PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Invest Ophthalmol Vis Sci. Author manuscript; available in PMC 2010 July 1.
Published in final edited form as:
PMCID: PMC2714044
NIHMSID: NIHMS128727

Short-Term Adaptive Modification of Dynamic Ocular Accommodation

Abstract

Purpose

Indirect observations suggest that the neural control of accommodation may undergo adaptive recalibration in response to age-related biomechanical changes in the accommodative system. However, there has been no direct demonstration of such an adaptive capability. This investigation was conducted to demonstrate short-term adaptation of accommodative step response dynamics to optically induced changes in neuromuscular demands.

Methods

Repetitive changes in accommodative effort were induced in 15 subjects (18–34 years) with a double-step adaptation paradigm wherein an initial 2-D step change in blur was followed 350 ms later by either a 2-D step increase in blur (increasing-step paradigm) or a 1.75-D step decrease in blur (decreasing-step paradigm). Peak velocity, peak acceleration, and latency of 2-D single-step test responses were assessed before and after 1.5 hours of training with these paradigms.

Results

Peak velocity and peak acceleration of 2-D step responses increased after adaptation to the increasing-step paradigm (9/12 subjects), and they decreased after adaptation to the decreasing-step paradigm (4/9 subjects). Adaptive changes in peak velocity and peak acceleration generalized to responses that were smaller (1 D) and larger (3 D) than the 2-D adaptation stimulus. The magnitude of adaptation correlated poorly with the subject's age, but it was significantly negatively correlated with the preadaptation dynamics. Response latency decreased after adaptation, irrespective of the direction of adaptation.

Conclusions

Short-term adaptive changes in accommodative step response dynamics could be induced, at least in some of our subjects between 18 and 34 years, with a directional bias toward increasing rather than decreasing the dynamics.

Recalibration of neural signals allows a biological control system to restore and maintain optimal motor performance in the event of any age-, pathology-, or environment-related changes in the biomechanical plant. Ocular accommodation of humans and nonhuman primates is a good example of a biological control system whose biomechanical plant (comprising the crystalline lens, ciliary muscle, choroid, and zonules) becomes increasingly noncompliant with age, ultimately leading to a complete loss of focusing ability (absolute presbyopia).1 The loss of focusing ability can be primarily attributed to a decrease in lenticular compliance (compliance = 1/elasticity),24 increase in lenticular viscoelasticity,4 mechanical restriction in ciliary muscle movement,5,6 and morphologic changes in the accommodative plant.79 During the incipient stages of presbyopia, the increasingly noncompliant accommodative plant increases the neuromuscular demands required to generate the desired amount of accommodation.10

Several indirect observations suggest that the neural control of accommodation may undergo adaptive recalibration to suit the biomechanical changes in the accommodative plant. For instance, neural recalibration to age-related loss of plant compliance is suggested by an increase in the myodiopter (the neuromuscular effort required to generate a diopter of accommodation11) during incipient presbyopia.10,12 An increase in the myodiopter is indicated by an age-related increase in the response AC/A ratio (the amount of vergence generated by a diopter of accommodation) and an age-related decline in the response CA/C ratio (the amount of accommodation generated by a meter angle of vergence), within the linear range of the accommodative stimulus–response function.10,13,14 Neural recalibration to an age-related increase in lenticular viscoelasticity is suggested by the lack of a dramatic reduction in the peak velocity of small accommodative step responses (1–3 D).1517 On a more general level, neural recalibration of disaccommodation (near-to-far accommodation) is suggested by the serendipitous observation of Bharadwaj and Schor18 that repeated pilot experiments over 1 year reduced the speed and acceleration of small disaccommodative step responses of the first author, minimizing the instabilities that are usually associated with such responses. Neural recalibration may have occurred to optimize stability by slowing down the responses.18 Finally, adaptable dynamics are also suggested by the success of orthoptic training exercises (e.g., accommodative flippers and pencil push-ups) that decrease the time constant, speed up abnormally sluggish accommodative responses, and increase the rate at which accommodation responds to repetitive positive and negative focusing errors (accommodative flipper rates).19,20

Apart from these indirect observations, to our knowledge, there has been no direct demonstration of the adaptation capability of accommodative step response dynamics. The main purpose of this study was to provide direct evidence for such an adaptive capacity by employing a double-step adaptation paradigm that optically stimulates increases (increasing-step paradigm) or decreases (decreasing-step paradigm) in neuromuscular efforts. Our results illustrate that the dynamics of accommodative step responses can be modified on a short-term basis, at least in some subjects ranging in age from 18 to 34 years, with increases in neuromuscular effort adapting more readily than decreases in neuromuscular effort.

Methods

Fifteen naïve observers (except subject SRB) between 18 and 34 years (11 emmetropic subjects and 4 fully corrected myopic subjects) participated after signing an informed consent form approved by the Committee for Protection of Human Subjects, University of California at Berkeley. The study also adhered to the tenets of the Declaration of Helsinki. Fourteen and 12 subjects participated in the increasing-step and decreasing-step paradigms, respectively. Eight subjects participated in both paradigms.

Subjects, positioned on a bite bar and forehead rest, focused on a high-contrast, black-and-white Maltese cross. Their left eye's accommodative responses were measured (the right eye was occluded) with an infrared recording optometer (200 Hz sampling frequency; Generation V SRI International, Menlo Park, CA).21 The left pupil of each subject remained fully dilated throughout the experiment with phenylephrine hydrochloride. The concentration (2.5%) was much weaker than that shown to influence accommodative dynamics (10%).22 Step changes in optical blur were stimulated using a Badal stimulus optometer (SRI International).21 Accommodative responses were calibrated thrice on each subject by stimulating 0 to 4 D of accommodation and disaccommodation in 1-D steps and the averaged output voltages were linearly regressed against the accommodative stimuli to convert the optometer voltages into dioptric units.18,23 For older subjects, only 1-to 3-D stimuli were used for calibration. Because of the typically observed lag of accommodation,24 the calibration slope is artificially steepened when the voltage signals are regressed against the accommodative stimuli. Steepening of the calibration slope, however, would concurrently amplify both the accommodative magnitude and dynamics for a given stimuli, without altering the relative relation between amplitude and dynamics. Furthermore, since the calibration slope of a given subject did not typically change during the experiment, accommodative response characteristics measured before and after training remained unaffected relative to each other.

Changes in accommodative step response dynamics (peak velocity and peak acceleration) were induced with a double-step adaptation paradigm, similar to those used to induce vergence25,26 and saccade27,28 adaptation (Fig. 1). Each experimental session, lasting approximately 1.5 to 2 hours, consisted of four sessions (preadaptation session, adaptation session, sustained session, and postadaptation session). Each trial in a session lasted for 4 seconds, and the inter-trial interval was 2.5 seconds (Fig. 1). Limited periods of rest were given to subjects, during which time they were asked to keep their eyes closed to avoid normal visual stimulation.

Figure 1
Experimental design for the increasing- and decreasing-step adaptation paradigms.

In the preadaptation session (first phase of the experiment) and the postadaptation session (last phase in the experiment), consisting of 100 and 50 trials, respectively, single-step stimuli ranging from 1 to 4 D (test stimuli) were presented in steps of 0.5 D in pseudorandomized order (Fig. 1). In the two oldest subjects (31 and 34 years of age), the stimuli ranged from 1 to 3 D in 0.5-D steps. These stimulus amplitudes typically fell within 75% of the maximum age-matched objective and subjective amplitude of accommodation described for humans.1,29,30 To minimize the ability of subjects to anticipate the start of the trial, the stimulus onset was randomized by a time delay (range, 0–1000 ms) after the subject initiated the trial with a button press.

Two different adaptation sessions (second phase of the experiment), each consisting of 150 trials, were run on separate days. In the adaptation trials of the increasing-step paradigm, an initial 2-D step change in blur was followed 350 ms later by a 2-D step increase in blur (100% increment; Fig. 1). In the adaptation trials of the decreasing-step paradigm, an initial 2-D step change in blur was followed 350 ms later by a 1.75-D step decrease in blur (87.5% decrement; Fig. 1). The second-step decrement in blur was less than 100%, to ensure that the stimulus was not perceived as a transient change in blur. In four subjects, adaptation to a larger amplitude stimulus decrement was also investigated, wherein an initial 4-D step change in blur was followed 350 ms later by a 2-D step decrement in blur (not shown in Fig. 1). The results for the large-step decrement were similar to those of the small-step decrement, and hence they are not presented separately. The sustained session of the experiment (third phase of the experiment), consisting of 150 trials, was similar to the adaptation session, but 2-D single-step test trials were interleaved with the double-step trials with a ratio of five double-step trials to one single-step trial to assess the changes in response characteristics induced during adaptation by the double-step trials (Fig. 1). Each trial in the adaptation and sustain sessions of the experiment was initiated 2.5 seconds after the end of the previous trial (Fig. 1). To minimize the ability of subjects to anticipate the start of the trial, the stimulus onset was randomized by a time delay (ranging between 0–1000 ms) after the 2.5 seconds elapsed.

The second step change in the double-step stimulus was always presented at a constant delay of 350 ms, which corresponded approximately to the end of the latency period of first accommodative step response. The second-step change is likely to induce an unanticipated focus error at the end of the latency period of the first accommodative step response. This defocus error is corrected by a change in the accommodative neuromuscular effort that is guided by blur feedback. An increased neuromuscular effort is required by a double step that increases the blur during the initial accommodative response (increasing-step paradigm) and a decreased neuromuscular effort is required by a double step that decreases the blur during the initial accommodative response (decreasing-step paradigm). Repeated presentation of the double-step condition is intended to stimulate a recalibration of the relationship between the initial amount of defocus and the initial neuromuscular effort needed to correct the focus error. This recalibration may be reflected as changes in the velocity and acceleration of single-step responses as observed for double-step adaptation of disparity vergence25,26 and saccades.27,28 Ideally, the timing of the second-step stimulus would be controlled to vary in each trial with the online estimates of response latency. However, it was not possible to obtain reliable online estimates of response latency because of the relatively large fluctuations in the steady state accommodation.31 The second step was therefore always presented 350 ms after the first step (Fig. 1).

The increasing- and decreasing-step paradigms were conducted within 2 weeks of each other in eight subjects who took part in both training paradigms. Since response dynamics that are collected within a few weeks of each other were found not to change significantly (Bharadwaj SR, unpublished observations, 2003–2005), only one baseline (preadapted) session was conducted in these subjects, typically on the day that the increasing-step paradigm was conducted. This strategy was used to minimize fatigue and the total time required of the subject to complete the experiment. Separate adaptation, sustaining, and postadaptation sessions were conducted (in that order) for the increasing- and decreasing-step paradigms on different days, and the postadaptation data from each paradigm was compared to the preadaptation data collected a few days earlier.

Data were analyzed using three commercial statistical packages (Excel; Microsoft, Redmond, WA, MatLab; The MathWorks, Natick, MA, and SPSS, SPSS, Chicago, IL). Accommodative position traces were differentiated and smoothed with a 100-ms running-average window to obtain velocity and acceleration profiles.23 The start and end of the accommodative response was identified with a velocity–criterion algorithm described in detail elsewhere.23,32 In the increasing-step paradigm, the postadaptation responses of some subjects showed instabilities and in such cases, the end of the response was confirmed through visual inspection. Response amplitude was calculated as the dioptric difference between the averaged (100 ms) accommodative states before the start of the response and after the end of the response. Single-step accommodative responses with very short latencies (<100 ms) and responses that changed nonmonotonically (~10% to 15% of the responses in each subject) were discarded from the analyses because reliable estimates of response parameters could not be obtained from these traces.

Adaptive changes in response dynamics were quantified by comparing the peak velocity, peak acceleration, and response latency before and after double-step training (Fig. 2). In the pre- and postadaptation sessions, the peak velocity and peak acceleration were grouped for each subject into 1-D bins, and all the responses within the bin were averaged. Since the response latency did not show any significant variation across response amplitude, all the single-step response latencies within an experimental session were averaged, irrespective of the amplitude of the response. The peak velocity, peak acceleration and latency of individual responses within the preadaptation and postadaptation sessions did not show any systematic trial dependent variations and therefore it is unlikely that the averaging obscured any obvious trends in the data. The pre- and postadaptation dynamics of each subject were statistically compared on an individual basis by using a two-factor repeated-measures ANOVA test. Statistical significance (defined here as P < 0.01) of factor 1 (adaptation magnitude) would indicate that, when the data across all response amplitudes are considered together, the preadaptation and postadaptation peak velocity and peak acceleration were significantly different from each other. Statistical significance of factor 2 (response amplitude) indicates that, when all the preadaptation and postadaptation data were considered together, the peak velocity and peak acceleration changed significantly across response amplitude. Statistically significant interaction between factors 1 and 2 indicates that the change in peak velocity and peak acceleration after adaptation was not uniform across all response amplitudes. For data sets that violated the assumption of sphericity (i.e., the difference scores between various levels of the repeated-measures factor had unequal variances; 2 of the 12 subjects in the increasing-step paradigm and 3 of the 9 subjects in the decreasing-step paradigm), appropriate corrections to the F-value (Greenhouse-Geisser correction) was applied before computing statistical significance. The pre- and postadaptation response latencies were compared by using a Student's t-test.

Figure 2
Raw traces of accommodative position (top), velocity (middle), and acceleration (bottom) for a 2-D step stimulus from three representative individuals in the increasing-step paradigm. The position traces are averages of five temporally aligned individual ...

Results

Increasing-Step Paradigm

Usable data were collected from 12 subjects. Data from the two remaining subjects were discarded because most of their responses had extremely short latencies (<100 ms). Figure 2 shows representative raw traces from three subjects who gave usable data in the increasing-step paradigm. The static and dynamic (peak velocity and peak acceleration) characteristics of preadaptation single-step responses were comparable to those described in previous experiments23,33 (Fig. 2). The increase in dynamics after adaptation (9/12 subjects) were associated with response instabilities in three subjects (e.g., subject DK in Fig. 2) but not in the others (e.g., subject KC in Fig. 2). In most subjects, the latency of the single-step responses decreased after double-step stimulation, irrespective of whether the peak velocity and peak acceleration showed adaptive changes (Fig. 2, Table 1).

Table 1
Response Latency of Accommodation before and after Training in the Two Paradigms

The preadaptation peak velocity increased significantly with response amplitude in six subjects (ANOVA factor 2, P < 0.001; Figs. 3b, 3c, 3e, 3h–j), saturated at larger response amplitudes in three subjects (ANOVA factor 2, P < 0.01; Figs. 3a, 3d, 3f) and did not change with response amplitude in the remaining three subjects (ANOVA factor 2, P > 0.01; Figs. 3g, 3k, 3l). Unlike the preadaptation peak velocity, the preadaptation peak acceleration did not change significantly with response amplitude in the majority of the subjects (ANOVA factor 2, P = 0.03−0.94 across subjects; Fig. 4). Similar to earlier studies,23,34 the standard deviations of preadaptation peak acceleration were larger than those of preadaptation peak velocity (Figs. 3, ,4).4). However, when the standard deviations of peak velocity and peak acceleration were normalized to their respective means (coefficient of variation), the statistical variability was similar for both parameters.

Figure 3
(al) Mean (±SD) peak velocity of accommodation as a function of response amplitude for the preadaptation and postadaptation sessions for each subject in the increasing-step paradigm. (k) Repeatability data from subject PRM. The numbers ...
Figure 4
(al) Mean (±1 SD) peak acceleration of accommodation as a function of response amplitude for the preadaptation and postadaptation sessions for each subject in the increasing-step paradigm. (k) Repeatability data from subject PRM. All ...

The postadaptation peak velocity was significantly larger than the preadaptation peak velocity across different response amplitudes in nine subjects, irrespective of whether the preadaptation peak velocity increased or saturated as a function of response amplitude (ANOVA factor 1, P < 0.001; Fig. 3). The pre- and postadaptation peak velocities were similar to each other in two subjects (ANOVA factor 1, P = 0.71; Figs. 3c, 3i) and the postadaptation peak velocity was significantly lower than the preadaptation peak velocity in the remaining one subject (ANOVA factor 1, P < 0.001; Fig. 3d). Six of 12 subjects showed a statistically significant increase in peak acceleration with adaptation (ANOVA factor 1, P < 0.01; Fig. 4), whereas the remaining showed either no change (ANOVA factor 1, P > 0.01; Figs. 4c, 4e, 4g, 4h) or a decrease (ANOVA factor 1, P < 0.01; Figs. 4d, 4i) in peak acceleration with adaptation. For peak velocity and peak acceleration, the interaction between adaptation magnitude and response amplitude was not statistically significant in most subjects, confirming that the change in peak velocity and peak acceleration with adaptation was similar across different response amplitudes (Figs. 3, ,44).

In those subjects who showed statistically significant changes with adaptation, both peak velocity and peak acceleration typically changed simultaneously with adaptation (e.g., compare data from subjects DK and EH in Figs. 3, ,4),4), with the average magnitude of change being similar for both parameters (Figs. 3, ,4).4). Although the change in peak velocity and peak acceleration appeared to spread across the three different response amplitudes, the magnitude of adaptation was definitely variable across these response amplitudes and across different subjects.

Decreasing-Step Paradigm

Usable data were collected from nine subjects. Data from the three remaining subjects were discarded because most of their responses had extremely short response latencies (<100 ms). In general, responses in the adaptation and postadaptation sessions of the decreasing-step paradigm were more variable than those in the increasing-step paradigm. The changes in peak velocity and peak acceleration in the postadaptation session were less prevalent in the decreasing-step paradigm when compared with the increasing-step paradigm. The peak velocity decreased after decreasing-step adaptation in four of the nine subjects (ANOVA factor 1, P < 0.001; raw trace from subject JW in Figs. 5, 6b, 6c, 6e, 6g). Similarly, the peak acceleration decreased significantly after adaptation in three of the nine subjects (ANOVA factor 1, P < 0.01; Figs. 7b, 7c, 7g). In the remaining subjects, the postadaptation peak velocity and peak acceleration either remained the same (ANOVA factor 1, P > 0.01) or increased (ANOVA factor 1, P < 0.01) after adaptation (raw traces from subjects DK and KC in Figs. 5, ,6,6, 7a, 7d, 7i). The interaction between adaptation magnitude and response amplitude was not statistically significant for both peak velocity and peak acceleration in most subjects indicating that the change in dynamics after adaptation was generalizable across response amplitudes (P = 0.20 − 0.96 across subjects; Figs. 6, ,7).7). The response latency reduced with adaptation irrespective of whether the other dynamic parameters also changed (Fig. 5, Table 1).

Figure 5
Raw traces of accommodative position (top), velocity (middle), and acceleration (bottom) for a 2-D step stimulus from three representative individuals in the decreasing-step paradigm. The position traces are averages of five temporally aligned individual ...
Figure 6
(ai) Mean (±1 SD) peak velocity of accommodation as a function of response amplitude for the preadaptation and postadaptation sessions for each subject in the decreasing-step paradigm. The numbers above each pair of black and gray circles ...
Figure 7
(ai) Mean (±1 SD) peak acceleration of accommodation as a function of response amplitude for the preadaptation and postadaptation sessions for each subject in the decreasing-step paradigm. (c) The repeatability data from subject JW. All ...

The increasing- and decreasing-step paradigms were repeated in subject PRM (Figs. 3k, ,4k)4k) and subject JW (Figs. 6c, ,7c),7c), respectively, by repeating the adaptation-, sustaining-, and postadaptation phases of the experiment. Repeatability was tested after 15 and 8 days after the first session was conducted in subjects PRM and JW, respectively. In subject PRM, the peak velocity and peak acceleration in the increasing-step paradigm increased significantly after adaptation in both sessions (Figs. 3k, ,4k).4k). The increase in peak velocity and peak acceleration in the two sessions were not significantly different from each other (peak velocity: P = 0.51; peak acceleration: P = 0.89). In this subject, the response latency also increased after adaptation in both sessions (Table 1). The peak velocity and peak acceleration in the decreasing-step paradigm of subject JW decreased significantly in the first session (Figs. 6, ,7),7), but they failed to decrease significantly in the second session (Figs. 6, ,7).7). The response latency increased after adaptation in the first session, whereas it decreased after adaptation in the second session (Table 1). Overall, the changes in response dynamics for subject JW in the decreasing-step paradigm was less repeatable than the change in dynamics for subject PRM in the increasing-step paradigm (Figs. 3, ,44).

To assess the presence of any age-related trends in the data, we computed a correlation coefficient between the magnitude of change in peak velocity and peak acceleration of 2-D step responses with the subject's age (18–34 years). Subjects whose dynamics did not change significantly with adaptation and whose adaptive changes were in the direction opposite than expected (e.g., reduction of peak velocity in the increasing-step paradigm) were not included in the correlation calculation. Since the data of three of the nine subjects in the decreasing-step paradigm were excluded for these reasons, no correlation coefficients were computed for this paradigm. There was only a modest, statistically insignificant correlation between the age of the subject and the magnitude of adaptation of peak velocity (r = 0.48; P = 0.09) and peak acceleration (r = 0.39; P = 0.21) in the increasing-step paradigm. No such age-related trend was apparent in the magnitude of adaptation of peak velocity and peak acceleration in the decreasing-step paradigm.

To assess whether the preadapted dynamics of the individual determined the magnitude of change in dynamics that occurred after training, a correlation coefficient was computed between the preadapted peak velocity and peak acceleration and the magnitude of change in these parameters after adaptation (Fig. 8). The dynamics of only the 2-D step responses were included in this analysis and the exclusion criteria imposed in the age correlation (see earlier paragraph) were applied to this analysis too. There was a statistically significant negative correlation between preadaptation peak velocity (Fig. 8a) and peak acceleration (Fig. 8b) with the respective magnitudes of adaptation in the increasing step paradigm. No such trends were apparent in the decreasing-step paradigm.

Figure 8
Magnitude of change (%) in peak (a) velocity and (b) acceleration in the increasing-step adaptation paradigm plotted as a function of the preadaptation dynamics. Positive and negative values in the ordinate scale indicate increase and decrease in peak ...

Discussion

Until now, only indirect observations have suggested that the accommodative control system may recalibrate in response to changes in neuromuscular demand.10,13,18,19 By employing a double-step adaptation paradigm, we have directly demonstrated that the peak velocity and peak acceleration of single-step accommodative responses can change on a short-term basis, at least in some subjects ranging from 18 to 34 years, in response to optically stimulated changes in neuromuscular demand (Figs. 3, ,4,4, ,6,6, ,7).7). Unlike response latency (Table 1), adaptive changes in peak velocity and peak acceleration were dependent on the direction of training (Figs. 3, ,6),6), suggesting that stimulus anticipation cannot explain the results completely. In many subjects who showed adaptive changes in the anticipated direction, both peak velocity and peak acceleration changed after training, suggesting that adaptive changes in peak velocity were brought about by integrated changes in the peak acceleration (Figs. 3, ,4,4, ,6,6, ,77).17 Of note, the peak velocity increased after increasing-step adaptation even in those subjects where the preadaptation peak velocity saturated at larger response amplitudes (Figs. 3a, 3d, 3f). This suggests that the saturation of peak velocity observed here and in previous experiments17,33,35 does not reflect an upper limit of velocity that the accommodative system is capable of generating for a given response amplitude.36 There was no obvious correlation between the magnitude of adaptation and the subject's age (18–34 years) in the small number of individuals we tested (n = 15). This result must be confirmed on a larger sample size.

Adaptive changes were larger in the increasing-step paradigm than in the decreasing-step paradigm in several subjects, including those for whom data were available from both training paradigms. This directional bias is unlikely to be due to accommodative fatigue from prolonged periods of stimulation,3739 for fatigue would only bias the magnitude of adaptation toward lower values in both paradigms. When compared to the increasing-step paradigm, a slightly smaller adaptation stimulus was used in the decreasing-step paradigm (1.75 D vs. 2 D in the increasing-step paradigm) to ensure that the adaptation stimulus was not perceived as a transient change in blur. Although a smaller adaptation stimulus may be expected to produce a smaller change in dynamics, the difference of 0.25 D between the two paradigms appears too small to account for the observed directional bias in adaptation. Further, the results of the decreasing-step paradigm using a 2-D adaptation stimulus (first step size of 4 D) were similar to those seen with a 1.75-D adaptation stimulus in four subjects (results not shown separately). This suggests that the smaller stimulus size did not account for the reduced adaptation in the decreasing-step paradigm. A second possibility for the smaller adaptation response in the decreasing-step paradigm may be that, with the lag of accommodation, the second step decrement in blur brought the visual target into focus while the system was responding to the first stimulus, thereby negating the need for accommodation to respond to the second step. The decreasing-step paradigm may therefore not have stimulated any change in neuromuscular demand. The directional bias in adaptation may also reflect strategies used to minimize errors encountered during the normal growth and ageing processes. For instance, robust adaptation of horizontal vergence40 and the accommodation–vergence coupling gains41 to sustained convergence stimuli (than to sustained divergence stimuli) may be a strategy to correct the physiological exophoria and minimize any undue demands on the fusional vergence system during the growth of the cranium.42,43 Perhaps, the bias in adapting more effectively to increases in neuromuscular effort is also an age-related strategy to compensate for the biomechanical changes in the accommodative plant (e.g., increased lenticular viscoelasticity,4 increased posterior restriction of ciliary muscle6). Accommodation could therefore be predisposed to increase the neural gain to the increasing step paradigm used in our experiment. Further experiments are warranted to explore these different possibilities.

The results of this study and the indirect observations made by earlier studies10,13,1517,19,20 qualitatively support the presence of adaptive capability in accommodation. However, our results cannot be directly compared to these earlier studies because the time scale and the magnitude over which neuromuscular effort was modified were very different in these studies. For instance, adaptation to age-related changes would involve a gradual change in neuromuscular effort occurring over several years.10,12,13,15,16 Similarly, the change in the neural control of disaccommodation observed by Bharadwaj and Schor18 and the speeding of abnormally sluggish accommodation after orthoptic training19,20 occurred over a span of several weeks. In contrast, in our experiment, adaptive changes occurred in response to dramatic changes in neuromuscular effort that occurred over a very short period (e.g., 100% change in effort within a few hours in the increasing-step paradigm). It is therefore possible that the neural mechanisms underlying adaptation in our experiment could be different from the mechanisms that triggered adaptation in the earlier studies. Indeed, different neural mechanisms have been proposed for adaptive regulation of saccadic accuracy to errors after muscle paresis (long-term changes) and to errors induced experimentally using a double-step paradigm (short-term changes).44 Whatever the underlying adaptive mechanisms may be, our results indicate that the accommodative system possesses the capacity to change its dynamic neural control pattern for a given response amplitude and that this capacity may be used to compensate for any age- or environment-related biomechanical changes in the accommodative plant.

Similar to accommodation, the peak velocity and peak acceleration of vergence step responses also increase after double-step increase in disparity and the peak velocity decreases after double-step decrease in disparity.25,26 The dynamics of saccades also increase after double-step increase in target eccentricity and they reduce after double-step decrease in target eccentricity.27,28 Qualitative similarities in the adaptive characteristics of accommodative and vergence step responses are somewhat expected given their neural coupling45 and the similarity in their neural control strategies.17,46 Adaptive changes in accommodation (Figs. 2, ,4)4) and vergence25,26 are different from those of saccades in that, unlike accommodation and vergence, adaptive changes in saccade amplitude and dynamics occur in conjunction with each other.27,28 This difference perhaps stems from the ballistic neural control of saccades.47 The dynamics of saccades is determined by a preprogramed pulse innervation and its response amplitude is determined by integrating the pulse innervation to produce a step innervation.47 Any change in the pulse innervation would therefore influence both the dynamics and the amplitude of saccades.47 However, accommodation17,48 and vergence46 step responses are only partially ballistic, with an initial open-loop pulse innervation controlling the response dynamics and an independent closed-loop step innervation that determines the response amplitude. Any change in response amplitude induced by changes in the characteristics of the pulse innervation would therefore be corrected by the adjusting the size of the step innervation in response to blur and disparity feedback.17,46

Conclusions

The peak velocity and peak acceleration of accommodative step responses can be modified, at least in some subjects ranging from 18 to 34 years, in response to short-term changes in optically stimulated neuromuscular demands. There appears to be an adaptive bias toward increasing than decreasing the dynamics of accommodative step responses.

Acknowledgments

The authors thank Ethan Rossi, Evan Hass, and James Maxwell and the two anonymous reviewers for contributions to the manuscript.

Supported by National Eye Institute Grant EY017678 (CMS).

Footnotes

Disclosure: S.R. Bharadwaj, None; I. Vedamurthy, None; C.M. Schor, None

References

1. Glasser A, Kaufman PL. Accommodation and presbyopia. In: Kaufman PL, Alm A, editors. Adler's Physiology of the Eye. St. Louis: Mosby; 2003. pp. 197–237.
2. Heys KR, Cram SL, Truscott RJ. Massive increase in the stiffness of the human lens nucleus with age: the basis for presbyopia? Mol Vis. 2004;10:956–963. [PubMed]
3. van Alphen GW, Graebel WP. Elasticity of tissues involved in accommodation. Vision Res. 1991;31:1417–1438. [PubMed]
4. Weeber HA, Eckert G, Soergel F, Meyer CH, Pechhold W, van der Heijde RG. Dynamic mechanical properties of human lenses. Exp Eye Res. 2005;80:425–434. [PubMed]
5. Neider MW, Crawford K, Kaufman PL, Bito LZ. In vivo videography of the rhesus monkey accommodative apparatus: age-related loss of ciliary muscle response to central stimulation. Arch Ophthalmol. 1990;108:69–74. [PubMed]
6. Tamm E, Croft MA, Jungkunz W, Lütjen-Drecoll E, Kaufman PL. Age-related loss of ciliary muscle mobility in the rhesus monkey: role of the choroid. Arch Ophthalmol. 1992;110:871–876. [PubMed]
7. Koretz JF, Cook CA, Kaufman PL. Accommodation and presbyopia in the human eye: changes in the anterior segment and crystalline lens with focus. Invest Ophthalmol Vis Sci. 1997;38:569–578. [PubMed]
8. Strenk SA, Semmlow JL, Strenk LM, Munoz P, Gronlund-Jacob J, DeMarco JK. Age-related changes in human ciliary muscle and lens: a magnetic resonance imaging study. Invest Ophthalmol Vis Sci. 1999;40:1162–1169. [PubMed]
9. Strenk SA, Strenk LM, Semmlow JL, DeMarco JK. Magnetic resonance imaging study of the effects of age and accommodation on the human lens cross-sectional area. Invest Ophthalmol Vis Sci. 2004;45:539–545. [PubMed]
10. Fincham EF. The proportion of ciliary muscular force required for accommodation. J Physiol. 1955;128:99–112. [PubMed]
11. van der Hoeve J, Flieringa HJ. Accommodation. Br J Ophthalmol. 1924;8:97–106. [PMC free article] [PubMed]
12. Fincham EF, Walton J. The reciprocal actions of accommodation and convergence. J Physiol. 1957;137:488–508. [PubMed]
13. Bruce AS, Atchison DA, Bhoola H. Accommodation: convergence relationships and age. Invest Ophthalmol Vis Sci. 1995;36:406–413. [PubMed]
14. Eskridge JB. Age and the ac-A ratio. Am J Optom Arch Am Acad Optom. 1973;50:105–107. [PubMed]
15. Heron G, Charman WN, Schor C. Dynamics of the accommodation response to abrupt changes in target vergence as a function of age. Vision Res. 2001;41:507–519. [PubMed]
16. Mordi JA, Ciuffreda KJ. Dynamic aspects of accommodation: age and presbyopia. Vision Res. 2004;44:591–601. [PubMed]
17. Schor CM, Bharadwaj SR. A pulse-step model of accommodation dynamics in the aging eye. Vision Res. 2005;45:1237–1254. [PubMed]
18. Bharadwaj SR, Schor CM. Dynamic control of ocular disaccommodation: first and second-order dynamics. Vision Res. 2006;46:1019–1037. [PMC free article] [PubMed]
19. Bobier WR, Sivak JG. Orthoptic treatment of subjects showing slow accommodative responses. Am J Optom Physiol Opt. 1983;60:678–687. [PubMed]
20. Liu JS, Lee M, Jang J, et al. Objective assessment of accommodation orthoptics. I. Dynamic insufficiency. Am J Optom Physiol Opt. 1979;56:285–294. [PubMed]
21. Crane HD, Steele CM. Generation-V dual-Purkinje-image eye-tracker. Appl Opt. 1985;24:527. [PubMed]
22. Mordi J, Tucker J, Charman WN. Effects of 0.1% cyclopentolate or 10% phenylephrine on pupil diameter and accommodation. Ophthalmic Physiol Opt. 1986;6:221–227. [PubMed]
23. Bharadwaj SR, Schor CM. Acceleration characteristics of human ocular accommodation. Vision Res. 2005;45:17–28. [PubMed]
24. Ciuffreda KJ. The Glenn A. Fry invited lecture: accommodation to gratings and more naturalistic stimuli. Optom Vis Sci. 1991;68:243–260. [PubMed]
25. Munoz P, Semmlow JL, Yuan W, Alvarez TL. Short term modification of disparity vergence eye movements. Vision Res. 1999;39:1695–1705. [PubMed]
26. Takagi M, Oyamada H, Abe H, et al. Adaptive changes in dynamic properties of human disparity-induced vergence. Invest Ophthalmol Vis Sci. 2001;42:1479–1486. [PubMed]
27. Collins T, Semroud A, Orriols E, Dore-Mazars K. Saccade dynamics before, during, and after saccadic adaptation in humans. Invest Ophthalmol Vis Sci. 2008;49:604–612. [PubMed]
28. Straube A, Deubel H. Rapid gain adaptation affects the dynamics of saccadic eye movements in humans. Vision Res. 1995;35:3451–3458. [PubMed]
29. Ostrin LA, Glasser A. Accommodation measurements in a prepresbyopic and presbyopic population. J Cataract Refract Surg. 2004;30:1435–1444. [PubMed]
30. Wold JE, Hu A, Chen S, Glasser A. Subjective and objective measurement of human accommodative amplitude. J Cataract Refract Surg. 2003;29:1878–1888. [PubMed]
31. Charman WN, Heron G. Fluctuations in accommodation: a review. Ophthalmic Physiol Opt. 1988;8:153–164. [PubMed]
32. Schor CM, Lott LA, Pope D, Graham AD. Saccades reduce latency and increase velocity of ocular accommodation. Vision Res. 1999;39:3769–3795. [PubMed]
33. Kasthurirangan S, Vilupuru AS, Glasser A. Amplitude dependent accommodative dynamics in humans. Vision Res. 2003;43:2945–2956. [PubMed]
34. Suryakumar R, Meyers JP, Irving EL, Bobier WR. Vergence accommodation and monocular closed loop blur accommodation have similar dynamic characteristics. Vision Res. 2007;47:327–337. [PubMed]
35. Kasthurirangan S, Glasser A. Age related changes in accommodative dynamics in humans. Vision Res. 2006;46:1507–1519. [PubMed]
36. Baumeister M, Wendt M, Glasser A. Edinger-Westphal stimulated accommodative dynamics in anesthetized, middle-aged rhesus monkeys. Exp Eye Res. 2008;86:25–33. [PMC free article] [PubMed]
37. Bharadwaj SR, Schor CM. Initial destination of the disaccommodation step response. Vision Res. 2006;46:1959–1972. [PubMed]
38. Takeda T, Ostberg O, Fukui Y, Iida T. Dynamic accommodation measurements for objective assessment of eyestrain and visual fatigue. J Hum Ergol (Tokyo) 1988;17:21–35. [PubMed]
39. Vilupuru AS, Kasthurirangan S, Glasser A. Dynamics of accommodative fatigue in rhesus monkeys and humans. Vision Res. 2005;45:181–191. [PMC free article] [PubMed]
40. Schor CM. The influence of rapid prism adaptation upon fixation disparity. Vision Res. 1979;19:757–765. [PubMed]
41. Miles FA. Adaptive regulation in the vergence and accommodation control systems. In: Berthoz, Jones Melvill G., editors. Adaptive Mechanism in Gaze Control: Facts and Theories. Amsterdam: Elsevier Science; 1985. pp. 81–94. [PubMed]
42. Aslin RN, Jackson RW. Accommodative-convergence in young infants: development of a synergistic sensory-motor system. Can J Psychol. 1979;33:222–231. [PubMed]
43. Rethy I. Development of the simultaneous fixation from the divergent anatomic eye-position of the neonate. J Pediatr Ophthalmol. 1969;6:92–96.
44. Fitzgibbon EJ, Goldberg ME, Seagraves MA. Short term saccadic adpatation in the monkey. In: Keller EL, Zee DS, editors. Adaptive Processes in Visual and Oculomotor Systems. New York: Pergamon; 1986. pp. 329–334.
45. Mays LE, Gamlin PD. Neuronal circuitry controlling the near response. Curr Opin Neurobiol. 1995;5:763–768. [PubMed]
46. Yuan W, Semmlow JL, Alvarez TL, Munoz P. Dynamics of the disparity vergence step response: a model-based analysis. IEEE Trans Biomed Eng. 1999;46:1191–1198. [PubMed]
47. Robinson DA. Models of the saccadic eye movement control system. Kybernetik. 1973;14:71–83. [PubMed]
48. Khosroyani M, Hung GK. A dual-mode dynamic model of the human accommodation system. Bull Math Biol. 2002;64(2):285–299. [PubMed]