Characteristics of Human Corneal Psychophysical Channels

Abstract

purpose. To characterize human corneal psychophysical channels.

methods. Twenty subjects participated in this study. A Belmonte pneumatic esthesiometer was used to deliver stimuli, and the ascending method of limits and the method of constant stimuli were used to estimate thresholds. Sensation was characterized for different stimuli. Corneal mechanical and chemical thresholds were measured at different temperatures.

results. The qualities of the sensations induced by stimuli with different temperatures were different, and the corresponding detection thresholds of the pneumatic stimuli at four temperatures gradually increased (repeated measures ANOVA (F_(3,12) = 10.326, P = 0.000). There were no temperature effects on chemical thresholds (repeated measures ANOVA F_(3,12) = 0.235, P = 0.870) or mechanical discomfort thresholds from 20°C to 50°C (paired t ₍₁₄₎= −0.233, P = 0.818). There were strong interactions when chemical and mechanical stimuli were added. Chemical thresholds were progressively lower when the flow rate increased and mechanical thresholds went down as the percentage of added CO₂ increased (repeated measures ANOVA F_(3,12) = 6.407, P = 0.007, F_(4,16) = 19.904, P = 0.000).

conclusions. This study suggests that humans sense corneal non-noxious cold and noxious mechanical and chemical stimuli, and that the sensitivity of some submodalities can be modulated by others. There are at least five psychophysical channels (non-noxious cold, noxious mechanical, noxious chemical-H⁺, noxious heat, and itching) processing corneal sensory information. Both decreased corneal chemical thresholds at high flow rates and decreased mechanical thresholds with an added chemical stimulation demonstrate that corneal psychophysical channels are not independent.

Psychophysical channel theory has been applied in many sensory systems such as vision, audition, and somatosensation. ^{1 2} In the somatosensory system, different submodalities such as temperature and pressure are hypothesized to be processed by different psychophysical channels and produce different sensations—for example, cold, touch, and pain. ^{1 2 3 4} Unlike the visual and auditory systems, however, in which different submodalities stimulation are processed independently, psychophysical evidence of the contribution of non-nociceptive stimulation to nociception, for example, suggests that the somatosensory psychophysical channels are not independent. ^{5 6}  

The cornea is innervated by the first branch of the trigeminal nerve. Its afferent information is delivered by the trigeminal nerve to the trigeminal spinal nucleus. The nerve bundles then project to the somatosensory cortex through the thalamus. ^{7 8} Free nerve endings are the only nerve terminals in the corneal epithelium, which is different from the case in the skin. ^{9 10 11} This morphologic difference lead von Frey to his specific nerve and function theory, and the concept that nociception is the only sensation that could be evoked from the human cornea has dominated this field for approximately 100 years. ^{8 12}  

Although there have been challenges of von Frey’s conclusion about corneal nociceptive sensation, ^{13 14 15 16} only recently, with the Belmonte pneumatic esthesiometer, which could deliver mechanical, chemical, and thermal stimuli, has it become possible to characterize systematically the sensation induced by the three submodalities and the sensitivity of the human cornea to these stimuli. ^{17 18 19 20}  

Electrophysiological studies have shown that the corneal nerves respond to thermal, mechanical, and chemical stimulation. ^{12 18 21 22 23 24 25} Molecular biological techniques have shown that several molecular receptors are expressed by trigeminal neurons, ^{26 27 28 29 30 31} demonstrating that different submodalities may be sensed or processed differently at the peripheral level. We believe that this multiple representation of receptors provides the substrate for corneal psychophysical channels. With this idea in mind, we designed this study to investigate the characteristics of human corneal psychophysical channels at threshold. Specifically, experiment I was designed to study the interactions of three submodalities at detection threshold, and experiment II was designed to study whether there is a thermal effect on mechanical sensitivity. 

Methods

Subjects

This study followed the tenets of the Declaration of Helsinki for research involving human subjects and received approval from the University of Waterloo, Office of Research Ethics (Waterloo, Ontario, Canada). The inclusion criteria were: age 20 to 40 years with no history of contact lens wear, and no ocular or systemic disease. 

The Computer-Controlled Belmonte Pneumatic Esthesiometer

The original Belmonte esthesiometer was manually controlled. ^{17 20} Mechanical stimuli are generated by pneumatic flow and the chemical stimuli by the controlled proportion of CO₂ in the air dissolving in the tears to produce carbonic acid thus decreasing tear pH. ²⁵ The air temperature is varied to deliver (calibrated) thermal stimulation of the ocular surface. There are two flow controllers regulating air and CO₂ flow rate and a thermostat controls stimulus temperature. A solenoid is used to control the duration of the pneumatic stimuli. We developed a computerized controller and software that added greater flexibility and ease of use. 

A microcomputer controlled the experiment: It received responses from the subject’s response box and regulated the flow rate and concentration of the CO₂ during the experiment. It also received feedback from the temperature sensor and regulated the stimulus temperature. The distance between the tip of the nozzle and the ocular surface was continually monitored by means of a calibrated signal from a video camera and set at 5 mm. 

Procedure

Measurements were taken on the left eye and, during measurements, subjects were seated comfortably in front of the instrument. The subjects were asked to look at a target using the untested eye, and the investigator positioned the tip of the esthesiometer to stimulate the central cornea. 

Experiment I was designed to measure the mechanical and chemical sensory threshold at different temperatures as well as the quality of the sensation. Five male subjects (30–38 years) participated in this measurement. The method of constant stimuli was used for estimating sensory threshold. The ascending method of limits was used to get an approximate stimulus range, which was then used to set five levels of the stimuli for the method of constant stimuli. The step between each stimulus level was 10 mL/min for mechanical and 10% CO₂ for chemical stimulation, respectively. The stimuli were presented randomly, with each stimulus level presented six times. Stimulus duration was 2 seconds. This stimulus duration was selected because it was long enough to allow particularly the slowly developing chemical sensation to occur, not long enough to induce discomfort while subjects kept their eyes open during stimulus delivery and brief enough so that the method of constant stimuli was not too protracted. 

The threshold was defined as the flow rate or percentage of CO₂ which the subject sensed 50% of the time. The psychometric data were fitted using a logistic function  

\[y{=}1\ {-}\ 1{/}{[}1\ {+}\ (x{/}t)^{S}{]}\]

from which the threshold (t) was estimated. 

The measurements were made at 20°C, 30°C, 40°C, and 50°C. Only two subjects were able to complete the 60°C measurements. Flow thresholds were estimated with 0% CO₂ added as well as with systematically increased (10% step size), randomly presented additional CO₂. For each temperature, therefore, the data were collected in a cube with the z-axis being the percentage correct and the x- and y-axis being flow rate and percentage of CO₂, respectively. 

Experiment II was designed to test thermal effects on mechanical threshold. Four female and 11 male subjects (age range, 20–40 years) participated. The ascending method of limits was used to estimate threshold. The lowest flow rate was set at 20 mL/min. If the stimulus was not detected, the flow rate was increased in steps of 20 mL/min. When the subject felt the stimulus, the intensity was decreased by 20 mL/min and then increased in 10-mL/min steps. If it was detected again, the intensity was recorded. Then the procedure was repeated twice, and the threshold was defined as the average intensity of the three detections. The stimulus duration was 2 seconds. Detection thresholds were measured at 20°C and 50°C. Irritation threshold was measured at 20°C. 

Subjects reported whether the stimulus was detected using a response box. In addition, when subjects reported the stimulus as present, they were required to report the attributes of the sensation. This was noted, and, when required, additional questions were asked by the experimenter to categorize more clearly the information from the subject. 

Statistical Analysis

Repeated measures ANOVA and post hoc tests were used to evaluate the statistical significance in experiment I, set at P < 0.05. Two-tailed paired t-tests with Bonferroni correction were used in experiment II to estimate the significance, set at P < 0.017. 

Results

Experiment I demonstrated that the quality of sensation at threshold changed gradually from cold to irritation (occasionally itching) when the temperature of the air without CO₂ was gradually increased from room temperature to 50°C. The sensation evoked by CO₂ was a stinging or burning pain that could be easily differentiated from the sensation evoked by air without CO₂. When the air was set at 60°C, it evoked a pricking or stinging pain. Figure 1 shows that without CO₂, the detection thresholds at four temperatures gradually increased. Repeated-measures ANOVA showed a significant temperature effect (F_(3,12) = 10.326, P = 0.000). Post hoc tests showed significant differences in thresholds between 20°C and 40°C, 20°C and 50°C, and 30°C and 50°C. 

There was no significant temperature effect on chemical threshold from 20°C to 50°C (Fig. 2 , repeated measures ANOVA F_(3,12) = 0.235, P = 0.870). The chemical threshold, however, was decreased at 60°C (without statistical analysis, as only two subjects could complete this experiment). There were strong interactions when chemical and mechanical stimuli were added (Fig. 3 , repeated measures ANOVA F_(3,12) = 6.407, P = 0.007, Fig. 4 , F_(4,16) = 19.904, P = 0.000). Chemical thresholds were progressively lower when the flow rate increased and mechanical thresholds went down as the percentage of added CO₂ increased. 

When the flow rate was at threshold at 20°C, the burning pain evoked by CO₂ contained a cooling component. This thermal cooling sensation, however, gradually disappeared when the concentration of CO₂ reached higher levels. This is illustrated in Figures 5 and 6 , showing the data of subjects 2 and 5, respectively. 

In experiment II, in addition to detection thresholds at 20°C and 50°C, we measured each subject’s irritation threshold at 20°C. The corneal cold threshold was 56.1 ± 4.9 mL/min (mean ± SE), and mechanical irritation thresholds were 88.0 ± 7.2 mL/min (20°C) and 89.8 ± 6.3 mL/min (50°C) (Fig. 7) . Two-tailed paired t-tests with Bonferroni correction showed that there was no difference between the irritation thresholds at the two temperatures (t ₍₁₄₎= −0.233, P = 0.818). There was, however, a significant difference between cold and irritation thresholds at both temperatures (t ₍₁₄₎ = −5.789, P = 0.000 and t ₍₁₄₎= −5.004, P = 0.000). 

Discussion

The four different kinds of stimuli—mechanical, thermal cold, thermal heat, and chemical CO₂—delivered by the Belmonte pneumatic esthesiometer provide us with an ideal method to test the parallel processing hypothesis in the human cornea. In the present study, at 20°C, the initially detected sensation was described as cool or cold, which is the same as reported by us and others. ^{15 16} This is not surprising, because there is substantial new evidence that the sensation evoked by stimulating the human cornea with room temperature air is one of cold. ^{15 16 19 32 33} This detection is most likely subserved by the reportedly 10% of fibers in the cornea that respond to downward change in temperature in the non-noxious range. ^{18 34} The cold sensation gradually changed to irritation with less report of cold, as the stimulus temperatures increased from 20° to 50°C. This too is consistent with recent evidence that pneumatic stimulation of the cornea excites nociceptors that are either broadly tuned to a variety of stimuli (polymodal nociceptors) or are more specific mechanonociceptors. ^{12 21 22 24 34 35 36} The sensation evoked by CO₂ was stinging or burning pain, as has been reported in other studies. ^{17 18 19 20} The putative neuronal substrate for this is polymodal nociceptors or chemonociceptors. ^{22 23 24 25 35 36}  

The sensory quality change with different submodalities at threshold supports the view that nociceptive sensation is not the only sensation that may be evoked in the human cornea, ^{13 19} and it provides psychophysical evidence that subcomponents of corneal sensory information are sensed and processed differently. If corneal sensory processing were similar to the skin, with mechanical stimuli processed by different psychophysical channels, ^{3 4 37} then corneal sensory information would be processed by different psychophysical channels, which, according to psychophysical channel theory, ¹ possess different sensitivities. This actually was demonstrated by this study in which it was shown that the threshold of cold (20°C) was different from that of irritation (50°C), and it was also supported by the finding that, at 20°C and 50°C, the irritation thresholds were approximately equal, but each was different from the cold threshold. 

In the visual system, different psychophysical channels are supposed to be independent at threshold. ^{1 38} This, however, is not the case in the skin. For example, it has been demonstrated that heat-induced pain may diminish vibrotactile perception, ⁶ and heat may increase the sense of pain. ⁵ To demonstrate whether there is a thermal effect on mechanical sensitivity, experiment II was designed to test mechanical irritation threshold at 20°C and 50°C, respectively. No effect of temperature on mechanical sensitivity was found. 

Corneal chemical sensitivity was not affected by temperature when the stimulus temperature ranged from 20°C to 50°C. This demonstrates that the non-noxious temperature change has no effect on chemical channels. On the other hand, when the air was set at 60°C, the chemical threshold decreased, implying that noxious heat could facilitate the sensing of chemical stimulation. These results are supported by the electrophysiological finding that the cold stimulus would not have reached noxious levels and therefore would not be expected to stimulate polymodal nociceptors, whereas chemical and noxious thermal stimuli could each affect corneal polymodal nociceptors, therefore inducing the chemical and thermal threshold interaction. ^{12 18 21 24 25}  

Another result demonstrating an interaction was that as the flow rate increased, the chemical threshold decreased, and as the concentration of CO₂ increased, the mechanical threshold decreased. This additive effect demonstrates that these two submodalities are not processed independently at threshold. At the neuron level, this is also supported by the observation that both mechanical and chemical stimuli excited corneal polymodal nociceptors. ^{22 23 24} This may also reflect mechanical and chemical receptors interacting with each other at the molecular level and/or an interaction at a more proximal level in the neurons of the spine, thalamus, or cortex. There is an additional possibility that the higher mechanical stimulation caused the greater availability of CO₂ molecules to dissolve in the tears, and this mechanical stimulus may have reduced the thickness of tear film and thus increased the concentration of CO₂. This stimulus-based explanation does not account for the fact that CO₂ affected the mechanical threshold, however. 

Electrophysiological and molecular studies provide evidence of the anatomic and physiological substrates of psychophysical channels. These include different molecular receptors and nerves that respond to different stimuli. Based collectively on these anatomical and physiological experiments and our psychophysical experiments, we propose that the psychophysical corneal innocuous cold channel (ICC) processes cold stimulation mediated by cold menthol receptor type 1 (CMR1)/transient receptor potential menthol 8 (TRPM8) which responds from 8°C to 30°C and is expressed on trigeminal ganglion C fibers ^{30 31} or the complex interaction between ion channels, ³⁹ expressed on the cold neurons, producing the cooling sensation. A corneal noxious heat channel (NHC) mediates thermal pain, and its possible substrate may be the vanilloid receptor (VR)-1, which is expressed on Aδ and C fibers. ²⁶ The corneal chemical channel-H⁺ (CCH⁺) senses hydrogen ion through VR-1 and/or the acid-sensitive ion channels (ASICs) ²⁷ that are expressed on Aδ and C fibers and produce stinging or burning pain, and the corneal itching channel (CIC) that possibly processes histamine and other itch stimuli is itch specific. ^{40 41}  

The irritative sensation evoked by air (without CO₂) is nociceptive, but is different from that of stinging or burning pain caused by CO₂. It may be proposed that this difference occurs because both mechanonociceptors and polymodal nociceptors are recruited by mechanical stimulation, and the joint excitation of these two kinds of sensory nerves may produce a slightly different sensation from that produced by the polymodal nociceptor excitation by the chemical stimulus. ¹⁸ This seems unnecessarily complex and involves a novel interaction between mechano- and polymodal nociceptors never before described. In contrast, a simpler hypothesis, derived from the physiological description of the rabbit cornea is that the chemical stimuli excite predominantly C-fiber chemonociceptors, whereas the mechanical stimuli excite predominantly Aδ driving mechanonociceptors. ^{35 36} This hypothesis is in line with the phenomenology of the sensations inasmuch as chemical stimuli typically burn and mechanical stimuli give rise to foreign-body (sharp) sensations ⁴² as well as the temporal characteristics of these neurons, with C-fiber stimulation being associated with late (“second”) pain and Aδ stimulation occurring more quickly (“first” pain). ⁴³ At the molecular level, the mechanical irritation may be mediated by mechanical receptors rather than VR-1 at the peripheral level. Mechanically sensitive receptors such as the degenerin/epithelial sodium channel (DEG/ENaC) family, which mediate mechanical force transduction and are expressed by trigeminal ganglion Aδ and possibly C fibers, ^{28 29} may be the potential substrate of the corneal mechanical channel (MC). 

The interaction between thermal and chemical and between mechanical and chemical modalities presented by this study demonstrates that these corneal channels are not completely independent at threshold. In experiment I, when the air was set at 20°C, the cold component of the sensation went down if the concentration of CO₂ was high (Figs. 5 6) . The suppression of innocuous cold by CO₂-induced pain may share a mechanism similar to that of heat-induced pain that diminishes touch sensitivity. ⁶ This possibly reflects “gate control theory” at the spinal level in which noxious and innocuous inputs are controlled. ⁴⁴ We could not, however, exclude the possibility that this may happen at the peripheral level, as it has been shown that approximately 50% of CMR1-expressing trigeminal neurons also express VR1, and it has been proposed that cold and heat may elicit different neuronal activities. ³⁰ In addition, when one stimulus contains two components, such as in experiment I, the activity of these two receptors may be modulated by another receptor. 

When the air was set at 60°C, the subjects became more sensitive to CO₂. This may occur at the neuron level by both heat and chemical stimuli exciting corneal polymodal nociceptors because at the molecular receptor level, both heat and H⁺ excite the VR-1 receptor. ^{26 45} Alternatively, it could be that chemical and warm stimuli have a facilitative interaction when they activate separate molecular systems. ⁵ The mechanical and chemical interaction may also imply a similar underlying mechanism. Although ASICs have been demonstrated to be the substrate of hydrogen ion sensing, ²⁷ as members of DEG/ENaC family, they have been proposed to be a mechanoreceptor. ⁴⁶ Furthermore, it has been found that both VR-1 and ASICs are expressed on the same neurons in rat dorsal root ganglia (DRG). ⁴⁷  

It has been shown that cooling (with a temperature decrease of 17°C) inhibits capsaicin-induced currents in rat dorsal root ganglion neurons. ⁴⁸ In addition, strong cooling relieves dermal pain. ⁴⁹ In the present study, neither corneal chemical sensitivity nor mechanical sensitivity was decreased by ocular surface cooling. This possibly implies that decreasing the corneal surface temperature within the limited range, as in this experiment, is insufficient to affect VR1 or mechanical receptors. 

Most natural external stimuli do not contain only one submodality. It seems that corneal nerves represent the sensory information simultaneously and in parallel when external stimuli are presented on the corneal surface—for example, cold, and stinging. However, unlike somatosensory psychophysical channels arising particularly from the skin possessing receptors with different nerve morphology, ^{3 4} corneal nociceptors and cold nerve fibers posses the same morphology (i.e., they are bare terminals). ^{10 11 12 18 22 23 25} Despite this similar microscopic architecture, they give rise to distinct sensations and they therefore must be distinctive in ways not yet revealed using microscopy. 

How may these individual neurons integrate multidimensional sensory information peripherally and transmit these different submodalities to the brain to give rise to different sensations? It is possible that in the periphery, the molecular receptors such as CMR1, VR-1, ASIC, and DEG/ENaC, are expressed on the same neuron and excite it in different dynamic ways, such as has been suggested with CMR1 and VR1. ³⁰ This may be related to the observation that electrophysiologically the rates of discharge of cat corneal nerves increased sharply when the corneal temperature was higher than 44°C, ^{12 14} and this temperature is close to the threshold of noxious thermal pain sensation in humans. In addition, there may be some intermediate neurons or neuronal networks that operate on the neuronal signal by, for example, a lateral inhibitory mechanism. A higher process would then need to decode this multiplexed temporal signal. ⁵⁰ In contrast, at the somatosensory cortex, the afferent sensory information that arose from physically distinct sensory (corneal) receptors may arrive in physically distinct neurons. It then would require higher processing to deal with physically segregated sensory dimensions. 

The human cornea is innervated by Aδ and C fibers and both are believed to convey nociceptive information. ²⁴ In a previous study, we reported that the burning pain evoked by CO₂ was delayed in comparison with the irritation evoked by a purely mechanical stimulus. ²⁰ We have proposed a number of hypothetical reasons for this observation. The more rapidly appearing foreign-body, scratching sensation and the more slowly emerging, stinging sensation evoked by CO₂ suggest that mechanical and chemical stimuli trigger fast and slow pain mechanisms that are mediated by Aδ and C fibers. ⁴³ In addition, the fast and slow perceptions may be attributable to the superficial and deep distribution of these two kinds of nerve fibers. ^{21 35 36} Both of these though, provide physiological evidence of the substrates of the parallel processing of mechanical and chemical stimuli in humans. The delay in the reported burning pain evoked by CO₂, however, may also be related to the time necessary for CO₂ to react in the tears to produce protons. 

In summary, this study demonstrates the interaction of different submodalities at sensory threshold, psychophysically. We propose that there are at least five corneal psychophysical sensory channels, an innocuous cold channel, a noxious heat channel, a nociceptive mechanical channel, a chemical channel⁺, and an itching channel. These channels are not completely independent. The interaction between the different psychophysical channels could be related to the type of corneal receptors stimulated, the relative distribution of these membrane receptors expressed on their respective nerve terminals, and the interaction between the receptors and/or nerves. 

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

 Submitted for publication February 2, 2004; revised March 25, 2004; accepted April 15, 2004.

 Disclosure: Y. Feng, None; T.L. Simpson, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Corneal mechanical threshold and temperature. The higher the temperature of the stimuli, the higher the threshold. Error bars are 95% confidence intervals for the means in each of the relevant figures.

Figure 2.

 

Corneal chemical threshold at different temperatures. Each line represents chemical thresholds at different temperatures.

Figure 3.

 

Corneal chemical thresholds and flow rates (20°C–50°C). As the flow rate increased, the chemical threshold decreased.

Figure 4.

 

The relationship between mechanical threshold and the concentration of CO₂. As the concentration of CO₂ increased, the mechanical threshold decreased (the concentration of CO₂ is the percent of CO₂ of the total volume).

Figure 5.

 

The inhibitory effect of burning pain on cooling sensation (50 mL/min).

Figure 6.

 

The inhibitory effect of burning pain on cooling sensation (60 mL/min).

Figure 7.

 

The effect of temperature on corneal mechanical thresholds. There was no temperature effect (20°C and 50°C) on corneal mechanical irritation thresholds.

The authors thank the subjects who participated in this study and the reviewers for their suggestions.