Baseline TA2 (D) Post-task TA2 (D) Binocular Reading AR (D) Binocular TA Adaptation (D)
42 41 40 40 41 39 40 38
1.84 2.05 2.27 0.23 1.79 1.70 2.39 0.15
1.38 1.72 2.25 0.17 1.64 1.47 2.33 0.04
1.29 1.13 0.54 0.87 1.14 0.99 0.53 0.71
0.43 0.06 1.49 1.63 0.19 0.22 1.65 2.60
5.14 4.73 3.87 2.68 4.86 4.22 3.64 1.47
myopic or hyperopic shift (Fig. 5). Comparing group means alone might mask the individual effect. Therefore, we calculated linear regressions between reading AR (predictor variable) and the individual TA adaptation (predicted variable) for monocular and binocular reading tasks, respectively. Data are presented in Table 3. Results suggested that near work-induced TA adaptation was not significantly correlated with AR error. These results support the conclusion that only the tonic accommodation level, not near work-induced TA adaptation, was associated with the accommodative response during the near visual task. 4. Discussion
Fig. 2. Frequency distribution of baseline tonic accommodation (TA1).
and Baseline TA1 from Fig. 2 (Fig. 4A). The linear regression equations for 3 D continuous reading task are presented in Table 2. Similar to the static target at the same demand level, Baseline TA1 accounted for significant amounts (45%) of the variance in AR under the monocular reading task (Fig. 4B). The correlation between Baseline TA2 and AR during reading is weakened under the binocular reading condition, accounting for a smaller amount (24.3%) of variance in AR (Fig. 4C). AR during binocular reading is slightly higher than that during monocular reading (Table 1), although a paired t-test showed that the difference was not statistically significant (t = 1.627; p = 0.112). The slopes of these regression plots suggest that every diopter of baseline TA is associated with about one-quarter to one-third of a diopter to the AR during natural reading conditions. Together, results from Table 2, Fig. 3 and Fig. 4 suggest that Baseline TA is significantly correlated with closed-loop accommodative responses at different viewing distances and different accommodative task. 3.4. Relationship between near work-induced TA adaptation and Closed-loop AR during near visual tasks To study the relationship between near work-induced TA adaptation and AR, we examined if ten minutes of reading at a 3 D distance had induced a significant change in TA level. For monocular reading task, Post-task TA1 showed only a small increase (myopic shift) of 0.23 D compared to Baseline TA1. A paired t-test showed that Post-task TA1 was not significantly different from Baseline TA1 (t = 1.650; p = 0.107). For a binocular reading task, Post-task TA2 showed a small decrease (hyperopic shift) of 0.15 D comparing to Baseline TA2. Similar to monocular reading, a paired t-test showed that Post-task TA2 was not significantly different from Baseline TA2 (t = 1.268; p = 0.213). Although group means of TA adaptation showed no significant changes after monocular or binocular reading tasks, individual subjects exhibited different adaptation patterns, showing either
This study found a strong association between open-loop TA and closed-loop AR. Every diopter of Baseline TA1 was associated with approximately one-third of a diopter of steady-state accommodative response during a monocular accommodative task. This ratio was consistent across a variety of viewing distances and different near work tasks. These correlations do not prove that tonic accommodation directly contributes to the steady state accommodation response. Nevertheless, they are consistent with qualitative and quantitative models in which the individual tonic accommodation level is the starting point (or home) from which various stimuli such as defocus, proximity, voluntary effort, and vergence accommodation operate to provide an accommodation response closer to the stimulus (Charman, 1986; Heath, 1956; Hung, 1992; Hung, Ciuffreda, & Rosenfield, 1996). Importantly, these current findings also demonstrate that the steady-state accommodation response is not as fully independent of the tonic accommodation starting point as is commonly believed. Our findings suggest that the assumptions made for the classic model of steady-state accommodation developed by Hung and Semmlow (1980) might need to be re-examined. We argue for the adoption of individual rather than group mean values for estimating model parameters. Hung and Semmlow’s model parameters were built on data obtained from only four subjects and two were over 30 years-old (age ranging from 18 to 37, TA level ranging from 0.42 D to 2.2 D). The TA or ABIAS (the term used in this model) was incorporated into the steady state AR calculation model as well as calculation for Accommodative Controller Gain for Positive Accommodative Error (ACGP) (see the following equations cited from the Hung & Semmlow, 1980 study. AS represents accommodative stimulus; DSP represents deadspace in accommodation, also known as ‘‘depth-of-focus”).
ACGP ACGP DSP 1 þ ACGP 1 þ ACGP 1 þ ABIAS 1 þ ACGP
AR ¼ AS
ACG ¼
AR ABIAS AE DSP
ðFormula 1Þ ðFormula 2Þ
C. Liu et al. / Vision Research 129 (2016) 25–32
29
Fig. 3. Scatterplot of the effect of TA1 on accommodation responses while viewing static targets at different viewing distances. Linear regression lines are plotted as dotted lines; the target’s accommodative stimulus demand is plotted as a solid, horizontal line.
Based on Subject JS’s data, who was 37 year-old and had a low ABIAS (or TA) of 0.42 D and high ACGP of 8.53, Hung and Semmlow concluded that ABIAS contributed minimally to AR, only accounted for 1.8% of the AR for a 3 D accommodative stimulus.
Rosenfield et al.’s (1993) review assessed the role of TA in closed-loop accommodation by using Hung’s model. Based on the work of Mordi and Ciuffreda (1998), Rosenfield et al. (1993) assumed that the ACG value was 10 and concluded that TA only
30
C. Liu et al. / Vision Research 129 (2016) 25–32
Table 2 The relationship between baseline TA (x variable) and AR (y variable) at different viewing conditions. Target distance
Linear regression equation
95% CI for slope
p-value for nonzero slope
0.1 D Static* 2 D Static* 3 D Static* 4 D Static* 5 D Static* 3 D Monocular reading* 3 D Binocular readingy
Y = 0.296*X 0.10 Y = 0.335*X + 1.02 Y = 0.355*X + 1.87 Y = 0.408*X + 2.69 Y = 0.425*X + 3.57 Y = 0.277*X + 1.77
0.18, 0.20, 0.19, 0.21, 0.25, 0.17,
0.41 0.47 0.53 0.60 0.60 0.38
<0.005 <0.005 <0.005 <0.005 <0.005 <0.005
Y = 0.229*X + 1.98
0.10, 0.36
<0.005
* Monocular tasks, Baseline TA1 was used as independent variable (x) in regression analysis. y Binocular task, Baseline TA2 was used as independent variable (x) in regression analysis.
contributed less than 0.1 D in AR to a 5 D stimulus. Using a similar approach, Hung assessed the impact of ABIAS (TA) on the accommodative stimulus–response function by varying ABIAS (TA) values while holding ACG at a constant value of 10 in their computer model. (Hung, 1998) Computer simulation revealed that when holding ACG at constant value of 10, varying ABIAS (TA) had minimum impact on the slope of accommodative stimulus–response function. In contrast, when holding ABIAS (TA) at constant value of 1 D, varying the ACG value (from 1 to 20) had a significant effect on the slope, with increased ACG value associated with a steeper slope. Based on this result, Hung et al. drew the same conclusion that TA had minimum influence on AR under normal closed-loop condition. These modeling results, however, did not explore the possible co-variation between TA and ACG. According to Formula 2 and the trend of empirical data from Table 1 in Hung and Semmlow (1980) paper, ABIAS (TA) inversely correlates to ACG so that when TA increases, ACG drops significantly. Thus, the higher the TA value, the larger the proportion of AR will be contributed from TA. Holding the ACG parameter constant in the simulation when empirical data shows that ACG varies individually might not be optimal for the modeling the dynamics of the accommodative system. For example, when subject JM’s data (ABIAS (TA) = 2.2 D, ACGP = 2.32) from Hung and Semmlow (1980) paper is used for model calculation, ABIAS (TA) contributed 33.4% to AR for the same 3 D accommodative stimulus. This subject’s ABIAS (TA) values lie
close to our group mean for TA (1.84 D), which might explain why these subjects have results that are similar to those reported in our study. In our study we adopted a unique approach, studying the effects of individual variability and co-variation between TA and closedloop AR. In addition, all subjects’ data in our study, including both those with normal or abnormal accommodative function, were included in the analysis, allowing us to examine this correlation under a wide spectrum of functioning level. As described above, most previous studies assumed that group means for these values applied to all individuals in the sample. Using the group mean for TA value presumes that all adopt an intermediate posture of 1.84 D, yet our subjects showed a very widely distributed TA level, ranging from 0.43 D to +5.14 D. The variability found in our experiment is consistent with other studies that assessed TA using different methodologies (Gilmartin et al., 1984; Heron et al., 1981; Leibowitz & Owens, 1975, 1978; Maddock et al., 1981). Similarly, a wide distribution of ACG values has been previously reported for normal adult subjects, ranging from 1.5 to 21.0 with mean value around 10 (Ciuffreda, Hokoda, Hung, & Semmlow, 1984; Mordi & Ciuffreda, 1998; Ong, Ciuffreda, & Tannen, 1993). Although we did not calculate the ACG value in the current study, one can assume a similar distribution pattern of ACG in our sample. Models using a fixed value of 10 for the ACG and 1 D for TA may mask the true correlation between TA level and AR. Consistent with Miller (1980), our results suggested that the correlation between TA and closed-loop AR is more prominent under monocular viewing conditions comparing to binocular viewing conditions. Both the R-square value and the regression equation slope in regression analysis are lower for binocular reading conditions. This finding confirmed that the vergence system would also have a significant impact on closed-loop AR through the CA pathway which would provide additional stimulus to the accommodative system under binocular viewing conditions. Previous studies have found that the accommodative lag at near is greater when viewing under monocular viewing conditions (Jiang, Tea, & O’Donnell, 2007; Nakatsuka, Hasebe, Nonaka, & Ohtsuki, 2003). Although not statistically significant, our data did show a trend of a slight increase in AR during binocular reading compared to monocular reading, consistent with the previous literature. It is possible that some individual may utilize CA to compensate for weak accommodative system. If this indeed is the case, one might hypothesize that the compensatory effect would only be possible
Fig. 4. Scatterplot of the effect of TA1 on the accommodation response while viewing: (A) a monocular 3D static target and (B) continuous monocular reading at 3D. The effect of Baseline TA2 on the accommodation response while performing continuous binocular reading at 3D is shown in Fig. 4C. Linear regression lines are plotted as dotted lines; the target is plotted as a solid, horizontal line.
C. Liu et al. / Vision Research 129 (2016) 25–32
31
responses observed in our group. Further study is necessary to clarify this issue. To our knowledge, this is the first study to examine the relationship between TA and the steady state AR, using an objective instrument to continuously monitor the accommodation responses during both sustained reading task and to static targets at different viewing distances. Our results suggest that the steady-state accommodation response is not as independent of individual TA as is commonly thought. A significant correlation is found between the two variables. These findings encourage a re-examination of the classic model of steady-state accommodation. Disclosure C. Liu, None; S. Drew, None; E. Borsting, None; A. Escobar, None; L. Stark, None; C. Chase, None. Acknowledgments This research project is supported by National Eye Institute grant (1R15EY021021 Chase, C. (PI)). References
Fig. 5. Frequency distribution of near work-induced TA adaptation: (A) monocular reading task, (B) binocular reading task.
Table 3 The relationship between near work induced-TA adaptation (x variable) and AR during reading (y variable). Task Monocular reading Binocular readingy
Linear regression equation *
*
Y = 0.077 X + 2.314 Y = 0.052*X + 2.404
R2
p-value
0.017 0.005
0.441 0.685
* Monocular reading, Monocular TA Adaptation was used as independent variable (x) in regression analysis; y Binocular reading task, Binocular TA Adaptation was used as independent variable (x) in regression analysis.
in individual who has strong vergence system. Further study is necessary to test this hypothesis. In contrast to some of the previous literature, our subject group did not show a significant near-work induced TA adaptation when comparing Post-task TA to Baseline TA, although large individual variability exists within group with both myopic and hyperopic TA adaptation. Further correlational analysis also failed to find any statistically significant relation between individual AR and TA adaptation. Instead, our data indicated that only the tonic TA level itself correlated to the accuracy of the AR. Our adapting task was similar to Rosenfield and Gilmartin (1999) who used a 3 D target and 10 min of near viewing. In the Rosenfeld et al. study a significant number of subjects did not adapt to the near task and in turn did not show a reduction in accommodative lag. It is possible that our sample of subjects could have included a significant number of non-adaptors since we recruited adults with a broad range of accommodative function. This may account for the lack of adaptive
Andre, J. T., & Owens, D. A. (1999). Predicting optimal accommodative performance from measures of the dark focus of accommodation. Human Factors, 41(1), 139–145. Bullimore, M. A. M., & Gilmartin, B. (1989). The measurement of adaptation of tonic accommodation under two open-loop conditions. Ophthalmic & Physiological Optics, 9(1), 72–75. http://dx.doi.org/10.1111/j.1475-1313.1989.tb00810.x. Bullimore, M. A., Gilmartin, B., & Hogan, R. E. (1986). Objective and subjective measurment of tonic accommodation. Ophthalmic & Physiological Optics, 6(1), 57–62. http://dx.doi.org/10.1111/j.1475-1313.1986.tb00700.x. Charman, W. N. (1986). Static accommodation and the minimum angle of resolution. American Journal of Optometry and Physiological Optics, 63, 915–921. Ciuffreda, K. J., Hokoda, S. C., Hung, G. K., & Semmlow, J. L. (1984). Accommodative stimulus/response function in human amblyopia. Documenta Ophthalmologica, 56(4), 303–326. Gilmartin, B., Hogan, R. E., & Thompson, S. M. (1984). The effect of Timolol Maleate on tonic accommodation, tonic vergence, and pupil diameter. Investigative Ophthalmology & Visual Science, 25(6), 763–770. Gray, L., Strang, N., Winfield, N., Gilmartin, B., & Winn, B. (1998). The magnitude and distribution of open-loop accommodation using three different methods of opening the loop. Optometry and Vision Science, 75(12), 897–902. Hasebe, S., Nonaka, F., & Ohtsuki, H. (2005). Accuracy of accommodation in heterophoric patients: Testing an interaction model in a large clinical sample. Ophthalmic & Physiological Optics, 25, 582–591. Hazel, C. A., Cox, M. J., & Strange, N. C. (2003). Wavefront aberration and its relationship to the accommodative stimulus-response function in myopic subjects. Optometry and Vision Science, 80(2), 151–158. Heath, G. G. (1956). Components of accommodation. American Journal of Optometry and Archives of American Academy of Optometry, 33, 569–579. Heron, G., Smith, A. C., & Winn, B. (1981). The influence of method on the stability of dark focus position of accommodation. Ophthalmic & Physiological Optics, 1(2), 79–90. http://dx.doi.org/10.1111/j.1475-1313.1981.tb00151.x. Hung, G. K. (1992). Adaptation model of accommodation and vergence. Ophthalmic & Physiological Optics, 12, 319–326. Hung, G. K. (1998). Sensitivity analysis of the stimulus-response function of a static nonlinear accommodation model. IEEE Transactions on Bio-medical Engineering, 45(3), 335–341. http://dx.doi.org/10.1109/10.661158. Hung, G. K., Ciuffreda, K. J., & Rosenfield, M. (1996). Proximal contribution to a linear static model of accommodation and vergence. Ophthalmic & Physiological Optics, 16, 31–41. Hung, G. K., & Semmlow, J. L. (1980). Static behavior of accommodation and vergence: Computer simulation of an interactive dual-feedback system. IEEE Transactions on Bio-medical Engineering, 27(8), 439–447. http://dx.doi.org/ 10.1109/TBME.1980.326752. Jiang, B.-C., Tea, Y. C., & O’Donnell, D. (2007). Changes in accommodative and vergence responses when viewing through near addition lenses. Optometry, 78 (3), 129–134. http://dx.doi.org/10.1016/j.optm.2006.08.017. Leibowitz, H. W., & Owens, D. A. D. (1975). Anomalous myopias and the intermediate dark focus of accommodation. Science, 189(4203), 646–648. Leibowitz, H. W. H., & Owens, D. A. D. (1978). New evidence for the intermediate position of relaxed accommodation. Documenta Ophthalmologica, 46(1), 133–147. Liu, C., Chase, C., Drew, S., & Castellanos, E. (2013). Tonic accommodation correlates with accommodative amplitude-scaled facility test for symptomatic graduate students. Investigative Ophthalmology & Visual Science, 54 (E-abstract 4259).
32
C. Liu et al. / Vision Research 129 (2016) 25–32
Maddock, R. J. R., Millodot, M. M., Leat, S. S., & Johnson, C. A. C. (1981). Accommodation responses and refractive error. Investigative Ophthalmology & Visual Science, 20(3), 387–391. Mallen, E. A. H., Wolffsohn, J. S., Gilmartin, B., & Tsujimura, S. (2001). Clinical evaluation of the Shin-Nippon SRW-5000 autorefractor in adults. Ophthalmic & Physiological Optics, 21, 101–107. McBrien, N. A., & Millodot, M. (1987). The relationship between tonic accommodation and refractive error. Investigative Ophthalmology & Visual Science, 28(6), 997–1004. Miller, R. J. (1980). Ocular vergence-induced accommodation and its relation to dark focus. Perception & Psychophysics, 28(2), 125–132. http://dx.doi.org/10.3758/ BF03204337. Mordi, J. A., & Ciuffreda, K. J. (1998). Static aspects of accommodation: Age and presbyopia. Vision Research, 38(11), 1643–1653. http://dx.doi.org/10.1016/ S0042-6989(97)00336-2. Nakatsuka, C., Hasebe, S., Nonaka, F., & Ohtsuki, H. (2003). Accommodative lag under habitual seeing conditions: Comparison between adult myopes and emmetropes. Japanese Journal of Ophthalmology, 47(3), 291–298. Ong, E., Ciuffreda, K. J., & Tannen, B. (1993). Static accommodation in congenital nystagmus. Investigative Ophthalmology & Visual Science, 34(1), 194–204. Owens, D. A., & Wolf-Kelly, K. (1987). Near work, visual fatigue, and variations of oculomotor tonus. Investigative Ophthalmology & Visual Science, 28(4), 743–749. Rosenfield, M., Ciuffreda, K. J., & Gilmartin, B. (1992). Factors influencing accommodative adaptation. Optometry and Vision Science, 69(4), 270–275. Rosenfield, M., Ciuffreda, K. J., Hung, G. K., & Gilmartin, B. (1993). Tonic accommodation: A review. I. Basic aspects. Ophthalmic & Physiological Optics, 13(3), 266–284.
Rosenfield, M., & Gilmartin, B. (1999). Accommodative error, adaptation and myopia. Ophthalmic & Physiological Optics, 19(2), 159–164. Schor, C. M., Kotulak, J. C., & Tsuetaki, T. (1986). Adaptation of tonic accommodation reduces accommodative lag and is masked in darkness. Investigative Ophthalmology & Visual Science, 27(5), 820–827. Strang, N., Gilmartin, B. S. B., Gray, L., Winfield, N. R. N., & Winn, B. (2000). Openloop accommodation in emmetropia and myopia. Current Eye Research, 20(3), 190–194. Toates, F. M. (1972). Accommodation function of the human eye. Physiological Reviews, 52(4), 828–863. Tosha, C., Borsting, E., Ridder, W. H., & Chase, C. (2009). Accommodation response and visual discomfort. Ophthalmic & Physiological Optics, 29(6), 625–633. http:// dx.doi.org/10.1111/j.1475-1313.2009.00687.x. Wang, B., & Ciuffreda, K. J. (2006). Depth-of-focus of the human eye: Theory and clinical implications. Survey of Ophthalmology, 51(1), 75–85. Win-Hall, D. M., & Glasser, A. (2009). Objective accommodation measurements in pseudophakic subjects using an autorefractor and an aberrometer. Journal of Cataract and Refractive Surgery, 35(2), 282–290. http://dx.doi.org/10.1016/j. jcrs.2008.10.033. Win-Hall, D. M., Houser, J., & Glasser, A. (2010). Static and dynamic accommodation measured using the WAM-5500 autorefractor. Optometry and Vision Science, 87 (11), 873–882. http://dx.doi.org/10.1097/OPX.0b013e3181f6f98f. Winn, B., Gilmartin, B., Mortimer, L. C., & Edwards, N. R. (1991). The effect of mental effort on open- and closed-loop accommodation. Ophthalmic & Physiological Optics, 11(4), 335–339.
Contents lists available at ScienceDirect
Vision Research journal homepage: www.elsevier.com/locate/visres
Tonic accommodation predicts closed-loop accommodation responses Chunming Liu a,⇑, Stefanie A. Drew a,b, Eric Borsting c, Amy Escobar a,1, Lawrence Stark c, Christopher Chase a,⇑ a
Western University of Health Sciences, College of Optometry, Pomona, CA, United States California State University Northridge, Northridge, CA, United States c Southern California College of Optometry at Marshall B. Ketchum University, Fullerton, CA, United States b
a r t i c l e
i n f o
Article history: Received 14 July 2015 Received in revised form 24 August 2016 Accepted 26 August 2016 Available online 1 November 2016 Keywords: Tonic accommodation Accommodation response TA-adaptation Vergence
a b s t r a c t The purpose of this study is to examine the potential relationship between tonic accommodation (TA), near work induced TA-adaptation and the steady state closed-loop accommodation response (AR). Forty-two graduate students participated in the study. Various aspects of their accommodation system were objectively measured using an open-field infrared auto-refractor (Grand Seiko WAM-5500). Tonic accommodation was assessed in a completely dark environment. The association between TA and closed-loop AR was assessed using linear regression correlations and t-test comparisons. Initial mean baseline TA was 1.84 diopter (D) (SD ± 1.29 D) with a wide distribution range (0.43 D to 5.14 D). For monocular visual tasks, baseline TA was significantly correlated with the closed-loop AR. The slope of the best fit line indicated that closed-loop AR varied by approximately 0.3 D for every 1 D change in TA. This ratio was consistent across a variety of viewing distances and different near work tasks, including both static targets and continuous reading. Binocular reading conditions weakened the correlation between baseline TA and AR, although results remained statistically significant. The 10 min near reading task with a 3 D demand did not reveal significant near work induced TA-adaptation for either monocular or binocular conditions. Consistently, the TA-adaptation did not show any correlation with AR during reading. This study found a strong association between open-loop TA and closed-loop AR across a variety of viewing distances and different near work tasks. Difference between the correlations under monocular and binocular reading condition suggests a potential role for vergence compensation during binocular closed-loop AR. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Under natural viewing conditions, the accommodative system of the human eye mainly employs a retinal blur signal to drive an accurate accommodative response (AR) using a closed-loop negative feedback system under monocular condition (Toates, 1972). In the absence of an adequate visual stimulus, the feedback loop is opened and accommodation rests at an intermediate myopic posture that typically varies individually between 0.5 diopter (D) and 4.5 D (Gilmartin, Hogan, & Thompson, 1984; Heron, Smith, & Winn, 1981; Leibowitz & Owens, 1975, 1978; Maddock, Millodot, Leat, & Johnson, 1981). This posture has been described as tonic accommodation (TA), (McBrien & Millodot, 1987) but terms such as dark focus (Leibowitz & Owens, 1975) or dark
⇑ Corresponding authors at: Western University of Health Sciences, College of Optometry, 309 E. Second St., Pomona, CA 91766, United States. E-mail addresses: [email protected] (C. Liu), [email protected] (C. Chase). 1 Current address: Coastline Community College, Fountain Valley, CA, United States. http://dx.doi.org/10.1016/j.visres.2016.08.010 0042-6989/Ó 2016 Elsevier Ltd. All rights reserved.
accommodation (Rosenfield, Ciuffreda, & Gilmartin, 1992) also have been used when the AR is measured in complete darkness under open-loop conditions. TA measurements are influenced by a number of factors, including which instrumentation method are used, task and environmental conditions, and mental state of the observer (see Rosenfield, Ciuffreda, Hung, & Gilmartin, 1993 for an extensive review). To open the accommodation loop and measure TA, one preferred method is to record accommodation posture in the dark with an open-field, infrared auto-refractor during passive viewing. In general, studies that have used the dark focus procedure find mean TA values in the range of 0.74–1.15 D (Andre & Owens, 1999; Bullimore & Gilmartin, 1989; Bullimore, Gilmartin, & Hogan, 1986; Gray, Strang, Winfield, Gilmartin, & Winn, 1998; McBrien & Millodot, 1987; Strang, Gilmartin, Gray, Winfield, & Winn, 2000). However, only a few of these studies reported correcting for distance residual uncorrected refraction error. Based on the combined results from two studies that adjusted for residual refractive error, a young adult sample of 226 students had a mean
26
C. Liu et al. / Vision Research 129 (2016) 25–32
dark TA value of 0.78 D (McBrien & Millodot, 1987; Strang et al., 2000). Many different factors have been found to influence the accuracy of AR under normal or abnormal binocular conditions, such as task instructions (Winn, Gilmartin, Mortimer, & Edwards, 1991), residual uncorrected refractive error (Hasebe, Nonaka, & Ohtsuki, 2005), higher-order aberrations of the eye (Hazel, Cox, & Strange, 2003), and pupil diameter (via mechanism of depth-offocus) (Wang & Ciuffreda, 2006). Based on the currently wellaccepted dual-interaction model of steady-state accommodation (Hung & Semmlow, 1980), the impact of TA on the steady state AR at near is usually considered to be minimal and the accommodative controller gain (ACG) would be the primary contributor to AR, assuming a normal depth of focus (Hung & Semmlow, 1980; Rosenfield et al., 1993). However, this conclusion was based on very limited empirical evidence (4 subjects’ data in Hung’s original 1980 paper) and computer simulation thereafter (Hung, 1998). In contrast to the conventional view that TA makes a minimal contribution to the AR, unpublished data from our research group showed a significant correlation between TA level and clinical accommodative testing such as accommodative facility (Liu, Chase, Drew, & Castellanos, 2013), leading us to re-examine the relationship. Among previous literature, Miller (1980) had revealed the only direct evidence that individuals’ AR to target located at various viewing distances was significantly correlated to their dark focus (TA) baseline. Miller examined this correlation under both monocular and binocular viewing conditions for the closed-loop AR. Results indicated that the relationship of AR to TA was stronger for the monocular condition than for the binocular condition. This difference indicates a potential role of the vergence system on AR through the convergence accommodation (CA) pathway. Some limitations exist in this study, such as low sample size (n = 13) and the lack of continuous monitoring of the AR during the task. Other studies also have examined the relationship between near work-induced TA adaptation and AR accuracy. Schor, Kotulak, and Tsuetaki (1986) found that the amplitude of near work-induced TA adaptation was reciprocally related to the amplitude of accommodative lag. But such an effect only manifested when the TA was assessed by opening the loop with Maxwellian view or Ganzfield, not with darkness. Owens and Wolf-Kelly (1987) also observed that one-hour of reading at 20 cm caused a myopic shift in both TA and monocular closed-loop AR. This increase in TA after reading was associated with less AR error. Rosenfield and Gilmartin (1999) assessed TA adaptation by comparing pre- and post-near task dark accommodation level. The adaptor group showed more than +0.30 D post-task adaptation in the initial 10 s and exhibited significant reduction in monocular closed-loop AR error during the near task. These results suggest that TA adaptation may improve the accuracy of AR under closed-loop conditions. The present study examined the potential impact of TA and near work induced TA-adaptation on the steady state AR. Most research groups have shown a large individual variability in their TA dataset similar to that of the original report from Leibowitz and coworkers (Leibowitz & Owens, 1975, 1978). Such variability is an important characteristic of dynamic biological system such as accommodation and vergence. Instead of looking at the group mean as a representative value, our approach has been to focus on the individual variability in different accommodative parameters and their correlations with each other. Accommodation responses during sustained monocular reading and to static targets at different viewing distances were compared to measures of baseline TA as well as post-reading TA adaptation. Under the vergenceaccommodation dual-interaction model, the vergence system would have an obvious impact on closed-loop AR under binocular viewing conditions. To isolate the accommodative system and to
focus on the pure influence of TA on steady state AR, most of our experiments were conducted under monocular conditions. To further test if vergence system involvement weakens this relationship, as shown previously by Miller, 1980, we also recorded AR during a sustained reading task under binocular conditions. 2. Method 2.1. Participants profile Forty-two first year graduate students, age between 22 and 29 years-old (24.5 ± 1.8), were recruited from the Western University of Health Sciences student body over a one-year period of time by solicitation during a first-year orientation session. There were 27 female and 15 male subjects. Western University of Health Sciences is a private institution that consists of nine graduate level colleges of different health professions. All participants received and signed informed consent approved by the Institutional Review Board at Western University. The research followed the tenets of the Declaration of Helsinki. All subjects were required to have a best-corrected monocular visual acuity of 20/25 or better, uncorrected (or residual) refractive error of 61.25 D hyperopia, 60.50 D myopia, 61.00 D uncorrected astigmatism or anisometropia. If correction was necessary for the subjects during the study, they were required to have habitual contact lens correction that had been prescribed and worn for a least one month. The requirement for contact lens correction during objective AR recording is necessitated by our study protocol of prolonged (10 min) continuous AR recording during reading task. The spectacle reflection resulted in excessive loss of data during autorefraction recording. The residual uncorrected refractive error was determined by making three static auto-refraction recordings while participants viewed the 20/25 row of a Snellen chart at 6 m from the right-eye. Recordings were made under subjects’ habitual viewing condition, either uncorrected, or wearing their habitual contact lens correction. Major exclusion criteria for this study included history of treatment for binocular disorder, corneal refractive surgery, epilepsy or head trauma, multiple sclerosis, Graves’s thyroid disease, myasthenia gravis, diabetes or Parkinson’s disease. Participants were also ineligible if they were currently taking non-SSRI anti-anxiety drugs, anti-arrhythmic agents, anticholinergic or tri-cyclic antidepressants. Additionally, participants that were deaf or stuttering were excluded. 2.2. Objective accommodation response measure All AR were measured monocularly from the right eye using the Grand Seiko WAM-5500, an open-field infrared auto-refractor (AIT Industries, Bensenville, IL, USA), set in either static recording mode or dynamic continuous recording mode. The left eye was occluded during the experiment unless otherwise stated. Although we did not directly monitor pupil size, the instrument limited data to be recorded only when pupil diameters were 3 mm or larger during recording (Mallen, Wolffsohn, Gilmartin, & Tsujimura, 2001), thus minimizing the effect of changes in pupil diameter on the AR. Calibration studies have shown the WAM to be accurate in both static and dynamic recording modes (Win-Hall & Glasser, 2009; Win-Hall, Houser, & Glasser, 2010). Several different accommodation measures were made during an hour-long test session. First, measurement of monocular AR for static targets was recorded. The procedures had previously been described (Tosha, Borsting, Ridder, & Chase, 2009). A 2 cm target of high-contrast (Michelson = 79%) star symbol was presented at five viewing distances measured from the corneal plane in a
C. Liu et al. / Vision Research 129 (2016) 25–32
fixed order: 600, 50, 33, 25, and 20 cm and subjects were instructed to keep the target in clear focus for the entire viewing time of two minutes. The accommodative response was recorded continuously with a 5 Hz sampling rate and data from the 2-min recording were averaged. Two sets of Baseline TA and post-task TA were recorded for monocular and binocular reading task, and were labeled as TA1 and TA2 respectively (please see the Fig. 1 below for experimental task sequence). After the static recordings were made, the subjects were dark adapted for 5 min in a dark room with a dark cloth covering the head and shoulders. During this time, the subjects were allowed to close their eyes. Baseline TA1 (monocular) was recorded continuously for 2 min after dark adaptation while the subjects were instructed to look straight ahead. Consistent with other studies, the early portion of the TA baseline recording was discarded while subjects adjusted to the dark, and only the last 90 s of the recording were averaged for TA baseline. After Baseline TA1 recordings, participants read monocularly for 10 min, during which continuous AR recordings were made (Monocular Reading AR). A short story was presented on a computer screen set at 33.3 cm, with each paragraph containing six lines of text with a print size where the height of a lower case letter ‘‘x” subtending a visual angle of 2.5 min of arc (Snellen equivalent of 20/50). The subtended horizontal visual angle of each line was small enough so that most subjects would make up to 3 horizontal saccades for reading each line. This ensures that the horizontal saccadic eye movement during reading would not interfere with continuous AR recording. Subjects pressed a button when they reached the end of the paragraph to present the next paragraph. Immediately after the completion of the 10 min reading task, post-task TA1 was re-assessed continuously for 2 min by completely darkening the room and covering the subjects’ head and shoulders with a dark cloth. This maneuver usually took less than 5 s. Subjects were instructed to keep looking straight ahead to maintain stable fixation. The last 90 s of the recording were averaged for post-task TA1. After the post-task TA1 recording, while remaining in total darkness, participants were asked three yes/no questions about the story to ensure that attention was maintained during reading. The subjects were dark adapted for another 5 min and Baseline TA2 (binocular) was recorded continuously for 2 min. Afterward, 10 min of 3 diopter reading experiment, with the target located at 33.3 cm distance, was repeated under binocular viewing condition (Binocular Reading AR), and post-task TA2 was recorded with the same procedure. This sequence of events is shown in the following graph (Fig. 1). TA adaptation induced by either the monocular or binocular reading task was calculated by subtracting subjects’ baseline TA from their post-task TA respectively. A positive value for TA adaptation indicates an increase in TA level after 10 min of continuous reading at the 3 D distance, thus a myopic shift. Conversely, a negative value for TA adaptation indicates a decrease in TA level, thus a hyperopic shift. The average of the residual uncorrected refractive error was subtracted from TA and AR to adjust accommodation measurements in the data analysis.
27
2.3. Data analysis Due to AR recording errors, data was missing for seven participants in some of the test conditions. Data were analyzed using all participants, resulting in differences in the sample size for conditions where data were missing (see Table 1). Data were analyzed by linear regression correlations and t-test comparisons made using Prism 6.0 (GraphPad Software, San Diego, CA). 3. Results 3.1. Descriptive statistics Table 1 provides descriptive statistics for different aspects of accommodative function. The baseline TA1 frequency distribution is shown in Fig. 2. Similar to previous studies, (Leibowitz & Owens, 1978; McBrien & Millodot, 1987) TA scores had an asymmetrical positive distribution (Skewness = 0.60). The mean TA of 1.84D is larger than previous studies (McBrien & Millodot, 1987; Strang et al., 2000) that used similar recording methods and in which data were corrected for distance refractive error (t(41) = 5.34, p < 0.0001, 95% CI[1.44, 2.24]). 3.2. Relationship between baseline TA and AR to static target Regression analysis between Baseline TA (independent variable) and objective measurement of AR during different tasks (dependent variable) are presented in Fig. 3 and Table 2. Fig. 3 shows scatterplots for accommodation responses (AR) to static targets and monocular baseline TA (TA1). Linear regression analysis shows that Baseline TA1 accounted for significant amounts (32–41%) of variance in monocular AR to static targets at every viewing distance. Accommodation errors for those with a higher TA level tended to lead the target (above the solid stimulus demand line), but those with a lower TA tended to show accommodative lags (below the solid stimulus demand line). Linear regression equations for different static target distances are presented in Table 2. Slopes of the regression equation suggest that each diopter of Baseline TA1 is associated with about onethird of a diopter change to the AR. 3.3. Relationship between TA and AR during continuous reading task One might argue that accommodative responses to static targets may not accurately represent subjects’ responses during natural reading tasks due to the factors such as cognitive demand and saccadic eye movement changes. To test this hypothesis, we compared the AR for the 3D static target with the AR during reading. Regression analysis between Baseline TA (predictor variable) and AR during reading tasks (predicted variable) were performed and the results are presented in Fig. 4 and Table 2. Fig. 4 shows a side-by-side comparison of the scatterplot for Baseline TA and AR to continuous reading task at 3 D, both monocular (Fig. 4B) and binocular (Fig. 4C), with the scatterplot for AR to 3 D static targets
Fig. 1. Schematic representation of experimental task sequence.
28
C. Liu et al. / Vision Research 129 (2016) 25–32
Table 1 Descriptive statistics for different AR measures described in Fig. 1. Measure
N
Mean
Median
SD
Min
Max
Baseline TA1 (D) Post-task TA1 (D) Monocular Reading AR (D) Monocular TA Adaptation (D) Baseline TA2 (D) Post-task TA2 (D) Binocular Reading AR (D) Binocular TA Adaptation (D)
42 41 40 40 41 39 40 38
1.84 2.05 2.27 0.23 1.79 1.70 2.39 0.15
1.38 1.72 2.25 0.17 1.64 1.47 2.33 0.04
1.29 1.13 0.54 0.87 1.14 0.99 0.53 0.71
0.43 0.06 1.49 1.63 0.19 0.22 1.65 2.60
5.14 4.73 3.87 2.68 4.86 4.22 3.64 1.47
myopic or hyperopic shift (Fig. 5). Comparing group means alone might mask the individual effect. Therefore, we calculated linear regressions between reading AR (predictor variable) and the individual TA adaptation (predicted variable) for monocular and binocular reading tasks, respectively. Data are presented in Table 3. Results suggested that near work-induced TA adaptation was not significantly correlated with AR error. These results support the conclusion that only the tonic accommodation level, not near work-induced TA adaptation, was associated with the accommodative response during the near visual task. 4. Discussion
Fig. 2. Frequency distribution of baseline tonic accommodation (TA1).
and Baseline TA1 from Fig. 2 (Fig. 4A). The linear regression equations for 3 D continuous reading task are presented in Table 2. Similar to the static target at the same demand level, Baseline TA1 accounted for significant amounts (45%) of the variance in AR under the monocular reading task (Fig. 4B). The correlation between Baseline TA2 and AR during reading is weakened under the binocular reading condition, accounting for a smaller amount (24.3%) of variance in AR (Fig. 4C). AR during binocular reading is slightly higher than that during monocular reading (Table 1), although a paired t-test showed that the difference was not statistically significant (t = 1.627; p = 0.112). The slopes of these regression plots suggest that every diopter of baseline TA is associated with about one-quarter to one-third of a diopter to the AR during natural reading conditions. Together, results from Table 2, Fig. 3 and Fig. 4 suggest that Baseline TA is significantly correlated with closed-loop accommodative responses at different viewing distances and different accommodative task. 3.4. Relationship between near work-induced TA adaptation and Closed-loop AR during near visual tasks To study the relationship between near work-induced TA adaptation and AR, we examined if ten minutes of reading at a 3 D distance had induced a significant change in TA level. For monocular reading task, Post-task TA1 showed only a small increase (myopic shift) of 0.23 D compared to Baseline TA1. A paired t-test showed that Post-task TA1 was not significantly different from Baseline TA1 (t = 1.650; p = 0.107). For a binocular reading task, Post-task TA2 showed a small decrease (hyperopic shift) of 0.15 D comparing to Baseline TA2. Similar to monocular reading, a paired t-test showed that Post-task TA2 was not significantly different from Baseline TA2 (t = 1.268; p = 0.213). Although group means of TA adaptation showed no significant changes after monocular or binocular reading tasks, individual subjects exhibited different adaptation patterns, showing either
This study found a strong association between open-loop TA and closed-loop AR. Every diopter of Baseline TA1 was associated with approximately one-third of a diopter of steady-state accommodative response during a monocular accommodative task. This ratio was consistent across a variety of viewing distances and different near work tasks. These correlations do not prove that tonic accommodation directly contributes to the steady state accommodation response. Nevertheless, they are consistent with qualitative and quantitative models in which the individual tonic accommodation level is the starting point (or home) from which various stimuli such as defocus, proximity, voluntary effort, and vergence accommodation operate to provide an accommodation response closer to the stimulus (Charman, 1986; Heath, 1956; Hung, 1992; Hung, Ciuffreda, & Rosenfield, 1996). Importantly, these current findings also demonstrate that the steady-state accommodation response is not as fully independent of the tonic accommodation starting point as is commonly believed. Our findings suggest that the assumptions made for the classic model of steady-state accommodation developed by Hung and Semmlow (1980) might need to be re-examined. We argue for the adoption of individual rather than group mean values for estimating model parameters. Hung and Semmlow’s model parameters were built on data obtained from only four subjects and two were over 30 years-old (age ranging from 18 to 37, TA level ranging from 0.42 D to 2.2 D). The TA or ABIAS (the term used in this model) was incorporated into the steady state AR calculation model as well as calculation for Accommodative Controller Gain for Positive Accommodative Error (ACGP) (see the following equations cited from the Hung & Semmlow, 1980 study. AS represents accommodative stimulus; DSP represents deadspace in accommodation, also known as ‘‘depth-of-focus”).
ACGP ACGP DSP 1 þ ACGP 1 þ ACGP 1 þ ABIAS 1 þ ACGP
AR ¼ AS
ACG ¼
AR ABIAS AE DSP
ðFormula 1Þ ðFormula 2Þ
C. Liu et al. / Vision Research 129 (2016) 25–32
29
Fig. 3. Scatterplot of the effect of TA1 on accommodation responses while viewing static targets at different viewing distances. Linear regression lines are plotted as dotted lines; the target’s accommodative stimulus demand is plotted as a solid, horizontal line.
Based on Subject JS’s data, who was 37 year-old and had a low ABIAS (or TA) of 0.42 D and high ACGP of 8.53, Hung and Semmlow concluded that ABIAS contributed minimally to AR, only accounted for 1.8% of the AR for a 3 D accommodative stimulus.
Rosenfield et al.’s (1993) review assessed the role of TA in closed-loop accommodation by using Hung’s model. Based on the work of Mordi and Ciuffreda (1998), Rosenfield et al. (1993) assumed that the ACG value was 10 and concluded that TA only
30
C. Liu et al. / Vision Research 129 (2016) 25–32
Table 2 The relationship between baseline TA (x variable) and AR (y variable) at different viewing conditions. Target distance
Linear regression equation
95% CI for slope
p-value for nonzero slope
0.1 D Static* 2 D Static* 3 D Static* 4 D Static* 5 D Static* 3 D Monocular reading* 3 D Binocular readingy
Y = 0.296*X 0.10 Y = 0.335*X + 1.02 Y = 0.355*X + 1.87 Y = 0.408*X + 2.69 Y = 0.425*X + 3.57 Y = 0.277*X + 1.77
0.18, 0.20, 0.19, 0.21, 0.25, 0.17,
0.41 0.47 0.53 0.60 0.60 0.38
<0.005 <0.005 <0.005 <0.005 <0.005 <0.005
Y = 0.229*X + 1.98
0.10, 0.36
<0.005
* Monocular tasks, Baseline TA1 was used as independent variable (x) in regression analysis. y Binocular task, Baseline TA2 was used as independent variable (x) in regression analysis.
contributed less than 0.1 D in AR to a 5 D stimulus. Using a similar approach, Hung assessed the impact of ABIAS (TA) on the accommodative stimulus–response function by varying ABIAS (TA) values while holding ACG at a constant value of 10 in their computer model. (Hung, 1998) Computer simulation revealed that when holding ACG at constant value of 10, varying ABIAS (TA) had minimum impact on the slope of accommodative stimulus–response function. In contrast, when holding ABIAS (TA) at constant value of 1 D, varying the ACG value (from 1 to 20) had a significant effect on the slope, with increased ACG value associated with a steeper slope. Based on this result, Hung et al. drew the same conclusion that TA had minimum influence on AR under normal closed-loop condition. These modeling results, however, did not explore the possible co-variation between TA and ACG. According to Formula 2 and the trend of empirical data from Table 1 in Hung and Semmlow (1980) paper, ABIAS (TA) inversely correlates to ACG so that when TA increases, ACG drops significantly. Thus, the higher the TA value, the larger the proportion of AR will be contributed from TA. Holding the ACG parameter constant in the simulation when empirical data shows that ACG varies individually might not be optimal for the modeling the dynamics of the accommodative system. For example, when subject JM’s data (ABIAS (TA) = 2.2 D, ACGP = 2.32) from Hung and Semmlow (1980) paper is used for model calculation, ABIAS (TA) contributed 33.4% to AR for the same 3 D accommodative stimulus. This subject’s ABIAS (TA) values lie
close to our group mean for TA (1.84 D), which might explain why these subjects have results that are similar to those reported in our study. In our study we adopted a unique approach, studying the effects of individual variability and co-variation between TA and closedloop AR. In addition, all subjects’ data in our study, including both those with normal or abnormal accommodative function, were included in the analysis, allowing us to examine this correlation under a wide spectrum of functioning level. As described above, most previous studies assumed that group means for these values applied to all individuals in the sample. Using the group mean for TA value presumes that all adopt an intermediate posture of 1.84 D, yet our subjects showed a very widely distributed TA level, ranging from 0.43 D to +5.14 D. The variability found in our experiment is consistent with other studies that assessed TA using different methodologies (Gilmartin et al., 1984; Heron et al., 1981; Leibowitz & Owens, 1975, 1978; Maddock et al., 1981). Similarly, a wide distribution of ACG values has been previously reported for normal adult subjects, ranging from 1.5 to 21.0 with mean value around 10 (Ciuffreda, Hokoda, Hung, & Semmlow, 1984; Mordi & Ciuffreda, 1998; Ong, Ciuffreda, & Tannen, 1993). Although we did not calculate the ACG value in the current study, one can assume a similar distribution pattern of ACG in our sample. Models using a fixed value of 10 for the ACG and 1 D for TA may mask the true correlation between TA level and AR. Consistent with Miller (1980), our results suggested that the correlation between TA and closed-loop AR is more prominent under monocular viewing conditions comparing to binocular viewing conditions. Both the R-square value and the regression equation slope in regression analysis are lower for binocular reading conditions. This finding confirmed that the vergence system would also have a significant impact on closed-loop AR through the CA pathway which would provide additional stimulus to the accommodative system under binocular viewing conditions. Previous studies have found that the accommodative lag at near is greater when viewing under monocular viewing conditions (Jiang, Tea, & O’Donnell, 2007; Nakatsuka, Hasebe, Nonaka, & Ohtsuki, 2003). Although not statistically significant, our data did show a trend of a slight increase in AR during binocular reading compared to monocular reading, consistent with the previous literature. It is possible that some individual may utilize CA to compensate for weak accommodative system. If this indeed is the case, one might hypothesize that the compensatory effect would only be possible
Fig. 4. Scatterplot of the effect of TA1 on the accommodation response while viewing: (A) a monocular 3D static target and (B) continuous monocular reading at 3D. The effect of Baseline TA2 on the accommodation response while performing continuous binocular reading at 3D is shown in Fig. 4C. Linear regression lines are plotted as dotted lines; the target is plotted as a solid, horizontal line.
C. Liu et al. / Vision Research 129 (2016) 25–32
31
responses observed in our group. Further study is necessary to clarify this issue. To our knowledge, this is the first study to examine the relationship between TA and the steady state AR, using an objective instrument to continuously monitor the accommodation responses during both sustained reading task and to static targets at different viewing distances. Our results suggest that the steady-state accommodation response is not as independent of individual TA as is commonly thought. A significant correlation is found between the two variables. These findings encourage a re-examination of the classic model of steady-state accommodation. Disclosure C. Liu, None; S. Drew, None; E. Borsting, None; A. Escobar, None; L. Stark, None; C. Chase, None. Acknowledgments This research project is supported by National Eye Institute grant (1R15EY021021 Chase, C. (PI)). References
Fig. 5. Frequency distribution of near work-induced TA adaptation: (A) monocular reading task, (B) binocular reading task.
Table 3 The relationship between near work induced-TA adaptation (x variable) and AR during reading (y variable). Task Monocular reading Binocular readingy
Linear regression equation *
*
Y = 0.077 X + 2.314 Y = 0.052*X + 2.404
R2
p-value
0.017 0.005
0.441 0.685
* Monocular reading, Monocular TA Adaptation was used as independent variable (x) in regression analysis; y Binocular reading task, Binocular TA Adaptation was used as independent variable (x) in regression analysis.
in individual who has strong vergence system. Further study is necessary to test this hypothesis. In contrast to some of the previous literature, our subject group did not show a significant near-work induced TA adaptation when comparing Post-task TA to Baseline TA, although large individual variability exists within group with both myopic and hyperopic TA adaptation. Further correlational analysis also failed to find any statistically significant relation between individual AR and TA adaptation. Instead, our data indicated that only the tonic TA level itself correlated to the accuracy of the AR. Our adapting task was similar to Rosenfield and Gilmartin (1999) who used a 3 D target and 10 min of near viewing. In the Rosenfeld et al. study a significant number of subjects did not adapt to the near task and in turn did not show a reduction in accommodative lag. It is possible that our sample of subjects could have included a significant number of non-adaptors since we recruited adults with a broad range of accommodative function. This may account for the lack of adaptive
Andre, J. T., & Owens, D. A. (1999). Predicting optimal accommodative performance from measures of the dark focus of accommodation. Human Factors, 41(1), 139–145. Bullimore, M. A. M., & Gilmartin, B. (1989). The measurement of adaptation of tonic accommodation under two open-loop conditions. Ophthalmic & Physiological Optics, 9(1), 72–75. http://dx.doi.org/10.1111/j.1475-1313.1989.tb00810.x. Bullimore, M. A., Gilmartin, B., & Hogan, R. E. (1986). Objective and subjective measurment of tonic accommodation. Ophthalmic & Physiological Optics, 6(1), 57–62. http://dx.doi.org/10.1111/j.1475-1313.1986.tb00700.x. Charman, W. N. (1986). Static accommodation and the minimum angle of resolution. American Journal of Optometry and Physiological Optics, 63, 915–921. Ciuffreda, K. J., Hokoda, S. C., Hung, G. K., & Semmlow, J. L. (1984). Accommodative stimulus/response function in human amblyopia. Documenta Ophthalmologica, 56(4), 303–326. Gilmartin, B., Hogan, R. E., & Thompson, S. M. (1984). The effect of Timolol Maleate on tonic accommodation, tonic vergence, and pupil diameter. Investigative Ophthalmology & Visual Science, 25(6), 763–770. Gray, L., Strang, N., Winfield, N., Gilmartin, B., & Winn, B. (1998). The magnitude and distribution of open-loop accommodation using three different methods of opening the loop. Optometry and Vision Science, 75(12), 897–902. Hasebe, S., Nonaka, F., & Ohtsuki, H. (2005). Accuracy of accommodation in heterophoric patients: Testing an interaction model in a large clinical sample. Ophthalmic & Physiological Optics, 25, 582–591. Hazel, C. A., Cox, M. J., & Strange, N. C. (2003). Wavefront aberration and its relationship to the accommodative stimulus-response function in myopic subjects. Optometry and Vision Science, 80(2), 151–158. Heath, G. G. (1956). Components of accommodation. American Journal of Optometry and Archives of American Academy of Optometry, 33, 569–579. Heron, G., Smith, A. C., & Winn, B. (1981). The influence of method on the stability of dark focus position of accommodation. Ophthalmic & Physiological Optics, 1(2), 79–90. http://dx.doi.org/10.1111/j.1475-1313.1981.tb00151.x. Hung, G. K. (1992). Adaptation model of accommodation and vergence. Ophthalmic & Physiological Optics, 12, 319–326. Hung, G. K. (1998). Sensitivity analysis of the stimulus-response function of a static nonlinear accommodation model. IEEE Transactions on Bio-medical Engineering, 45(3), 335–341. http://dx.doi.org/10.1109/10.661158. Hung, G. K., Ciuffreda, K. J., & Rosenfield, M. (1996). Proximal contribution to a linear static model of accommodation and vergence. Ophthalmic & Physiological Optics, 16, 31–41. Hung, G. K., & Semmlow, J. L. (1980). Static behavior of accommodation and vergence: Computer simulation of an interactive dual-feedback system. IEEE Transactions on Bio-medical Engineering, 27(8), 439–447. http://dx.doi.org/ 10.1109/TBME.1980.326752. Jiang, B.-C., Tea, Y. C., & O’Donnell, D. (2007). Changes in accommodative and vergence responses when viewing through near addition lenses. Optometry, 78 (3), 129–134. http://dx.doi.org/10.1016/j.optm.2006.08.017. Leibowitz, H. W., & Owens, D. A. D. (1975). Anomalous myopias and the intermediate dark focus of accommodation. Science, 189(4203), 646–648. Leibowitz, H. W. H., & Owens, D. A. D. (1978). New evidence for the intermediate position of relaxed accommodation. Documenta Ophthalmologica, 46(1), 133–147. Liu, C., Chase, C., Drew, S., & Castellanos, E. (2013). Tonic accommodation correlates with accommodative amplitude-scaled facility test for symptomatic graduate students. Investigative Ophthalmology & Visual Science, 54 (E-abstract 4259).
32
C. Liu et al. / Vision Research 129 (2016) 25–32
Maddock, R. J. R., Millodot, M. M., Leat, S. S., & Johnson, C. A. C. (1981). Accommodation responses and refractive error. Investigative Ophthalmology & Visual Science, 20(3), 387–391. Mallen, E. A. H., Wolffsohn, J. S., Gilmartin, B., & Tsujimura, S. (2001). Clinical evaluation of the Shin-Nippon SRW-5000 autorefractor in adults. Ophthalmic & Physiological Optics, 21, 101–107. McBrien, N. A., & Millodot, M. (1987). The relationship between tonic accommodation and refractive error. Investigative Ophthalmology & Visual Science, 28(6), 997–1004. Miller, R. J. (1980). Ocular vergence-induced accommodation and its relation to dark focus. Perception & Psychophysics, 28(2), 125–132. http://dx.doi.org/10.3758/ BF03204337. Mordi, J. A., & Ciuffreda, K. J. (1998). Static aspects of accommodation: Age and presbyopia. Vision Research, 38(11), 1643–1653. http://dx.doi.org/10.1016/ S0042-6989(97)00336-2. Nakatsuka, C., Hasebe, S., Nonaka, F., & Ohtsuki, H. (2003). Accommodative lag under habitual seeing conditions: Comparison between adult myopes and emmetropes. Japanese Journal of Ophthalmology, 47(3), 291–298. Ong, E., Ciuffreda, K. J., & Tannen, B. (1993). Static accommodation in congenital nystagmus. Investigative Ophthalmology & Visual Science, 34(1), 194–204. Owens, D. A., & Wolf-Kelly, K. (1987). Near work, visual fatigue, and variations of oculomotor tonus. Investigative Ophthalmology & Visual Science, 28(4), 743–749. Rosenfield, M., Ciuffreda, K. J., & Gilmartin, B. (1992). Factors influencing accommodative adaptation. Optometry and Vision Science, 69(4), 270–275. Rosenfield, M., Ciuffreda, K. J., Hung, G. K., & Gilmartin, B. (1993). Tonic accommodation: A review. I. Basic aspects. Ophthalmic & Physiological Optics, 13(3), 266–284.
Rosenfield, M., & Gilmartin, B. (1999). Accommodative error, adaptation and myopia. Ophthalmic & Physiological Optics, 19(2), 159–164. Schor, C. M., Kotulak, J. C., & Tsuetaki, T. (1986). Adaptation of tonic accommodation reduces accommodative lag and is masked in darkness. Investigative Ophthalmology & Visual Science, 27(5), 820–827. Strang, N., Gilmartin, B. S. B., Gray, L., Winfield, N. R. N., & Winn, B. (2000). Openloop accommodation in emmetropia and myopia. Current Eye Research, 20(3), 190–194. Toates, F. M. (1972). Accommodation function of the human eye. Physiological Reviews, 52(4), 828–863. Tosha, C., Borsting, E., Ridder, W. H., & Chase, C. (2009). Accommodation response and visual discomfort. Ophthalmic & Physiological Optics, 29(6), 625–633. http:// dx.doi.org/10.1111/j.1475-1313.2009.00687.x. Wang, B., & Ciuffreda, K. J. (2006). Depth-of-focus of the human eye: Theory and clinical implications. Survey of Ophthalmology, 51(1), 75–85. Win-Hall, D. M., & Glasser, A. (2009). Objective accommodation measurements in pseudophakic subjects using an autorefractor and an aberrometer. Journal of Cataract and Refractive Surgery, 35(2), 282–290. http://dx.doi.org/10.1016/j. jcrs.2008.10.033. Win-Hall, D. M., Houser, J., & Glasser, A. (2010). Static and dynamic accommodation measured using the WAM-5500 autorefractor. Optometry and Vision Science, 87 (11), 873–882. http://dx.doi.org/10.1097/OPX.0b013e3181f6f98f. Winn, B., Gilmartin, B., Mortimer, L. C., & Edwards, N. R. (1991). The effect of mental effort on open- and closed-loop accommodation. Ophthalmic & Physiological Optics, 11(4), 335–339.