Measurement of Difference Thresholds on the Ocular Surface

Abstract

Purpose.: To establish difference thresholds of the central cornea and compare thresholds between contact lens wearers and noncontact lens wearers.

Methods.: Mechanical sensitivity of the central cornea was determined in 12 lens wearers and 12 nonlens wearers using a modified Belmonte pneumatic esthesiometer and method of limits. Then, a series of systematically increasing stimuli were presented, with the first stimulus being 25% less than threshold. Subjects compared intensity of each stimulus with the preceding one and reported if any difference in intensity was detectable. Intensities at which an increase was perceived from the prior stimulus were recorded and the difference between the intensities was the difference threshold (DL). Five DLs were measured for each subject. Weber's constants that relate the size of difference threshold to stimulus intensity were derived for each DL level. Repeated-measures ANOVA was used to compare Weber's constants between lens wearers and nonlens wearers.

Results.: A significant main effect of DL level on Weber's constant (P < 0.001) was observed, with the first DL being higher than following DLs. Lens wearers had higher Weber's constants than nonlens wearers (P = 0.02) However, no interaction was found between DL level and group type on Weber's constants (P = 0.38).

Conclusions.: Differential sensitivity of the ocular surface can be successfully measured with a pneumatic esthesiometer and it appears that Weber's law holds for corneal nociceptive sensory processing. There are subtle differences in mechanical difference thresholds between lens wearers and nonlens wearers, suggesting the possibility of different neural activity levels in the two groups.

Introduction

Measurement of threshold plays a pivotal role in the assessment of any sensory system. Absolute threshold refers to the smallest amount of stimulus energy necessary to produce a sensation, while difference threshold indicates the amount of change in a stimulus (ΔΦ) required to produce a just noticeable difference (JND) in sensation. ¹ Typically, greater change (ΔΦ) is required to detect changes in higher-intensity stimuli than stimuli of lower intensities. The increase or decrease in the intensity of the stimuli that is just noticeably different (ΔΦ) is a constant fraction (c) of the starting intensity of the stimulus (Φ), as stated by Weber's law. ¹  

 

 

Sensitivity is unique to each sensory system, and within each sensory modality, the stimulus dimension differs in intensity, quality, and duration. Although discrimination thresholds (ΔΦ) cannot be compared across various sensory systems and stimulus dimensions, their relative sensitivities can be compared by means of Weber's fraction. ¹ Experiments have been conducted in vision, ² hearing, ^3–7 smell, ⁸ and tactile sensation, ^9–11 among others, to study the sensory discrimination differences. Weber's fractions range from 0.3 for the pitch of pure tones ¹² to 25 for odor intensity, ⁸ and the size of the Weber's fraction is hypothesized to be associated with the ability of the sensory system to detect differences. Vision and hearing have been found to be the keenest senses, while taste and smell are the dullest. ¹  

The only ocular surface thresholds ever reported have been absolute thresholds. Early detection thresholds have been estimated by using Cochet-Bonnet type esthesiometers ¹³ that involve touching of the cornea with a fine nylon filament. This procedure of tactile stimulation suffers from several drawbacks ^14–16 and it has by large been superseded by pneumatic esthesiometers. ¹⁷ Although absolute threshold measurements with the pneumatic esthesiometer are found to be reliable and repeatable, ^15,18,19 there are discrepancies in the thresholds reported by different studies. A few of the contributing factors for the variability ^{14,15,17–19} include differences in psychophysical techniques used, the characteristics of the instruments used, the time of the day the measurements were obtained, the type of stimuli used, the distance of the probe from the ocular surface, and the duration of the stimulus. 

Studies that involve tactile stimulation show a reduction in corneal sensitivity with contact lens wear, ^20–22 with the decline in sensitivity associated with the use of low Dk lenses, ²³ decreased oxygen permeability, ^22,23 duration of contact lens wear, ^21,24,25 and type of lenses worn. ^26,27 However, measurements taken with the pneumatic esthesiometers show mixed results. Comparison of corneal sensitivity between lens wearers (soft and rigid gas permeable) and nonlens wearers indicates a difference in sensitivity between the groups, ²⁸ while there are no differences between soft and rigid lens wearers. Counter to this, though, is a report by Stapleton et al. ¹⁸ that there is no difference in corneal sensitivity between nonlens wearers and users of low Dk and high Dk lenses. Another study ²⁹ shows that wearers of silicone hydrogel lenses exhibit an increase in sensitivity. 

On the ocular surface, investigations of JND have been limited to the interocular difference in comfort assessment using contact lenses. ³⁰ The authors suggest that subjects could detect 7 to 8 units differences (on a 100-point numerical scale) in comfort between the two eyes. The study perhaps provides some sort of preliminary information about the interocular discrimination capacity of the ocular surface, but traditional discrimination data cannot be determined as contact lenses are not a source of fixed stimulus strengths that can be quantified (the stimulus intensity is not related to the sensory difference). Therefore, to establish the difference thresholds for the ocular surface, suprathreshold stimuli of quantifiable intensities need to be used. Our study fills in this inadequacy by measuring the difference thresholds of the central cornea by using a modified Belmonte pneumatic esthesiometer and investigated if there are any differences in thresholds between contact lens wearers and noncontact lens wearers. 

Methods

Subjects

Twenty-four participants were enrolled for the study; 12 were adapted soft contact lens wearers (7 females, 5 males) and 12 were noncontact lens wearers (6 females, 6 males). The age of the participants ranged from 21 to 33 years. Subjects in good ocular and systemic health were chosen for the study in order to establish a baseline for difference threshold measurements on the ocular surface. Contact lens wearers wore either daily or biweekly disposable hydrogel or silicone hydrogel lenses and were asked to cease lens wear on the night before the test procedures were performed. These subjects used lenses for more than 1 year and wore lenses for at least 3 days a week for a minimum of 5 hours a day. The study was conducted in accordance to the guidelines of the Declaration of Helsinki, and ethics clearance was obtained from the University of Waterloo, Office of Research Ethics (Waterloo, Ontario, Canada). Informed consent was obtained from each participant. 

Instruments

The construction of the computer-controlled Belmonte esthesiometer has been described in various studies. ^17,31,32 The modified pneumatic Belmonte esthesiometer was used to deliver mechanical stimuli to the central cornea. It contained computerized controllers regulating flow and components for mixing air and CO₂ and controlling the stimulus temperature. The mechanical stimuli consisted of series of air pulses (in this instance with 0% added CO₂) with flow rate ranging from 0 to 200 mL/min. The temperature of the stimulus was set at 50°C, which was 33°C at the ocular surface. ³³ The distance between the ocular surface and the tip of the esthesiometer probe was set at 5 mm and was continuously monitored by a calibrated video camera. The stimulus duration was 2 seconds, with a 2-second interval between the subject's response and the next stimulus. Custom software was used to automate the appropriate psychophysical procedure, set the stimulus intensity and duration of the stimulus, and record the subjects' responses from a button box. 

Psychophysical Procedures

Figure 1 schematically illustrates the measurement of detection and difference thresholds. The experiment was carried out on a randomly selected eye of each participant. A mechanical detection threshold of the central cornea was first determined by using the ascending method of limits. The initial stimulus intensity was 10 mL/min and increased in 10 mL/min-steps until the participant reported the presence of the pneumatic stimulus. The detection threshold was then fine-tuned by decreasing the flow rate by 20 mL/min and then increasing in 5 mL/min-steps until the stimulus was detected. The procedure was carried out three times and the absolute threshold was an average of the three fine reversals. Following the measurement of absolute threshold, a series of systematically increasing stimuli (in 5-mL increments) were presented, with the first stimulus being 25% less than the absolute threshold. The subjects were asked to compare the intensity of each stimulus with the preceding one and report if any difference in intensity was detectable. The intensities at which the subjects perceived an increase from the previous stimulus were recorded. The difference threshold (DL) was the differences between the stimulus intensities at which an increase was perceived, and five DLs were measured for each participant. 

Figure 1

 

Schematic presentation of the detection and difference threshold measurements used in the study. ★ represents the intensity at which the subject reports the presence of the stimulus. Background intensity refers to the intensity of the stimulus applied to the cornea in addition to the background activity of the nervous system. Each DL represents the difference between the two stimulus intensities at which an increase in intensity is perceived. The minor irregularities in the levels in the figure refer to the neural stimulation as a consequence of the interaction between pneumatic stimulus, tear film, and epithelial neural tissue.

Data Analyses

Statistical analysis for this study was performed with Statistica 9.0 (Statsoft Inc., Tulsa, OK) and R 3.0.1. ³⁴ Independent sample Student's t-test and Bayesian analyses were used to examine detection thresholds. Repeated-measures ANOVA was used to compare the difference thresholds between lens wearers and nonlens wearers. P < 0.05 was considered to be statistically significant for all the tests. Bilinear functions using least squares nonlinear regression were fit to the group data of lens and nonlens wearers. 

Results

Comparison of mean absolute threshold in lens- and nonlens-wearing groups did not show any statistically significant difference (P = 0.07). To further test this absence of a difference, two additional Bayesian analyses were used: The scaled JZS Bayes factor ³⁵ was 0.85 and the 95% highest density interval ³⁶ included zero, both suggesting that there were no differences in detection thresholds between the groups. Repeated-measures ANOVA revealed a significant main effect of DL level on the Weber's constant, with the Weber's constant at the first DL being higher than the following DLs (all P < 0.001). The Weber's constant for the second DL was also statistically different from the last DL (P < 0.001). A significant main effect of the group type was also observed, with lens wearers showing higher Weber's constants than nonlens wearers (P = 0.02). However, there was no interaction between DL level and group type on Weber's constants (P = 0.38). The results from the repeated-measures ANOVA are illustrated in Figure 2. 

Figure 2

 

Mean difference thresholds in lens and nonlens wearers. The Weber's constants for different DL levels appear to follow a similar pattern in both groups, while lens wearers exhibit higher Weber's constants than nonlens wearers. Vertical bars denote 95% confidence intervals (CI) of mean thresholds.

The bilinear fit of the Weber's constants in the control and lens-wearing samples are shown in Figures 3a and 3b, respectively. In the bilinear models, there is an early window of stimulus intensities during which the Weber's constants are steeply declining, followed by a range of intensities during which the function appears to be flat. In nonlens wearers, the intersection of the two linear functions occurred at 80 mL/min, while in lens wearers, the break was at 96 mL/min. Also, the range of intensities used by nonlens wearers extended up to 170 mL/min, whereas the maximum intensity in lens wearers was 100 mL/min. 

Figure 3

 

Weber's fractions for different stimulus intensities in the group of (a) nonlens wearers and (b) lens wearers. The graph represents individual results obtained from all the noncontact lens wearers in the study, pooled together to show the pattern of Weber's constants over the range of stimulus intensities used. Each data point indicates the Weber's fraction obtained for the respective stimulus intensity. The dashed line represents the bilinear line of best fit. Background intensity refers to the intensity of the stimulus applied to the cornea.

Discussion

Investigations on the absolute thresholds of the cornea and conjunctiva, ^{14,16–19,31,32} and the ocular surface response to suprathreshold stimulus of different modalities demonstrate that the pneumatic esthesiometer can be effectively used to study the sensory processing on the ocular surface. ^17,37–39 Our study in addition provided important insight about how the ocular surface detects differences in stimulus intensities. The results of the study indicated that the difference thresholds of the central cornea can be successfully measured by using a pneumatic esthesiometer and there can be differences in how lens wearers and nonlens wearers appreciate stimulus changes. Understanding this functioning is vital to complete the characterization of ocular surface sensory processing. Detection and discrimination can be scientifically hypothesized to involve separate neural mechanisms inasmuch as the decision about a detection of stimulus is in the “absence” of a background or only in the presence of (usually) low levels of background noise, while a different decision is required to detect a stimulus when the background is more prominent. For example, the signal to noise ratio in the former (detection) condition is almost infinite, whereas with a substantial background, the signal to noise ratio is finite. On the other hand, detection and discrimination can be described as identical (detection of stimulus increment against physiological background or physiological plus additional stimulus background). Regardless of one's point of view about detection and discrimination, almost nothing in the world occurs only at detection threshold levels. This truism exists for any sensory processing system ⁴⁰ and therefore, a complete description of how the ocular surface sensory system performs cannot be done without understanding discrimination. Finally, since any treatment is presumably going to be developed to treat uncomfortable/painful conditions (e.g., dry eye), every sensory change (e.g., a decrement in symptoms) is a discrimination judgment (i.e., an improvement of existing symptom intensity), and understanding discrimination is imperative in order to quantify these potential therapeutic effects. 

The Weber's fraction is a useful index of sensory discrimination for comparison across different modalities and stimulus conditions. In this experiment, the Weber's fraction (ΔΦ/Φ) approaches a constant value for high intensities and increases rapidly for low-stimulus intensities close to the absolute threshold. A similar phenomenon has also been observed in the discrimination of auditory tones ⁵ and tactile vibration. ¹⁰ In a study ⁵ that has investigated the intensity discrimination of auditory tones, the Weber's fraction rapidly decreases with an increase in stimulus intensity and the fraction gradually decreases without becoming a constant. This deviation from the Weber's law for low-intensity stimuli is known as the “near miss” of the Weber's law. ⁶ Discrimination experiments examining loudness and noise ⁷ also have demonstrated the Weber's constant to be higher with stimulus intensities closer to threshold that became a constant with increasing levels of intensity. The increase in Weber's constant toward the lower-intensity stimuli can be due to the presence of noise in the stimulus or sensory/neural noise, fluctuations in the activity of the neurons that carry signals from the ocular surface to the central nervous system, ¹ which perturb the background sufficiently at low levels to interfere with discrimination thresholds and so elevate Weber's fraction. The background noise can also be present in the absence of the stimulus ⁴¹ and for a stimulus to be above detection threshold, the ocular sensory system would need to respond strongly enough for the sensation to be distinguishable from the activity in the absence of the stimulus (sensory noise). This theoretical interpretation aside, it is clear from the data in our experiment that the ocular surface tissue is another place where Weber's law also holds. 

The average difference thresholds were higher in lens wearers than nonlens wearers although there was no difference in absolute detection threshold between the two groups. Recent studies examining ocular surface sensitivity in silicone hydrogel contact lens wearers have demonstrated no changes in sensitivity ¹⁸ or an increased sensitivity ²⁹ with lens wear. Contact lenses have been hypothesized to depress the sensitivity of the cornea by the mechanical effects of the lens against the ocular surface or cause a change in corneal physiology perhaps due to hypoxia, ²⁸ and also because of lens wear effects on the tear film. ^42–44 Any or a combination of these factors might alter the equilibrium of the ocular surface, causing a change in sensory input in contact lens wearers, possibly increasing the neural noise in lens wearers and thereby giving rise to increased Weber's constants. The shift in neural activity levels between lens and nonlens wearers can also be observed in the bilinear fit of the pooled data as illustrated in Figures 3a and 3b. The relatively fewer data points for stimulus intensities higher than 100 mL/min in lens wearers might have been because only 5 DLs were measured for each participant or perhaps, lens wearers were able to judge intensity discrimination between closely spaced stimulus intensities, whereas nonlens wearers required a higher change in intensity to appreciate a difference. Despite this apparent limit, there still were two linear components to the bilinear fits as illustrated in Figure 3b. Another potential explanation for the observed difference in lens wearers is that the discrimination of pneumatic stimuli itself may be processed differently in the nervous system of lens and nonlens wearers owing to the differences in ocular surface physiology. 

Suprathreshold sensory processing on the ocular surface has been shown to be different for mechanical and chemical stimulus modalities, ^17,31,37 and with varying locations (Situ P, et al. IOVS 2007;48:ARVO E-Abstract 5387) but the influence of contact lens wear on suprathreshold stimuli is yet to be studied. At suprathreshold levels, as in discrimination of pneumatic intensities, the stimulus comparisons are based on the differences in the sensation magnitude of two stimuli, whereas absolute threshold involves the difference between no sensation and presence of a sensation (i.e., the detection of stimulus from the background), and therefore it is not unreasonable that discrimination and detection could be affected differently in lens wearers. 

Not finding a statistical difference in absolute thresholds between lens- and nonlens-wearing groups might be because there are no differences. Pneumatic esthesiometry has been used to examine detection thresholds and the results are equivocal, with some experiments finding no difference and others finding differences in one direction or the other. ^{20–22,25,27} No difference uncovered could also be an error, not revealed because we did not have statistical power in the experiment to uncover this effect, partly owing to the sample sizes. We did not use detection thresholds to estimate sample size since this outcome variable is not integral to our experiment and hypotheses. We measured it in order to scale suprathreshold discrimination stimuli appropriately and we tested detection threshold differences to determine if we were able to demonstrate differences statistically, but it is an ancillary outcome. In other sensory dimensions, ^11,45 dissociation between detection and discrimination thresholds has been reported. For example, difference thresholds have been found to be similar even though there were detection threshold differences. Therefore, even if the absence of a statistical difference between absolute thresholds is actually a type II error, it is not self-evident that the difference in discrimination thresholds is also an error arising from differences in stimulus detectability. In addition, the increased variance in detection thresholds in nonlens wearers might also have contributed to a smaller effect size (and therefore nonsignificant difference in detection thresholds between the two groups), while indeed there were actual differences existing between them (i.e., a type II error). The origin of this variance difference between groups is unclear and could itself perhaps be simply due to sampling differences rather than being meaningful physiologically. 

The clinical consequences of differences in discrimination thresholds is by no means clear, but it would suggest that contact lens–adapted subjects would in general be less able to tell differences apart than those not wearing lenses or require greater change in stimuli than those not wearing lenses to detect differences. For example, as many as half of lens wearers report discomfort toward the end of the day ^46,47 and the data in this study suggest that they perhaps have “adapted” to be more tolerant of uncomfortable stimuli so that greater shifts in uncomfortable stimuli would be required for them to report change. This reduced ability to detect change is also complicated by the phenomenon reported in symptomatic lens wearers who show less suprathreshold adaptation than asymptomatic lens wearers. ⁴⁸ Also, lens wearers might be less able to differentiate comfort (or other sensory effects) when wearing different lens types. 

In summary, we determined the difference thresholds of the ocular surface by using a pneumatic esthesiometer. In addition, ocular surface sensory processing is altered by contact lens wear, as exhibited by the differences in discrimination thresholds for pneumatic stimuli, with lens wearers demonstrating lower discrimination sensitivity. The importance of this difference needs to be established in larger groups of subjects and in those with underlying differences other than controls and lens wearers, but these results do point to a novel neural metric that might be used when trying to understand ocular surface sensory processing in a fuller context. 

Acknowledgments

Supported by an operating grant from Natural Sciences and Engineering Research Council of Canada (NSERC). 

Disclosure: S. Basuthkar Sundar Rao, None; T.L. Simpson, None