Abstract
A major problem for the rapidly growing population of older adults (age 65 and over) is age-related declines in vision, which have been associated with increased risk of falls and vehicle crashes. Research suggests that this increased risk is associated with declines in contrast sensitivity and visual acuity. We examined whether a perceptual-learning task could be used to improve age-related declines in contrast sensitivity. Older and younger adults were trained over 7 days using a forced-choice orientation-discrimination task with stimuli that varied in contrast with multiple levels of additive noise. Older adults performed as well after training as did college-age younger adults prior to training. Improvements transferred to performance for an untrained stimulus orientation and were not associated with changes in retinal illuminance. Improvements in far acuity in younger adults and in near acuity in older adults were also found. These findings indicate that behavioral interventions can greatly improve visual performance for older adults.
Researchers estimate that between 2000 and 2050, the number of older individuals (over the age of 65 years) in the United States will increase 135%, whereas the population under age 65 will increase only 32% (Wiener & Tilly, 2002). The number of individuals living to an advanced age (age 85 and older)—a group likely to need health- and long-term care services—is projected to increase by a staggering 350%. These percentages translate to an estimated increase of 47 million individuals age 65 and older (Wiener & Tilly, 2002). An important issue will be to develop strategies to address the health needs of this ever-growing sector of the population.
It is well documented in the literature that cognition significantly declines as people age. This decline includes changes in executive function, fluid intelligence and working memory, inhibitory control and attention, and language processing (see Craik & Salthouse, 2007, for a detailed review). In addition to cognitive declines, there are significant age-related declines in vision and visual processing that influence the health and well-being of an aging population. These declines have been identified as a major factor in the incidence of falls among the elderly (Lord, Smith, & Menant, 2010). In addition, age-related declines in vision have been associated with increased risk of motor-vehicle crashes (Owsley et al., 1998; Owsley, Stalvey, Wells, & Sloane, 1999); in fact, the crash risk for drivers over the age of 75 exceeds that for novice young drivers (Evans, 2004). The types of crashes that occur change with driver age: Compared with younger drivers, older drivers have more collisions with other moving vehicles and less single-vehicle or speed-related crashes (Langford & Koppel, 2006), which suggests that there are specific declines in visual function with age.
Given the health outcomes of age-related declines in vision, an important question is what aspects of vision and visual processing decline with age. A substantial corpus of research (see Owsley, 2011, and Andersen, 2012, for reviews) has shown age-related declines in a vast range of visual functions, including contrast sensitivity (Richards, 1977), orientation discrimination (Betts, Sekuler, & Bennett, 2007), visual acuity (Sekuler, Owsley, & Hutman, 1982), motion perception (Bennett, Sekuler, & Sekuler, 2007), form perception (Roudaia, Bennett, & Sekuler, 2008), and optic flow (Atchley & Andersen, 1998). While these age-related declines in visual function could be due to optical, retinal, cortical, or pathological changes, there is substantial evidence suggesting a cortical locus (Spear, 1993). Specifically, studies have suggested that these changes may be due to decreased cortical inhibition in visual cortex (Schmolesky, Wang, Pu, & Leventhal, 2000).
Of all these age-related declines in vision and visual processing, the decline in contrast sensitivity is one of the most pronounced. Declines in contrast sensitivity significantly affect visual function, including the ability to detect and resolve detail (e.g., Owsley, Sekuler, & Siemsen, 1983), process motion information important for balance (Sundermier, Woollacott, Jensen, & Moore, 1996), and process information when driving (Liutkevicˇiene., Cebatoriene., Liutkevicˇiene., Jasˇinskas, & Zaliu-niene., 2013). Declines in contrast sensitivity have also been found to be related to the likelihood of falls among older adults (Lord, Clark, & Webster, 1991). Contrast-sensitivity declines are most apparent at high spatial frequencies, and while it has been suggested that this decrease in contrast sensitivity is primarily due to optical factors, the decline in contrast sensitivity for moving stimuli may have a significant neural component (Burton, Owsley, & Sloane, 1993). Given these declines, are there any interventions for improving contrast sensitivity among the elderly? In the present study, we examined whether behavioral training, or perceptual learning, can counteract age-related declines in contrast sensitivity.
Perceptual learning refers to improved visual performance as a result of repeated exposure to stimuli—usually at or near an individual’s perceptual threshold. A number of perceptual-learning studies have shown that perceptual improvements are specific to the trained stimuli and do not transfer to other, nontrained stimuli or tasks. Specificity has been shown for orientation (Karni & Sagi, 1991), task difficulty (Ahissar & Hochstein, 1997), the presence or absence of external noise (Dosher & Lu, 2005), and location in the visual field (Schoups, Vogels, & Orban, 1995), as well as a combination of these factors (DeLoss, Watanabe, & Andersen, 2014). This specificity is an important issue to examine if perceptual-learning paradigms are to be used to counteract age-related declines in vision. For these reasons, it was important in the present study to examine specificity for the trained stimuli, as well as the possibility of transfer to other real-world tasks.
In the past decade, a number of studies have examined perceptual learning in older individuals (Andersen, Ni, Bower, & Watanabe, 2010; Bower & Andersen, 2012; Bower, Watanabe, & Andersen, 2013; DeLoss et al., 2014). Only one of these studies found complete specificity to an untrained retinal location (Andersen et al., 2010), while numerous other studies found partial to complete transfer to an untrained task, orientation, or stimulus (Bower & Andersen, 2012; Bower et al., 2013; DeLoss et al., 2014). This decreased specificity may be due to decreased neural inhibition, as well as decreased orientation and direction of neuronal selectivity (Schmolesky et al., 2000). Because of this decreased selectivity, broader ranges of neurons may be recruited during training, allowing for transfer not typically seen in younger adults. Although these age-related decreases in selectivity may make fine discrimination more difficult, they may result in a greater benefit from perceptual learning because of decreased specificity and greater generality for trained tasks and stimuli.
The present study assessed whether perceptual learning could be used as a possible intervention to counteract age-related declines in contrast sensitivity. Younger and older subjects performed an orientation-discrimination task using sine-wave gratings that varied in contrast. We assessed whether training improved performance for targets at a specific orientation, transferred to targets at an untrained orientation, and transferred to other tasks (e.g., near- and far-acuity tests). While previous studies have used contrast-training paradigms to improve vision in middle-age individuals (Polat, 2009), to our knowledge this is the first study to examine the use of perceptual learning to improve vision in adults 65 years of age and older. In addition, it is possible that improved performance could be due to changes in retinal illuminance—the amount of light reaching the retina. To examine this issue, we measured pupil diameter. If improved performance is due to increased light reaching the retina (pupil dilation), then retinal illuminance should be correlated with improvements in performance following training.
Method
Subjects
Sixteen younger adults (mean age = 22.43 years, SD = 1.16) from the University of California, Riverside (8 male and 8 female) and 16 older adults (mean age = 71.23 years, SD = 5.85) from the surrounding community (8 male and 8 female) participated in the study. The sample size was determined on the basis of pilot studies and previous research demonstrating an effect of training in older and younger adults (Andersen et al., 2010; Bower & Andersen, 2012; Bower et al., 2013; DeLoss et al., 2014). All subjects were compensated for their participation at a rate of $10 per hour. Older individuals received an additional $5 per hour to cover transportation costs to the campus. The sample size for each age group was comparable with sample sizes in other studies investigating the effects of perceptual learning. Subjects were naive to the purpose of the study, and all had normal or corrected-to-normal visual acuity.
All subjects were screened using an array of cognitive and perceptual tests. This procedure ensured that the older group did not exhibit any early or atypical cognitive decline (e.g., Alzheimer’s disease) that could influence the results of the study. Demographic information for the subjects is presented in Table 1. One-way analyses of variance revealed that prior to training, there were significant differences between the two age groups in acuity and on the Matrix Reasoning subtest of the third edition of the Wechsler Adult Intelligence Scale (Wechsler, 1997; Fs ≤ 2.5, ps > .05). While the older group’s performance was significantly different from that of the younger group on Matrix Reasoning, their score was well above the age norm found in previous research (Ardila, 2007). Subjects were also screened for eye disease and neurological disorders.
Demographic Information and Results From Cognitive and Perceptual Tests Administered Prior to the Start of the Study
Note: Contrast sensitivity was measured using the Pelli Robson Test (Pelli, Robson, & Wilkins, 1988). LogMAR = logarithm of the minimum angle of resolution. WAIS = Wechsler Adult Intelligence Scale–third edition (Wechsler, 1997).
For these tests, the difference between groups was significant (p < .05).
Apparatus
Stimuli were presented on a 21-in. CRT monitor (Viewsonic P225F) at a resolution of 1,024 × 768 pixels; the monitor had a refresh rate of 100 Hz (noninterlaced) and a mean luminance value of 53.82 cd/m2. Stimuli were generated on a Dell Vostro 430 PC equipped with an Intel Core i5 750 processor using the Windows XP (Service Pack 3) operating system. A GeForce GTS 240 graphics card (NVIDIA, Santa Clara, CA) was used along with a Bits++ system (Cambridge Research Systems, Rochester, Kent, United Kingdom) to achieve 14-bit gray scale (16,384 gray-scale levels). Custom experimental software was written in MATLAB (The MathWorks, Natick, MA); Psychophysics Toolbox extensions were also utilized (Brainard, 1997; Pelli, 1997). The monitor was calibrated using a ColorCal2 colorimeter (Cambridge Research Systems). Gamma correction was performed through linearization of the color lookup table.
Pupil size was measured on testing days using an EyeLink 1000 Tower Mount (SR Research, Ottawa, Ontario, Canada); subjects’ head position was also stabilized with this device. Subjects’ far acuity was measured using the 2000 Series Revised ETDRS Chart 2 (Precision Vision, La Salle, IL) at a distance of 3 m. Subjects’ near acuity was measured using the 2000 series New ETDRS Chart 3 at a distance of 40 cm. Contrast sensitivity was measured using the Pelli-Robson Contrast Sensitivity Chart (Precision Vision).
Stimuli and procedure
The experiment consisted of 1.5 hr per day of testing and training over 7 days. Subjects were required to complete the study within 3 weeks of their first testing session. Stimuli were viewed binocularly on a monitor at a distance of 94 cm. Corrective lenses normally worn by the subjects were allowed during the experiment. All stimuli were viewed through a plano-convex glass collimation lens (45.7 cm diameter) with a 19% magnification factor to minimize any age-related differences in accommodative focus. The size of the stimuli was corrected to account for this magnification factor. The experiment was run in a darkened room; the only light source was the monitor. Stimuli were Gabor patches presented at 1.5 cycles per degree of visual angle (0.65° standard deviation of the Gaussian mask), and the phase of the Gabor was randomized ±180° on each trial.
Familiarization task
On the first day of the study, before the experiment began, all subjects were given a 30-trial practice session to familiarize them with the task. These practice trials were presented without Gaussian noise. Subjects completed 15 trials in which the Gabor patch appeared in a standard orientation 45° clockwise off vertical and 15 trials in which the Gabor patch had a standard orientation 45° counterclockwise off vertical. At the beginning of each trial, subjects saw a fixation point in the center of the display for 1,600 ms. This fixation point alternated between black and white every 400 ms, to attract attention (Betts, Taylor, Sekuler, & Bennett, 2005). A Gabor patch then appeared at one of the two standard orientations for 100 ms. A second fixation point then appeared for 1,200 ms, alternating between black and white every 300 ms, followed by the target stimulus for 100 ms. During this phase, the target was rotated either 25° clockwise or 25° counterclockwise away from the standard orientation. After the target disappeared, the screen assumed a uniform midgray value indicating that the subject should judge whether the target stimulus was rotated clockwise or counterclockwise compared with the standard orientation (Fig. 1). Responses were made using the left and right arrow keys on the keyboard. Audio feedback was provided on each trial to indicate whether the subject was correct. Subjects were then prompted to “press any key to continue” to begin the next trial. During the familiarization task, subjects were instructed to get at least one trial incorrect to familiarize themselves with the corresponding auditory feedback.
Example of the task used in the study. In each trial, subjects saw a Gabor patch at one of two standard orientations—25° clockwise (shown here) or 25° counterclockwise for training and testing trials, 45° clockwise or 45° counterclockwise for familiarization trials. After this Gabor patch disappeared, subjects saw a second stimulus and had to judge whether it was rotated clockwise or counterclockwise in comparison with the standard orientation (the examples shown here are rotated 15° clockwise and counterclockwise off the standard orientation, respectively).
Testing
Orientation-discrimination thresholds were tested during the first and last days (Days 1 and 7) of the experiment. During testing, the Gabor patches were embedded in additive Gaussian noise in four of the five testing blocks. The first block consisted of a no-noise condition with the standard deviation of the additive Gaussian noise set to zero; the standard deviation of the Gaussian distribution was then increased by 0.084375 in each successive block to a maximum of 0.3375 in Block 5.
Standard stimulus orientations were either clockwise 25° or counterclockwise 25° off vertical (see Fig. 2 for an example of the stimuli). The testing order of these standard orientations was counterbalanced across subjects. These two orientations were chosen on the basis of previous research demonstrating that improvements in orientation sensitivity begin to degrade when the stimulus is rotated approximately 40° away from the trained orientation (Matthews, Liu, & Qian, 2001). These two specific standard orientations were chosen because they are 50° offset from one another.
Example of contrast and noise levels used in the experiment. Gabor patches are displayed at 75% contrast in the top row and at 25% contrast in the bottom row. Stimuli were presented in five blocks (examples shown from left to right). There was no noise in the first block, but starting with the second block, stimuli were presented in Gaussian noise, with the noise level increasing in each subsequent block.
During Days 1 and 7, pretraining Michelson contrast thresholds for the trained and untrained standard orientations were assessed at five noise levels using the QUEST procedure (Watson & Pelli, 1983). QUEST was initialized with a criterion level of 0.75 (β = 1.4, δ = 0.025, γ = 0.5). On testing days, the estimated contrast values for QUEST for each successive noise level were 10%, 15%, 20%, 25%, and 30%, with a standard deviation of 5%. Subjects completed 60 trials at each noise level during testing. All stimuli were viewed through a circular annulus with a radius of 8 degrees of visual angle that was placed against the surface of the monitor. The annulus removed any edge cues that may have been used in the orientation-discrimination task. The background between trials and during interstimulus intervals consisted of the midgray value of the monitor (for no-noise blocks) or additive Gaussian noise with the mean set at the midgray level of the monitor and a standard deviation that matched the additive Gaussian noise (for blocks with external noise).
Trials progressed in the same fashion as in the familiarization task. The orientation offset was fixed at 15° on all test trials. Posttraining thresholds for the trained and untrained orientations were measured on Day 7 using the same procedure as on Day 1. Pupil size was also measured on each trial during testing days to derive retinal-illuminance values.
Training
Training occurred on Days 2 through 6 using the same stimuli and a similar procedure as in the testing phase, with the key difference being that subjects were trained on only one standard orientation instead of the two they saw on testing days. During training, all subjects completed five blocks, one at each noise level. There were 150 trials per block, and subjects were allowed to take a short break after each block. During each training day, the QUEST procedure was run using the same parameters as during the testing days, but with one modification. The contrast threshold estimate for QUEST for each subject used the threshold derived on the previous day of testing or training as the initial estimate. Subsequent trials used the current threshold, taking into account each trial during the training sessions. Using this method, subjects were constantly trained at their threshold (75% correct) during each training day. This allowed training to be constantly optimized for any improvement that occurred between or within testing or training sessions. All subjects received 750 training trials per day, for a total of 3,750 training trials over the course of the experiment.
Results
Threshold analysis
A 2 (testing day: pretraining, posttraining) × 5 (noise level: 0–4) × 2 (age group: older, younger) × 2 (orientation: trained, untrained) mixed repeated measures analysis of variance was conducted on the thresholds obtained on the testing days. Two older individuals and three younger individuals had to be removed from the analysis, as QUEST was not able to converge on a stable threshold estimate in at least one of the conditions before or after training. Results revealed a significant effect of testing day, F(1, 25) = 30.696, p < .001, η p 2 = .551, which indicates that there was a significant improvement in perceptual thresholds after training. A significant Testing Day × Age Group interaction was also found, F(1, 25) = 6.583, p = .017, η p 2 = .208: Compared with younger adults, older adults showed a greater performance improvement after training, as shown in Figure 3. There was a significant difference in performance between older and younger adults prior to training, F(1, 25) = 8.314, p = .008, η p 2 = .250. However, after training, this effect of age was no longer significant, F(1, 25) = 3.996, p = .057, η p 2 = .138; performance for older and younger subjects was similar following training.
Mean Michelson contrast thresholds for older and younger adults as a function of noise level (values indicate the standard deviation of the Gaussian distribution) and testing day. Results are shown separately for trained and untrained orientations of the stimuli. Error bars indicate 95% confidence intervals.
There was also a significant effect of noise level, F(4, 100) = 198.463, p < .001, η p 2 = .888. As expected, as the level of external noise increased, so did subjects’ contrast threshold. A significant Noise Level × Age Group interaction was also found, F(4, 100) = 4.209, p = .035, η p 2 = .144. Older individuals showed lower tolerance to external noise, as exhibited by higher thresholds in the high-noise conditions (Blocks 4 and 5), both before and after training. There was also a significant Testing Day × Noise Level × Age Group effect, F(4, 100) = 2.506, p = .047, η p 2 = .091. Older adults exhibited greater reductions in thresholds in the high-noise conditions, compared with younger adults.
A possible factor that might account for the observed learning is that improvements in performance are a result of repeated testing with the stimuli and practice with the task. To examine this possibility, we analyzed the change in threshold between Days 1 and 2. The main effect of day was not significant for younger subjects, F(1, 13) = 2.04, p = .177, η p 2 = .135, or for older subjects, F(1, 12) = 3.23, p = .098, η p 2 = .212, which suggests that the improved performance over the full range of training was not the result of repeated testing with the stimuli or task practice.
Retinal-illuminance analysis
To examine whether a change in pupil size was correlated with the change in threshold, we took pupil-size measurements during each stimulus presentation, which were then converted into Trolands, the conventional unit of retinal illuminance. Percentage-change scores from before to after training for retinal illuminance and contrast thresholds were calculated. Any change greater than ±2.5 standard deviations was excluded from the analysis. All reported p values are one-tailed because we hypothesized that letting more light into subjects’ retinas would cause an improvement in performance for low-contrast stimuli, particularly in older individuals who have decreased retinal illuminance. Overall, there was a significant correlation between the percentage change in threshold and percentage change in retinal illuminance after training, r(224) = −.134, p = .019. However, further analysis indicated that this correlation was driven primarily by younger adults. When the data were split by age group, younger individuals continued to show a significant correlation, r(122) = −.228, p = .004, while older individuals showed no significant correlation, r(102) = −0.059, p = .276, as shown in Figure 4.
Scatter plot showing the percentage change in threshold as a function of the percentage change in retinal illuminance, separately for older and younger subjects. Each point represents a single measurement at one of the five noise levels. The lines indicate the best-fitting linear trend for each age group.
Previous research has shown that pupil size can alter the retinal-modulation transfer function (Van Nes & Bouman, 1967), which might enhance low and high spatial frequencies of the external noise and subsequently affect learning. To examine this issue, we correlated, between pretraining test and posttraining test, the change in pupil size with the change in contrast thresholds. The correlation was significant for younger adults, r(122) = −0.36, p < .001, as well as for older adults, r(102) = −0.18, p = .035. However, this accounted for only 13% of the variance in the improvements seen in younger adults and only 3.2% of the variance in older adults. Changes in pupil size can improve the ability to filter noise, and these findings suggest that any changes in the spatial-frequency characteristics of the external noise due to changes in pupil size are likely to be only a small factor in the overall effect of learning in younger individuals and explain very little of the overall effect of learning in older subjects.
Contrast-sensitivity and acuity analysis
Near-acuity, far-acuity, and contrast-sensitivity measurements were made before and after training. Contrast sensitivity was measured using the Pelli-Robson Test (Pelli, Robson, & Wilkins, 1988). One-tailed p values are reported for all measures. A single contrast-sensitivity measurement for 1 younger subject was lost and is missing from the analysis.
Neither younger nor older individuals demonstrated any significant improvement in contrast sensitivity, t(12) = −1.389, p = .190, and d = −0.385, t(12) = −1.477, p = .0825, d = −0.410, respectively (Fig. 5). To adjust for multiple comparisons when analyzing the acuity data, we used the Benjamini-Hochberg procedure to control the false discovery rate (FDR; Benjamini & Hochberg, 1995). Older individuals showed a significant improvement in near acuity, t(12) = 2.217, pFDR = .047, d = 0.615, whereas no significant change in near acuity was found in younger individuals, t(13) = 0.265, pFDR = .398, d = 0.071 (Fig. 6). Similarly, older individuals showed no significant change in far acuity, t(12) = 0.503, pFDR = .398, d = 0.140, whereas a significant improvement in far acuity was found for younger individuals, t(13) = 2.592, pFDR = .044, d = 0.693 (Fig. 7).
Mean log contrast sensitivity as a function of age group and testing day. The difference between pre- and posttraining sensitivity is also shown. Error bars indicate 95% confidence intervals.
Mean near acuity as a function of age group and testing day. The difference between pre- and posttraining acuity is also shown. Error bars indicate 95% confidence intervals. LogMAR = logarithm of the minimum angle of resolution.
Mean far acuity as a function of age group and testing day. The difference between pre- and posttraining acuity is also shown. Error bars indicate 95% confidence intervals. LogMAR = logarithm of the minimum angle of resolution.
Overall, these results are important for several reasons. First, contrast sensitivity can be improved for trained stimuli through behavioral training in both younger and older adults. The performance improvements for older adults are striking. Consider contrast-sensitivity thresholds for the highest-noise condition—the most difficult condition for both age groups. Prior to training, younger adults (mean threshold = 0.25) had significantly lower thresholds than older adults (mean threshold = 0.39) in the high-noise condition, F(1, 25) = 6.341, p < .01, η p 2 = .201. However, an analysis comparing performance in the high-noise condition for younger adults before training (mean threshold = 0.25) and older adults after training (mean threshold = 0.26) was not significant, F(1, 25) = 0.152, p > .05, η p 2 = .006. The change in the pattern of results is due to improved performance in older adults, as indicated by a significant effect of training for the high-noise condition (pretraining and posttraining thresholds = 0.39 and 0.26, respectively), F(1, 12) = 10.215, p < .01, η p 2 = .460.
Discussion
The present study examined the use of perceptual learning to improve older adults’ ability to see low-contrast stimuli. A major finding of the present study is that 5 days of training for older adults resulted in performance that was not statistically different from that of younger adults prior to training. This finding indicates that perceptual learning can be used to counteract age-related declines in contrast sensitivity. These improvements are the result of changes in sensory processing and not due to optical efficiency of the eye.
Although the magnitude of learning was considerable for processing the contrast of the trained stimuli, this learning did not to transfer to performance on a standard contrast-sensitivity chart. This may be because of the low fidelity of a chart-based test, in which small improvements in contrast sensitivity may not be detectable. Future research that measures the entire contrast-sensitivity function will be needed to determine whether there are any fine changes in contrast sensitivity to untrained stimuli.
A second important finding was that both younger and older individuals showed significant improvement in acuity with perceptual-learning training to improve contrast sensitivity. Notably, the improvements in acuity were associated with the range of acuity most problematic for each age group. Younger individuals tend to have near-optimal near acuity, as a result of a strong and fast accommodative response (Koretz, Kaufman, Neider, & Goeckner, 1989); in the present study, they showed a significant change only in far acuity. Older individuals suffer from losses in near acuity, resulting from significant declines in the accommodative response (Koretz et al., 1989). The improved performance in near acuity for older adults might be due to changes in improvements in the accommodative response. Previous research has found improved acuity after perceptual learning in middle-age adults (median = 51 years) when declines in accommodative focus or presbyopia occur (Polat et al., 2012). However, unlike in the present study, in which there were 5 days of training, the training for middle-age adults occurred three times a week for 3 months over 36 training sessions (Polat et al., 2012). An important issue for future research would be to measure the accommodative focus response of older adults (i.e., over the age of 65 years) in response to training in contrast sensitivity.
A third finding was that both age groups showed significant transfer of learning to the untrained orientation. This transfer was more pronounced at higher noise levels. This is likely because the primary limiting factor for discrimination in high-noise cases is the external noise itself, not the contrast of the stimulus. This also suggests that filtering of external noise is not specific to the trained stimulus and may be a more general process that can transfer to a variety of tasks. One other possible explanation for the high degree of transfer is that the task used in the present study is a fairly low-precision task. At high levels of contrast with no noise, subjects can perform the task consistently with near-perfect accuracy, aside from any attentional lapses in their responses, as the task requires only coarse orientation discrimination with a constant 15° offset. Previous research has shown that high-precision tasks tend to show greater specificity than do low-precision tasks (Jeter, Dosher, Petrov, & Lu, 2009).
Finally, no significant correlation between a change in pupil size and improvement in the task was found for older individuals. However, a significant correlation was found in younger individuals, though this accounted for only around 5% of the variance. While we proposed previously that older individuals may be compensating for a loss in retinal illuminance by dilating the pupils to allow more light to fall on the retina, the results from the retinal-illuminance data do not support this hypothesis. Given the small amount of variance explained by retinal illuminance in younger individuals, it is unlikely that this was a primary factor in the improvements seen in the task or the improvements seen in far acuity. It is intriguing that there was a significant change in retinal illuminance in younger adults; additional research will be needed to replicate this finding and determine the impact of this effect on training and improved vision.
In summary, the results of the study indicate that perceptual learning may be a viable intervention for age-related declines in vision. The present study consisted of only 5 days of training, whereas many perceptual-learning studies consist of much longer training periods, in some cases up to 30 days (Karni & Sagi, 1991; Levi, Polatt, & Polat, 1996). Given the short training period, the degree of improvement is quite impressive, particularly in the cases of near and far acuity, in which subjects were able to read an average of two to three additional letters on acuity charts after training. Further study is needed to determine how much more improvement in near and far acuity may be obtained by older adults, as there was not a long enough training period in the current study to determine whether asymptotic levels of performance were reached. However, the present findings are quite promising and provide further evidence of the plasticity of visual processing in advanced age.
Footnotes
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article.
Funding
This research was funded by National Institute on Aging Grant No. R01-AG031941.