Temporal Integration of Nasal Irritation from Ammonia at Threshold and Supra-threshold Levels
http://www.100md.com
《毒物学科学杂志》
Monell Chemical Senses Center, Philadelphia, Pennsylvania, 19104–3308
ABSTRACT
Two experiments examined integration of perceived irritation over short-term (100–4000 ms) delivery of ammonia into the nasal cavity of human subjects. Experiment 1 examined trade-offs between time and concentration at threshold level by means of nasal lateralization, a common measure of irritation threshold. Within experimental sessions, the duration of a fixed-concentration stimulus varied to determine the shortest, detectable pulse. Subjects could lateralize increasingly weaker concentrations with longer stimulus presentations. Experiment 2 examined an analogous trade-off for supra-threshold irritation. Subjects rated irritation from presentations of ammonia that varied both in concentration and in duration. Rated intensity for a given concentration increased with stimulus duration. Hence integration occurred at both threshold and supra-threshold levels. However, more than a twofold increase in duration was required to compensate for a twofold decrease in concentration to maintain threshold lateralization or a fixed level of perceived intensity. These results suggest that an imperfect mass-integrator model may be able to describe short-term integration of nasal irritation from ammonia at both the threshold and supra-threshold levels.
INTRODUCTION
We feel airborne chemicals when they interact with free endings of somatosensory nerves (Bryant and Silver, 2000; Doty and Cometto-Muiz, 2003). Sensations, which we term "irritation," include cooling, warming, burning, pungency, and stinging. The eyes, nose, throat, and lower airways prove most sensitive, because mucous membranes offer easier access to nerve endings than the dry, keratinized epithelium of the skin. In the eyes and nose, irritation is mediated by the trigeminal nerve (Bryant and Silver, 2000; Doty and Cometto-Muiz, 2003).
Government regulators view irritation as a material impairment of health and set many occupational exposure limits based on irritation (Cain, 1996; NIOSH, 1994). Accordingly, irritation has
One area where data are scarce is temporal integration in detection of nasal irritation. Sensory scientists use the term "temporal integration" to indicate that systems can integrate stimulus energy over time to detect weaker signals than they otherwise could (Baumgardt, 1972; Garner and Miller, 1947). The term "time-concentration trading" is also used to indicate that one can often achieve a given sensory effect either by presenting a relatively weak stimulus for a long time or a strong stimulus for a shorter duration. Because of integration, it is necessary to study the domain of time to fully understand any sensory system.
Investigators have demonstrated integration of nasal irritation in the supra-threshold range— i.e., for clearly detectable concentrations (see Frasnelli et al., 2003; Hummel, 2000; Hummell et al., 2003). Perceived irritation grows with stimulus duration. This can be seen both over the course of seconds (Anton et al., 1992; Cometto-Muiz and Cain, 1984; Wise et al., 2003) and over the course of minutes (Cain et al., 1986; Hempel-Jorgensen et al., 1999). Work at threshold level has been more limited, but one study examined short-term integration in nasal lateralization of carbon dioxide (Wise et al., 2004).
In nasal lateralization, subjects simultaneously receive chemical vapor in one nostril and clean air in the other. They try to determine which nostril
Wise and colleagues measured the minimum stimulus duration required for reliable lateralization of fixed concentrations (10–65%) of CO2 (Wise et al., 2004). Fixed-ratio increases in stimulus duration compensated for fixed-ratio decreases in concentration to maintain a constant level of lateralization, at least for durations up to 2.5 s (Wise et al., 2004). This finding suggests a model in which detection depends in some simple way on total mass delivered to the nose, i.e., the product of concentration (C) and time (T). Haber's rule, which states that C multiplied by T equals a constant for a fixed outcome, is a commonly used model of this type (see Miller et al., 2000).
However, for lateralization of CO2, a 3.4-fold increase in duration was necessary to compensate for a 2-fold decrease in concentration. This finding suggests an imperfect mass-integrator, consistent with a more general form of Haber's rule: CnT = k (see Bliss, 1940; Miller et al., 2000). Solving for T gives T = kC–n. The equation is a power function, so plotting stimulus duration versus concentration in log-log coordinates makes it possible to fit data with the following linear equation:
where Th represents the threshold stimulus duration needed for reliable lateralization, C represents concentration, and K is a constant. The slope, n, characterizes degree of integration. A slope of –1 indicates perfect integration. A steeper slope, i.e., less than –1, indicates imperfect integration. This model provided an excellent description of lateralization of CO2 over the range of concentrations studied (Wise et al., 2004).
Carbon dioxide stimulates nerve endings through tissue acidification (Hummel, 2000; Shusterman and Avila, 2003). Because liberation of H+ depends on carbonic anhydrase (reviewed in Tarun et al., 2003), CO2 may be unusual. Experiment 1 in the current report begins to test the hypothesis that a simple but imperfect mass integrator describes temporal integration for lateralization of other irritants. The alkaline compound ammonia (NH3), a common irritant in industrial and agricultural settings (Korhonen et al., 2004; Omland, 2002; Proctor et al., 1988), serves as a model stimulus.
An older study examined temporal integration for ratings of supra-threshold nasal irritation from NH3 (Cometto-Muiz and Cain, 1984). Subjects sniffed NH3 (47 to 434 ppm) for brief durations (1.25–3.75 s). As in the threshold-level study using CO2, Cometto-Muiz and Cain (1984) found simple but imperfect integration. This finding suggests that similar rules may govern short-term integration at both threshold levels and supra-threshold levels. In contrast to the threshold-level study of CO2, supra-threshold integration of irritation from NH3 was nearly perfect, i.e., a 2.1-fold increase in duration would compensate for a 2-fold decrease in concentration. Integration may be more nearly perfect for NH3 than for CO2. Integration may also be more nearly perfect for supra-threshold irritation. However, the fact that the two studies differed substantially in methodology makes comparisons problematic. Experiment 2 used the same apparatus as Experiment 1 to vary both concentration and duration of supra-threshold NH3 delivered to the nose. Subjects rated nasal irritation, and integration was assessed both by examining how ratings for given concentration increased with stimulus duration, and how concentration and time could be traded to maintain a fixed level of intensity.
In both experiments, a simple but imperfect mass-integrator model described integration for exposures up to 3–4 s reasonably well. Integration over such brief periods, i.e., integration that might occur within a single natural breath, may have limited immediate significance in predicting whether a given stimulus will cause irritation under natural conditions. Increased understanding of short-term integration may have more immediate relevance for interpretation of data from brief exposures in the laboratory (see General Discussion). Regardless, a model that provides such a concise description of the relationship between stimulus and sensation, based on measurements of just a few points on the integration function, is a potentially powerful tool for basic research.
EXPERIMENT 1
Experiment 1 examined the trade-off between concentration and stimulus duration in detection of nasal irritation from NH3.
Materials and Methods
Subjects.
Approval came from the institutional review board (IRB) of the University of Pennsylvania. Subjects provided written, informed consent on IRB approved forms prior to any manipulations. Four men (ages 26–52 years) and two women (ages 27 and 28 years) participated. Two of the men were authors P.W. and T.C. (subjects 1 and 6, respectively, in Figure 2). Other subjects were paid volunteers. Like all subjects, the authors were blind to any conditions that might cue responses trial-by-trial. Data from the authors resembled data from other subjects.
Olfactometer.
An olfactometer injected NH3 vapor into the nose. Injections of vapor were embedded in a steady background stream of air. Accordingly, stimulus onset caused little or no change in temperature, pressure, and humidity. Embedded stimuli allowed a controlled focus on the effects of chemical stimulation of somatosensory nerves.
Two parallel pressure sources generated flow (Fig. 1, top). The first was an air pump. Flow from the pump was dried, carbon-filtered, and then re-humidified by bubbling it through distilled water in a warm, temperature-controlled enclosure. The pump generated a steady background flow, which entered each nostril at 35°C, 85% RH, and 5 l/min. The pump also generated an air flow (4.7 l/min) used to dilute ammonia vapor from the second pressure source.
The second pressure source was compressed, medical-grade nitrogen. Nitrogen bubbled through odor vessels (two per nostril) at 0.3 l/min. For each nostril, one odor vessel contained an aqueous NH3 solution. The other odor vessel contained water. Nitrogen was chosen instead of air to flow through odor vessels to reduce oxidization. This precaution kept methods consistent with those of planned studies of other compounds, for which control of oxidation may be very important. Stimulus concentration was varied by changing the liquid-phase concentration of NH3 in the odor vessels. The solute was 30% ammonium hydroxide (reagent grade, Sigma-Aldrich). The solvent was Millipore-filtered water. Liquid-phase dilutions ranged from 0.5% to 8% (v/v). Experimenters prepared NH3 solutions under hoods, and they wore protective clothing to reduce the risk of exposure. Subjects were not exposed to liquid solutions.
Nitrogen that passed through an NH3 solution joined with 4.7 l/min of air from the pump to comprise a target, i.e., the stimulus to be lateralized. Nitrogen that passed through water joined with air to comprise a blank. Three-way solenoid values determined whether odorized or blank nitrogen joined with air from the pump, i.e., whether the nostril in question
Both the stimulus flow and the background flow passed to a switching mechanism located close to the nose (Fig. 1, bottom). The mechanism consisted of four 3-way solenoid-valves (two per nostril) with 15 ms response time. For each nostril, stimulus flow (ammonia vapor or blank) entered one valve and vented from the room. Background flow entered the other valve and flowed into the nostril. Energizing both valves simultaneously redirected the stimulus flow into the nostril and the background flow to the vent. Analog signals from a computer (PCI-6023E DAQ card; National Instruments; Austin, TX) provided switching signals with millisecond accuracy. Needle valves (not shown) equalized back pressure between nostril tubes and vent lines to minimize pressure transients. The path from the valves to the end of the tube in the nostril had a volume of 0.18 ml (2.2 ms transit-time at 5 l/min).
Stimuli were injected through flexible Tygon tubes (4 mm outer diameter) extending approximately 0.75 cm into the nostril. Flow exited the nostril around the tubing. Subjects practiced velopharyngeal closure during stimulation. This breathing technique isolates the nasal cavity from the rest of the upper and lower airways using the soft palate (Kobal and Hummel, 1991). Closure helps minimize exposure by preventing inhalation of vapor, and it also prevents fluctuations in pressure in the nasal cavity from respiration.
Experimenters checked flow rate (Gillibrator 2 flow meter; Gillian Instrument Corp.; Wayne, NJ), humidity (Digitron 2020R hygrometer; Topac Instruments; Hingham, MA), and temperature (BAT-12 thermocouple reader; Physiotemp Instruments; Clifton, NJ) at the output of the olfactometer after the device warmed up, and they rechecked flow rate periodically throughout the day. A fast-response pressure transducer (CyQ line, custom made; Cybersense; Nicholasville, KY) verified that minimal changes in flow occurred with switching between background and stimulus. A photo-ionization detector (MiniRAE 2000; RAE Systems; Sunnyvale, CA) measured vapor-phase concentration at the output of the olfactometer. Experimenters used the resulting calibration curve, viz., ppm at output versus liquid-phase concentration in odor vessels, to calculate vapor-phase concentrations. The following concentrations (in ppm) were used: 37, 48, 52, 67, 98, 131, 205, 289, and 721.
Procedure.
Subjects began a trial by placing the tubes in their nostrils, establishing velopharyngeal closure, and clicking a computer mouse. The mouse click began a 10-s countdown. The last three seconds of the countdown were accompanied by beeps, after which the computer triggered a stimulus-presentation of variable duration. The nostril that
Within runs, subjects
Concentrations ranged from the lowest each subject could reliably lateralize with pulses no longer than 10 s, to the highest concentration each subject could lateralize with presentations no briefer than 0.1 s. Concentrations for individuals were selected based on practice runs at 37, 52, 67, 97, 131, and 289 ppm. Subjects who lateralized at 37 ppm were also tested at lower concentrations. Below 37 ppm all subjects failed to lateralize. Based on practice runs, the highest concentration S1
Data analysis.
To further reduce the risk of spurious thresholds, only thresholds at or below the duration where subjects first achieved four consecutive correct responses counted; runs that failed to meet this criterion were repeated. Thresholds for each concentration were estimated by averaging the last five reversals for each run and averaging the results across runs. Threshold pulse duration was plotted versus concentration in log-log coordinates for each subject. Mass-integrator models (linear functions) were fit to the resulting curves by least-squares regression (see introduction).
Results
Figure 2 shows plots of threshol pulse duration versus concentration. Subjects could lateralize increasingly weaker concentrations if stimulus duration increased. Reliable lateralization failed at concentrations below 37 to 67 ppm, depending on the individual, even for long pulses (greater than 10 s). Threshold pulse duration for the lowest detectable concentration ranged from 1324 to 3840 ms, depending on the subject (average = 2691 ms). In log-log coordinates, linear functions accounted for 86–96% (mean = 91%) of the variance in thresholds. Geometric mean slope (calculated across subjects using absolute values) equaled –1.30 (95% confidence interval from –1.03 to –1.64). Averaged across subjects, it required an increase in duration of about 2.5-fold to compensate for a 2-fold decrease in concentration.
Discussion
We note three main features of the data. First, the finding that subjects could lateralize increasingly weaker pulses as duration increased demonstrates that the nasal trigeminal system can integrate NH3 over time at threshold level. Second, the finding that linear functions described time-concentration trading reasonably well suggests that lateralization is related in some simple way to total mass delivered to the nose; i.e., that some form of simple integrator model can describe the data, at least over the range of concentrations studied (but see below). Finally, the finding that slopes of linear functions were, on average, less than –1 suggests that an imperfect mass-integrator model is appropriate. All three findings are in qualitative agreement with both lateralization of carbon dioxide (Wise et al., 2004) and supra-threshold ratings of irritation from NH3 (Cometto-Muiz and Cain, 1984).
With the current method and stimulus, lateralization failed below about 37 ppm, even for pulses as long as 10 s. Based on the threshold pulse durations for the weakest stimuli that subjects could lateralize, the limit of temporal integration may be about 2.7 s. This agrees well with threshold pulse durations for the weakest concentration of CO2 that subjects could lateralize, 2.5 s (Wise et al., 2004). In both cases, one cannot rule out the possibility that less complete integration occurs after about 3 s, such that subjects could lateralize weaker concentrations with much longer pulses. Regardless, it is interesting that the apparent limit of short-term integration seems to coincide with the duration of a single, natural inspiration. Whether this observation has basic significance is unclear.
Considering the data in Figure 2, one might also wonder whether integration begins to break down even sooner than 2.7 s. Linear functions (simple integration model) fit reasonably well. However, long-duration thresholds might deviate from the trend. If thresholds longer than 2000 ms are excluded (dropping the leftmost point for subjects 2, 3, 5, and 6), linear fits accounted for an average of 95% of variance in ratings, as opposed to 91% with all data included. Further, average slope fell to –1.01 (95% CI –.88 to –1.13), suggesting perfect (99%) integration for short-duration (or high-concentration) stimuli. Admittedly, there is no clear justification for excluding the data points in question. However, the possibility that perfect integration occurs over some range for lateralization of NH3 remains open. Denser sampling of low concentrations could help settle the matter.
EXPERIMENT 2
Experiment 2 examined how supra-threshold irritation from NH3 changes as a function of stimulus duration.
Materials and Methods
Subjects.
Approval came from the IRB of the University of Pennsylvania. Subjects provided written, informed consent prior to any manipulations. Participants included 7 men and 13 women, 21 to 54 years of age. One subject (female) failed to complete the study because of time constraints. Her data were excluded. Author P.W. served as one subject. He was blind to the order of conditions. Analyses with and without P.W.'s data supported the same conclusions. Others were paid volunteers.
Olfactometer.
The olfactometer and most details of stimulus presentation matched those used in Experiment 1. The olfactometer was modified to present more than one concentration within a session. This was done by adding more odorant solutions with solenoid valves to select which concentration joined the humid air stream. Flow through all channels was constant, and solutions were replaced after every run to reduce depletion.
Presentation was bi-rhinal, i.e., identical concentrations of NH3 were delivered to both nostrils. The results of Experiment 1 were also based on both nostrils, but only one nostril
Procedure.
Subjects rated the strength of sensations via magnitude estimation. Subjects
Subjects
One duration of each concentration served as a modulus: 2276 ms of 165 ppm for six subjects, 988 ms of 304 ppm for six subjects, and 510 ms of 478 ppm for seven subjects. According to pilot work, these stimuli roughly matched in intensity and fell close to the middle of the intensity range for all stimuli in the study. Subjects were randomly assigned to a modulus, with the constraint of approximately equal numbers in each group. At the beginning of each session, subjects
Data analysis.
The arithmetic mean summarized replicate ratings within subjects. The geometric mean summarized data across subjects (Stevens, 1975). Next, experimenters fit linear functions, via least squares regression, to plots of log rated intensity versus log stimulus duration. Good linear fits would indicate simple integration, i.e., increasing stimulus duration by a fixed factor would increase intensity by a fixed factor. Unit slope would indicate perfect integration, whereas a shallower slope, i.e., less than 1, would indicate imperfect integration (N.B. this is opposite to the analysis in Experiment 1). This analysis assumes that ratings are perfectly proportional to perceived intensity.
A second analysis assumed only that equal magnitude estimates across two conditions indicated equal perceived intensity (Cometto-Muiz and Cain, 1984). Experimenters calculated the stimulus durations for each concentration needed to achieve fixed levels of perceived intensity. The second analysis is more comparable to that of Experiment 1, where duration and concentration traded to maintain a fixed level of detection. If one plots log stimulus duration versus log concentration, a slope of –1 indicates perfect integration and a slope less than –1 indicates imperfect integration.
Results
We note three main features of the data in Figure 3. First, rated intensity increased with duration for each concentration. Second, linear functions (in log-log coordinates) described the relationship between duration and rated intensity quite well, accounting for 97.5–99.9% of the variance in ratings. To a first approximation, a fixed-ratio increase in stimulus duration at a given concentration produced a fixed-ratio increase in rated irritation. Finally, slopes (with 95% confidence intervals based on regression) were 0.57 (0.29–0.85) for 165 ppm, 0.77 (0.68–0.85) for 304 ppm, and 0.82 (0.62–1.02) for 478 ppm. Geometric mean slope, across concentrations, was 0.71 (0.56–0.91). Averaged across concentrations, doubling duration at fixed concentration increased intensity by a factor of 1.65.
By interpolating between the points in Figure 3 using the linear fits, experimenters calculated the durations for each concentration needed to produce four fixed levels of perceived intensity. The resulting (iso-intensity) curves, in log-log coordinates, appear in Figure 4. Linear functions described the curves quite well. Slopes ranged from –1.47 to –1.81 for the range of perceived intensities that the three concentrations had in common. Slopes appeared to increase with perceived intensity. Average (across the common intensity-range) slope was –1.64., indicating that a 3.1-fold increase in duration was required to compensate for a 2-fold decrease in concentration to maintain a fixed level of perceived irritation.
Discussion
The finding that ratings of irritation increased with stimulus duration for a fixed concentration demonstrates that integration occurred (Fig. 3). The finding that fixed-ratio increases in duration led to fixed-ratio increases in ratings suggests that rated intensity is related in some simple way to total mass delivered to the nose, i.e., that some form of simple mass-integrator model can describe the data. The finding that iso-intensity curves in Figure 4 were also approximately linear supports the same conclusion. Finally, the finding that slopes of intensity versus duration functions were less than 1 (Fig. 3), and the finding that slopes of iso-intensity curves (Fig. 4) were less than –1, suggests that an imperfect mass-integrator model is appropriate. All three findings are in good qualitative agreement with lateralization of carbon dioxide (Wise et al., 2004), lateralization of NH3 (Experiment 1), and an older study of supra-threshold ratings of irritation from NH3 (Cometto-Muiz and Cain, 1984).
Inspection of Figure 3 reveals another point of agreement between the results of Experiment 2 and those of Cometto-Muiz and Cain (1984). In both experiments, slopes of intensity versus stimulus-duration functions went up as concentration increased. All in all, qualitative agreement among the studies is striking.
The results of Experiment 2 disagree with those of Cometto-Muiz and Cain (1984) in degree of integration. Experiment 2 suggests that a 3.1-fold increase in duration is required to compensate for a 2-fold decrease in concentration to maintain a fixed level of perceived intensity, whereas the older study suggested that a 2.1-fold increase in duration would suffice. Accordingly, the two supra-threshold studies of irritation from NH3 disagree with each other about as strongly as the findings of Cometto-Muiz and Cain (1984) disagree with the Wise et al. (2004) results on lateralization of CO2 (which suggested that a 3.4-fold increase in concentration is required to compensate for a 2-fold decrease in concentration). This finding strongly suggests that differences between the two older studies are not exclusively stimulus driven. If the differences between the older studies were instead driven largely by differences between threshold and supra-threshold integration, then one would expect better integration in Experiment 2 than in Experiment 1. In other words, given the same stimulus and similar methods, we would expect supra-threshold integration to be more perfect than threshold-level integration. Results of Experiment 1, which suggested that a 2.5-fold increase in duration is needed to compensate for a 2-fold decrease in concentration, actually came closer to perfect integration than did those of Experiment 2. In short, the results suggest that neither differences between NH3 and CO2 nor differences between threshold-level and supra-threshold level integration can completely account for differences in degree of integration seen in previous studies.
Some disagreement probably comes from methodological differences. In the present experiments, subjects
GENERAL DISCUSSION
Research on temporal integration in detection of nasal irritation has been scant. Wise and colleagues showed that an imperfect, mass-integrator model provides a good description of short-term integration in lateralization of CO2 (Wise et al., 2004). Carbon dioxide probably stimulates nerve endings through tissue acidification (Hummel, 2000; Shusterman and Avila, 2003). The current report extends the previous finding by demonstrating simple but imperfect integration in lateralization of the base, NH3. Furthermore, in agreement with a previous report (Cometto-Muiz and Cain, 1984), the current experiments suggest that an imperfect, mass-integrator model can also describe short-term integration for supra-threshold irritation from NH3. Studies of other acids and bases can determine how well the findings generalize. Studies of still other compounds, e.g., nicotine and menthol, can determine whether an imperfect, mass-integrator model can describe irritation from compounds that do not stimulate through changes in pH.
Limitations
The simple integration model is of the black-box type, and it includes all events from entry of the stimulus into the nostril to execution of the response. Dynamics of the generated stimulus will strongly influence dynamics of concentration in the peri-receptor environment, i.e., in epithelial or sub-epithelial layers of the mucosa (Finger et al., 1990). However, patterns of flow and diffusion through the nasal cavity and subsequent diffusion into the tissue will also influence dynamics of peri-receptor concentration. Online tracking of peri-receptor concentration, perhaps through a pH meter (Shusterman and Avila, 2003) or pH-sensitive dye, may help elucidate this component of the black box (see Wise et al., 2004, for more discussion). Psychophysiology, e.g., combinations of psychophysics and measurement of mucosal or cortical evoked potentials (Hummel, 2000; Hummel et al., 2003), may help elucidate other components of the black box.
Another limitation is instrumental, because the olfactometer could not reliably produce pulses briefer than about 100 ms. Perhaps integration would behave differently for very brief stimuli. The visual system, for example, displays perfect integration up to about 100 ms (200 ms for rod vision), and imperfect integration up to as long as 3 s (Baumgardt, 1972).
Finally, the stimulation technique, i.e., passive injection of vapor into the nose during velopharyngeal closure, is un-physiological. To elucidate basic response properties of the sensory system, the rigorous experimental control of the method has advantages. However, experimental conditions differed from natural breathing. For example, experimental flow rates fell below those of normal breathing. Further, stimulus flow was probably confined mostly to the nose itself, rather than being distributed throughout the nasal cavity. Even within the nose, patterns of flow almost certainly differed from those associated with natural breathing. Given the considerable body of literature which suggests that rates and patterns of flow strongly influence deposition and absorption of volatile compounds in the nasal cavity (e.g., Frederick et al., 1994, 1998; Kurtz et al., 2004; Morris, 2001), one should exercise caution in generalizing the results of the current studies to more natural conditions. Additional studies using natural breathing techniques would be useful in this regard.
Basic Significance
Simple but imperfect integration is potentially consistent with a variety of mechanisms (Cain, 1990; Wise et al., 2004). For example, integration may come from a build-up of stimulus molecules in the mucosa over time, but the build-up could be undermined by continual clearance or breakdown of molecules. Adaptation could also undermine perfect integration. Further, both unmyelinated C-fibers and thinly myelinated A-fibers in the nose can respond to chemical irritants (see Bryant and Silver, 2000). The two types of fibers have different temporal response properties and may give rise to different sensations, i.e., burning versus stinging (see Hummel, 2000). Some effects of varying stimulus duration could come from sequential stimulation of the two populations of fibers.
As suggested above under Limitations, it will require more than psychophysics to relate patterns of perception to specific mechanisms. However, to understand the relationship between physiology and sensation, it is necessary to measure sensation. The psychophysical models described above, or extensions thereof, can guide biophysicists, molecular biologists, and physiologists in the quest to elucidate the mechanisms of perceived irritation.
Practical Significance
Short-term integration, i.e., integration that might occur within the duration of a single, natural inspiration, may have limited applicability to integration in natural settings. However, increased understanding of short-term integration can have immediate implications for how data are collected and interpreted in the laboratory. In the laboratory, investigations of sensitivity to irritants, both for patients and for normal controls, are often based on presentations that last between.25 and 3 s (Hummel, 2000; Hummel et al., 2003; Shusterman, 2002). Often, subjects simply sniff from bottles, and duration is uncontrolled. As a result of integration, individual differences in sniff duration can become confounded with differences in sensitivity. One can control stimulus duration through olfactometery. Still, because of individual differences in slopes of integration functions (see Wise et al., 2004, and Figure 1 above), stimuli of different lengths may provide very different pictures of individual differences. Furthermore, much of the growing body of literature on the how irritant potency is related to molecular parameters is based on brief exposures in the laboratory (e.g., Abraham et al., 1998, 2003). If compounds differ in the slopes of their integration functions, sampling at a fixed duration will provide at best an incomplete picture of differences among compounds. Simple models that allow researchers to characterize entire time-concentration trading functions by measuring a few carefully selected points can be powerful tools that give researchers a better understanding of their data.
NOTES
The authors certify that all research involving human subjects was done under full compliance with all government policies and the Helsinki Declaration.
ACKNOWLEDGMENTS
Financial support came in part from grants P50 DC00214 and T32 DC00014 from the National Institute on Deafness and other Communication Disorders of the National Institutes of Health. We thank three anonymous reviewers for helpful comments. Conflict of interest: none declared.
REFERENCES
Abraham, M. H., Hassanisadi, M., Jalali-Heravi, M., Ghafourian, T., Cain, W. S., and Cometto-Muniz, J. E. (2003). Draize rabbit eye test compatibility with eye irritation thresholds in humans: A quantitative structure–activity relationship analysis. Toxicol. Sci., 76, 384–391.
Abraham, M. H., Kumarsingh, R., Cometto-Muiz, J. E., and Cain, W. S. (1998). An algorithm for nasal pungency thresholds in man. Arch. Toxicol. 72, 227–232.
Anton, F., Euchner, I., and Handwerker, H. O., (1992). Psychophysical examination of pain induced by defined CO2 pulses applied to the nasal mucosa. Pain 49, 53–60.
Baumgardt, E. (1972). Threshold quantal problems. In Visual Psychophysics (D. Jameson and J. Hurvich, Eds.), pp 29–55. Springer-Verlag, New York.
Bliss, C. I. (1940). The relation between exposure time, concentration and toxicity in experiments on insecticides. Ann. Entomol. Soc. Am. 33, 721–766.
Bryant, B. P., and Silver, W. L. (2000). Chemesthesis: The common chemical sense. In The Neurobiology of Taste and Smell (T. E. Finger, W. L. Silver, and D. Restrepo, Eds.), 2nd ed., pp. 73–100. Wiley-Liss, New York.
Cain, W. S. (1996). Overview: Odors and irritation in indoor pollution. In Indoor Air and Human Health (R. B. Gammage and B. A. Berven, Eds.), 2nd ed., pp. 23–30. Lewis, Boca Raton, FL.
Cain, W. S. (1990). Perceptual characteristics of nasal irritation. In Chemical Senses (Vol. 2): Irritation (B.G. Green, J. R. Mason, and M. R. Kare, Eds.), pp. 43–58. Marcel Dekker, New York.
Cain, W. S., See, L. C., and Tosun, T. (1986). Irritation and odor from formaldehyde: Chamber studies. In IAQ '86: Managing the Indoor Air for Health and Energy Conservation, pp. 126–137. American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE), Atlanta.
Caterina, M. J., and Julius, D. (2001). The vanilloid receptor: A molecular gateway to the pain pathway. Annu. Rev. Neurosci. 24, 487–517.
Cometto-Muiz, J. E., and Cain, W. S. (1984). Temporal integration of pungency. Chem. Senses 8, 315–327.
Doty, R. L., and Cometto-Muiz, J. E. (2003). Trigeminal chemosensation. In Handbook of Olfaction and Gustation (R. L. Doty, Ed.), 2nd Ed., pp. 981–999. Marcel Dekker, New York.
Doty, R. L., Cometto-Muniz, J. E., Jalowayski, A. A., Dalton, P., Kendal-Reed, M., and Hodgson, M. (2004). Assessment of upper respiratory tract and ocular irritative effects of volatile chemicals in humans. Crit. Rev. Toxicol. 34, 85–142.
Finger, T. E., Getchell, M. L., Getchell, T. V., and Kinnamon, J. C. (1990). Affector and effector functions of peptidergic innervation of the nasal cavity. In Chemical Senses (Vol. 2): Irritation (B. G. Green, J. R. Mason, and M. R. Kare, Eds.), pp. 1–20. Marcel Dekker, New York.
Frasnelli, J., Ltsch, J., and Hummel, T. (2003). Event-related potentials to intranasal trigeminal stimuli change in relation to stimulus concentration and stimulus duration. J. Clin. Neurophysiol. 20, 80–86.
Frederick, C. B., Bush, M. L., Lomax, L. G., Black, K. A., Finch, L., Kimbell, J. S., Morgan, K. T., Subramaniam, R. P., Morris, J. B., and Ultman, J. S. (1998). Application of a hybrid computational fluid dynamics and physiologically based inhalation model for interspecies dosimetry extrapolation of acidic vapors in the upper airways. Toxicol. Appl. Pharmacol. 152, 211–231.
Frederick, C. B., Morris, J. B., Kimbell, J. S., Morgan, K. T., and Scherer, P. T. (1994). Comparison of four biologically based dosimetry models for the deposition of rapidly metabolized vapors in the rodent nasal cavity. Inhal. Toxicol. 6(Suppl.), 135–157.
Garcia-Medina, M. R., and Cain, W. S. (1982). Bilateral integration in the common chemical sense. Physiol. Behav. 29, 349–353.
Garner, W. R., and Miller, G. A. (1947). The masked threshold of pure tones as a function of duration. J. Exp. Psychol. 37, 293–305.
Hempel-Jorgensen, A., Kjaergaard, S. K., Molhave, L., and Hudnell, H. K. (1999). Time course of sensory eye irritation in humans exposed to N-butanol and 1-octene. Arch. Environ. Health 54, 86–94.
Hummel, T. (2000). Assessment of intranasal trigeminal function. Int. J. Psychophysiol. 36, 147–155.
Hummel, T., Mohammadian, P., Marchl, R., Kobal, G., and Ltsch, J. (2003). Irritation of the nasal mucosa using short- and long-lasting painful stimuli. Int. J. Psychophysiol. 47, 147–158.
Kobal, G., and Hummel, T. (1991). Olfactory evoked potentials in humans. In Smell and Taste in Health and Disease (T. V. Getchell, R. L. Doty, L. M. Bartoshuck, and J. B. Snow, Eds.), pp. 255–275. Raven Press, New York.
Kobal, G., Van Toller, S., and Hummel, T. (1989). Is there directional smelling. Experientia 45, 130–132.
Korhonen, K., Liukkonen, T., Ahrens, W., Astrakianakis, G., Boffetta, P., Burdorf, A., Heederik, D., Kauppinen, T., Kogevinas, M., Osvoll, P., et al. (2004). Occupational exposure to chemical agents in the paper industry. Int. Arch. Occup. Environ. Health 77, 451–460.
Kurtz, D. B., Zhao, K., Hornung, D. E., and Scherer, P. (2004). Experimental and numerical determination of odorant solubility in nasal and olfactory mucosa. Chem. Senses 29, 763–773.
Miller, F. J., Schlosser, P. M., and Janszen, D. B. (2000). Haber's rule: A special case in a family of curves relating concentration and duration of exposure to a fixed level of response for a given endpoint. Toxicology 149, 21–34.
Morris, J. B. (2001). Overview of upper respiratory tract vapor uptake studies. Inhal. Toxicol. 13, 335–345.
NIOSH (National Institute for Occupational Safety and Health). (1994). NIOSH Pocket Guide to Chemical Hazards (Publication No. 94–116). National Institute for Occupational Safety and Health, Cincinnati.
Omland, O. (2002). Exposure and respiratory health in farming in temperate zones—A review of the literature. Ann. Agr. Environ. Med. 9, 119–136.
Proctor, N. H., Hughes, J. P., and Fischman, M. L. (1988). Chemical Hazards of the Workplace. J. B. Lippincott, Philadelphia.
Shusterman, D. (2002). Individual factors in nasal chemesthesis. Chem. Senses 27, 551–564.
Shusterman, D., and Avila, P. C. (2003). Real-time monitoring of nasal mucosal pH during carbon dioxide stimulation: Implications for stimulus dynamics. Chem. Senses 28, 595–601.
Stevens, S. S. (1975). Psychophysics: Introduction to Its Perceptual, Neural, and Social Prospects. Wiley, New York.
Tarun, A., Bryant, B., Zhai, W., Soloman, C., and Shusterman, D. (2003). Gene expression for carbonic anhydrase isoenzymes in human nasal mucosa. Chem. Senses 28, 621–629.
Tominaga, M., Caterina, M. J., Malmberg, A. B., Rosen T. A., Gilbert, H., Skinner, K., Raumann, B. E., Basbaum, A. I., and Julius, D. (1998). The cloned capsaicin receptor integrates multiple pain-producing stimuli. Neuron 21, 531–543.
Wetherill, G., and Levitt, H. (1965). Sequential estimation of points on a psychometric function. Br. J. Math. Stat. Psychol. 18, 1–10.
Wise, P. M., Radil., T., and Wysocki, C. J. (2004). Temporal integration in nasal lateralization and nasal detection of carbon dioxide. Chem. Senses 29, 137–142.
Wise, P. M., Wysocki, C. J., and Radil T. (2003). Time–intensity ratings of nasal irritation from carbon dioxide. Chem. Senses 28, 751–760.
Wysocki, C. J., and Wise, P. (2003). Methods, approaches, and caveats for functionally evaluating olfaction and chemesthesis. In Handbook of Flavor Characterization: Sensory, Chemical and Physiological (K. Deibler and J. F. Delwiche, Eds.), pp 1–40. Marcel Dekker, New York.(Paul M. Wise, Thomas M. C)
ABSTRACT
Two experiments examined integration of perceived irritation over short-term (100–4000 ms) delivery of ammonia into the nasal cavity of human subjects. Experiment 1 examined trade-offs between time and concentration at threshold level by means of nasal lateralization, a common measure of irritation threshold. Within experimental sessions, the duration of a fixed-concentration stimulus varied to determine the shortest, detectable pulse. Subjects could lateralize increasingly weaker concentrations with longer stimulus presentations. Experiment 2 examined an analogous trade-off for supra-threshold irritation. Subjects rated irritation from presentations of ammonia that varied both in concentration and in duration. Rated intensity for a given concentration increased with stimulus duration. Hence integration occurred at both threshold and supra-threshold levels. However, more than a twofold increase in duration was required to compensate for a twofold decrease in concentration to maintain threshold lateralization or a fixed level of perceived intensity. These results suggest that an imperfect mass-integrator model may be able to describe short-term integration of nasal irritation from ammonia at both the threshold and supra-threshold levels.
INTRODUCTION
We feel airborne chemicals when they interact with free endings of somatosensory nerves (Bryant and Silver, 2000; Doty and Cometto-Muiz, 2003). Sensations, which we term "irritation," include cooling, warming, burning, pungency, and stinging. The eyes, nose, throat, and lower airways prove most sensitive, because mucous membranes offer easier access to nerve endings than the dry, keratinized epithelium of the skin. In the eyes and nose, irritation is mediated by the trigeminal nerve (Bryant and Silver, 2000; Doty and Cometto-Muiz, 2003).
Government regulators view irritation as a material impairment of health and set many occupational exposure limits based on irritation (Cain, 1996; NIOSH, 1994). Accordingly, irritation has
One area where data are scarce is temporal integration in detection of nasal irritation. Sensory scientists use the term "temporal integration" to indicate that systems can integrate stimulus energy over time to detect weaker signals than they otherwise could (Baumgardt, 1972; Garner and Miller, 1947). The term "time-concentration trading" is also used to indicate that one can often achieve a given sensory effect either by presenting a relatively weak stimulus for a long time or a strong stimulus for a shorter duration. Because of integration, it is necessary to study the domain of time to fully understand any sensory system.
Investigators have demonstrated integration of nasal irritation in the supra-threshold range— i.e., for clearly detectable concentrations (see Frasnelli et al., 2003; Hummel, 2000; Hummell et al., 2003). Perceived irritation grows with stimulus duration. This can be seen both over the course of seconds (Anton et al., 1992; Cometto-Muiz and Cain, 1984; Wise et al., 2003) and over the course of minutes (Cain et al., 1986; Hempel-Jorgensen et al., 1999). Work at threshold level has been more limited, but one study examined short-term integration in nasal lateralization of carbon dioxide (Wise et al., 2004).
In nasal lateralization, subjects simultaneously receive chemical vapor in one nostril and clean air in the other. They try to determine which nostril
Wise and colleagues measured the minimum stimulus duration required for reliable lateralization of fixed concentrations (10–65%) of CO2 (Wise et al., 2004). Fixed-ratio increases in stimulus duration compensated for fixed-ratio decreases in concentration to maintain a constant level of lateralization, at least for durations up to 2.5 s (Wise et al., 2004). This finding suggests a model in which detection depends in some simple way on total mass delivered to the nose, i.e., the product of concentration (C) and time (T). Haber's rule, which states that C multiplied by T equals a constant for a fixed outcome, is a commonly used model of this type (see Miller et al., 2000).
However, for lateralization of CO2, a 3.4-fold increase in duration was necessary to compensate for a 2-fold decrease in concentration. This finding suggests an imperfect mass-integrator, consistent with a more general form of Haber's rule: CnT = k (see Bliss, 1940; Miller et al., 2000). Solving for T gives T = kC–n. The equation is a power function, so plotting stimulus duration versus concentration in log-log coordinates makes it possible to fit data with the following linear equation:
where Th represents the threshold stimulus duration needed for reliable lateralization, C represents concentration, and K is a constant. The slope, n, characterizes degree of integration. A slope of –1 indicates perfect integration. A steeper slope, i.e., less than –1, indicates imperfect integration. This model provided an excellent description of lateralization of CO2 over the range of concentrations studied (Wise et al., 2004).
Carbon dioxide stimulates nerve endings through tissue acidification (Hummel, 2000; Shusterman and Avila, 2003). Because liberation of H+ depends on carbonic anhydrase (reviewed in Tarun et al., 2003), CO2 may be unusual. Experiment 1 in the current report begins to test the hypothesis that a simple but imperfect mass integrator describes temporal integration for lateralization of other irritants. The alkaline compound ammonia (NH3), a common irritant in industrial and agricultural settings (Korhonen et al., 2004; Omland, 2002; Proctor et al., 1988), serves as a model stimulus.
An older study examined temporal integration for ratings of supra-threshold nasal irritation from NH3 (Cometto-Muiz and Cain, 1984). Subjects sniffed NH3 (47 to 434 ppm) for brief durations (1.25–3.75 s). As in the threshold-level study using CO2, Cometto-Muiz and Cain (1984) found simple but imperfect integration. This finding suggests that similar rules may govern short-term integration at both threshold levels and supra-threshold levels. In contrast to the threshold-level study of CO2, supra-threshold integration of irritation from NH3 was nearly perfect, i.e., a 2.1-fold increase in duration would compensate for a 2-fold decrease in concentration. Integration may be more nearly perfect for NH3 than for CO2. Integration may also be more nearly perfect for supra-threshold irritation. However, the fact that the two studies differed substantially in methodology makes comparisons problematic. Experiment 2 used the same apparatus as Experiment 1 to vary both concentration and duration of supra-threshold NH3 delivered to the nose. Subjects rated nasal irritation, and integration was assessed both by examining how ratings for given concentration increased with stimulus duration, and how concentration and time could be traded to maintain a fixed level of intensity.
In both experiments, a simple but imperfect mass-integrator model described integration for exposures up to 3–4 s reasonably well. Integration over such brief periods, i.e., integration that might occur within a single natural breath, may have limited immediate significance in predicting whether a given stimulus will cause irritation under natural conditions. Increased understanding of short-term integration may have more immediate relevance for interpretation of data from brief exposures in the laboratory (see General Discussion). Regardless, a model that provides such a concise description of the relationship between stimulus and sensation, based on measurements of just a few points on the integration function, is a potentially powerful tool for basic research.
EXPERIMENT 1
Experiment 1 examined the trade-off between concentration and stimulus duration in detection of nasal irritation from NH3.
Materials and Methods
Subjects.
Approval came from the institutional review board (IRB) of the University of Pennsylvania. Subjects provided written, informed consent on IRB approved forms prior to any manipulations. Four men (ages 26–52 years) and two women (ages 27 and 28 years) participated. Two of the men were authors P.W. and T.C. (subjects 1 and 6, respectively, in Figure 2). Other subjects were paid volunteers. Like all subjects, the authors were blind to any conditions that might cue responses trial-by-trial. Data from the authors resembled data from other subjects.
Olfactometer.
An olfactometer injected NH3 vapor into the nose. Injections of vapor were embedded in a steady background stream of air. Accordingly, stimulus onset caused little or no change in temperature, pressure, and humidity. Embedded stimuli allowed a controlled focus on the effects of chemical stimulation of somatosensory nerves.
Two parallel pressure sources generated flow (Fig. 1, top). The first was an air pump. Flow from the pump was dried, carbon-filtered, and then re-humidified by bubbling it through distilled water in a warm, temperature-controlled enclosure. The pump generated a steady background flow, which entered each nostril at 35°C, 85% RH, and 5 l/min. The pump also generated an air flow (4.7 l/min) used to dilute ammonia vapor from the second pressure source.
The second pressure source was compressed, medical-grade nitrogen. Nitrogen bubbled through odor vessels (two per nostril) at 0.3 l/min. For each nostril, one odor vessel contained an aqueous NH3 solution. The other odor vessel contained water. Nitrogen was chosen instead of air to flow through odor vessels to reduce oxidization. This precaution kept methods consistent with those of planned studies of other compounds, for which control of oxidation may be very important. Stimulus concentration was varied by changing the liquid-phase concentration of NH3 in the odor vessels. The solute was 30% ammonium hydroxide (reagent grade, Sigma-Aldrich). The solvent was Millipore-filtered water. Liquid-phase dilutions ranged from 0.5% to 8% (v/v). Experimenters prepared NH3 solutions under hoods, and they wore protective clothing to reduce the risk of exposure. Subjects were not exposed to liquid solutions.
Nitrogen that passed through an NH3 solution joined with 4.7 l/min of air from the pump to comprise a target, i.e., the stimulus to be lateralized. Nitrogen that passed through water joined with air to comprise a blank. Three-way solenoid values determined whether odorized or blank nitrogen joined with air from the pump, i.e., whether the nostril in question
Both the stimulus flow and the background flow passed to a switching mechanism located close to the nose (Fig. 1, bottom). The mechanism consisted of four 3-way solenoid-valves (two per nostril) with 15 ms response time. For each nostril, stimulus flow (ammonia vapor or blank) entered one valve and vented from the room. Background flow entered the other valve and flowed into the nostril. Energizing both valves simultaneously redirected the stimulus flow into the nostril and the background flow to the vent. Analog signals from a computer (PCI-6023E DAQ card; National Instruments; Austin, TX) provided switching signals with millisecond accuracy. Needle valves (not shown) equalized back pressure between nostril tubes and vent lines to minimize pressure transients. The path from the valves to the end of the tube in the nostril had a volume of 0.18 ml (2.2 ms transit-time at 5 l/min).
Stimuli were injected through flexible Tygon tubes (4 mm outer diameter) extending approximately 0.75 cm into the nostril. Flow exited the nostril around the tubing. Subjects practiced velopharyngeal closure during stimulation. This breathing technique isolates the nasal cavity from the rest of the upper and lower airways using the soft palate (Kobal and Hummel, 1991). Closure helps minimize exposure by preventing inhalation of vapor, and it also prevents fluctuations in pressure in the nasal cavity from respiration.
Experimenters checked flow rate (Gillibrator 2 flow meter; Gillian Instrument Corp.; Wayne, NJ), humidity (Digitron 2020R hygrometer; Topac Instruments; Hingham, MA), and temperature (BAT-12 thermocouple reader; Physiotemp Instruments; Clifton, NJ) at the output of the olfactometer after the device warmed up, and they rechecked flow rate periodically throughout the day. A fast-response pressure transducer (CyQ line, custom made; Cybersense; Nicholasville, KY) verified that minimal changes in flow occurred with switching between background and stimulus. A photo-ionization detector (MiniRAE 2000; RAE Systems; Sunnyvale, CA) measured vapor-phase concentration at the output of the olfactometer. Experimenters used the resulting calibration curve, viz., ppm at output versus liquid-phase concentration in odor vessels, to calculate vapor-phase concentrations. The following concentrations (in ppm) were used: 37, 48, 52, 67, 98, 131, 205, 289, and 721.
Procedure.
Subjects began a trial by placing the tubes in their nostrils, establishing velopharyngeal closure, and clicking a computer mouse. The mouse click began a 10-s countdown. The last three seconds of the countdown were accompanied by beeps, after which the computer triggered a stimulus-presentation of variable duration. The nostril that
Within runs, subjects
Concentrations ranged from the lowest each subject could reliably lateralize with pulses no longer than 10 s, to the highest concentration each subject could lateralize with presentations no briefer than 0.1 s. Concentrations for individuals were selected based on practice runs at 37, 52, 67, 97, 131, and 289 ppm. Subjects who lateralized at 37 ppm were also tested at lower concentrations. Below 37 ppm all subjects failed to lateralize. Based on practice runs, the highest concentration S1
Data analysis.
To further reduce the risk of spurious thresholds, only thresholds at or below the duration where subjects first achieved four consecutive correct responses counted; runs that failed to meet this criterion were repeated. Thresholds for each concentration were estimated by averaging the last five reversals for each run and averaging the results across runs. Threshold pulse duration was plotted versus concentration in log-log coordinates for each subject. Mass-integrator models (linear functions) were fit to the resulting curves by least-squares regression (see introduction).
Results
Figure 2 shows plots of threshol pulse duration versus concentration. Subjects could lateralize increasingly weaker concentrations if stimulus duration increased. Reliable lateralization failed at concentrations below 37 to 67 ppm, depending on the individual, even for long pulses (greater than 10 s). Threshold pulse duration for the lowest detectable concentration ranged from 1324 to 3840 ms, depending on the subject (average = 2691 ms). In log-log coordinates, linear functions accounted for 86–96% (mean = 91%) of the variance in thresholds. Geometric mean slope (calculated across subjects using absolute values) equaled –1.30 (95% confidence interval from –1.03 to –1.64). Averaged across subjects, it required an increase in duration of about 2.5-fold to compensate for a 2-fold decrease in concentration.
Discussion
We note three main features of the data. First, the finding that subjects could lateralize increasingly weaker pulses as duration increased demonstrates that the nasal trigeminal system can integrate NH3 over time at threshold level. Second, the finding that linear functions described time-concentration trading reasonably well suggests that lateralization is related in some simple way to total mass delivered to the nose; i.e., that some form of simple integrator model can describe the data, at least over the range of concentrations studied (but see below). Finally, the finding that slopes of linear functions were, on average, less than –1 suggests that an imperfect mass-integrator model is appropriate. All three findings are in qualitative agreement with both lateralization of carbon dioxide (Wise et al., 2004) and supra-threshold ratings of irritation from NH3 (Cometto-Muiz and Cain, 1984).
With the current method and stimulus, lateralization failed below about 37 ppm, even for pulses as long as 10 s. Based on the threshold pulse durations for the weakest stimuli that subjects could lateralize, the limit of temporal integration may be about 2.7 s. This agrees well with threshold pulse durations for the weakest concentration of CO2 that subjects could lateralize, 2.5 s (Wise et al., 2004). In both cases, one cannot rule out the possibility that less complete integration occurs after about 3 s, such that subjects could lateralize weaker concentrations with much longer pulses. Regardless, it is interesting that the apparent limit of short-term integration seems to coincide with the duration of a single, natural inspiration. Whether this observation has basic significance is unclear.
Considering the data in Figure 2, one might also wonder whether integration begins to break down even sooner than 2.7 s. Linear functions (simple integration model) fit reasonably well. However, long-duration thresholds might deviate from the trend. If thresholds longer than 2000 ms are excluded (dropping the leftmost point for subjects 2, 3, 5, and 6), linear fits accounted for an average of 95% of variance in ratings, as opposed to 91% with all data included. Further, average slope fell to –1.01 (95% CI –.88 to –1.13), suggesting perfect (99%) integration for short-duration (or high-concentration) stimuli. Admittedly, there is no clear justification for excluding the data points in question. However, the possibility that perfect integration occurs over some range for lateralization of NH3 remains open. Denser sampling of low concentrations could help settle the matter.
EXPERIMENT 2
Experiment 2 examined how supra-threshold irritation from NH3 changes as a function of stimulus duration.
Materials and Methods
Subjects.
Approval came from the IRB of the University of Pennsylvania. Subjects provided written, informed consent prior to any manipulations. Participants included 7 men and 13 women, 21 to 54 years of age. One subject (female) failed to complete the study because of time constraints. Her data were excluded. Author P.W. served as one subject. He was blind to the order of conditions. Analyses with and without P.W.'s data supported the same conclusions. Others were paid volunteers.
Olfactometer.
The olfactometer and most details of stimulus presentation matched those used in Experiment 1. The olfactometer was modified to present more than one concentration within a session. This was done by adding more odorant solutions with solenoid valves to select which concentration joined the humid air stream. Flow through all channels was constant, and solutions were replaced after every run to reduce depletion.
Presentation was bi-rhinal, i.e., identical concentrations of NH3 were delivered to both nostrils. The results of Experiment 1 were also based on both nostrils, but only one nostril
Procedure.
Subjects rated the strength of sensations via magnitude estimation. Subjects
Subjects
One duration of each concentration served as a modulus: 2276 ms of 165 ppm for six subjects, 988 ms of 304 ppm for six subjects, and 510 ms of 478 ppm for seven subjects. According to pilot work, these stimuli roughly matched in intensity and fell close to the middle of the intensity range for all stimuli in the study. Subjects were randomly assigned to a modulus, with the constraint of approximately equal numbers in each group. At the beginning of each session, subjects
Data analysis.
The arithmetic mean summarized replicate ratings within subjects. The geometric mean summarized data across subjects (Stevens, 1975). Next, experimenters fit linear functions, via least squares regression, to plots of log rated intensity versus log stimulus duration. Good linear fits would indicate simple integration, i.e., increasing stimulus duration by a fixed factor would increase intensity by a fixed factor. Unit slope would indicate perfect integration, whereas a shallower slope, i.e., less than 1, would indicate imperfect integration (N.B. this is opposite to the analysis in Experiment 1). This analysis assumes that ratings are perfectly proportional to perceived intensity.
A second analysis assumed only that equal magnitude estimates across two conditions indicated equal perceived intensity (Cometto-Muiz and Cain, 1984). Experimenters calculated the stimulus durations for each concentration needed to achieve fixed levels of perceived intensity. The second analysis is more comparable to that of Experiment 1, where duration and concentration traded to maintain a fixed level of detection. If one plots log stimulus duration versus log concentration, a slope of –1 indicates perfect integration and a slope less than –1 indicates imperfect integration.
Results
We note three main features of the data in Figure 3. First, rated intensity increased with duration for each concentration. Second, linear functions (in log-log coordinates) described the relationship between duration and rated intensity quite well, accounting for 97.5–99.9% of the variance in ratings. To a first approximation, a fixed-ratio increase in stimulus duration at a given concentration produced a fixed-ratio increase in rated irritation. Finally, slopes (with 95% confidence intervals based on regression) were 0.57 (0.29–0.85) for 165 ppm, 0.77 (0.68–0.85) for 304 ppm, and 0.82 (0.62–1.02) for 478 ppm. Geometric mean slope, across concentrations, was 0.71 (0.56–0.91). Averaged across concentrations, doubling duration at fixed concentration increased intensity by a factor of 1.65.
By interpolating between the points in Figure 3 using the linear fits, experimenters calculated the durations for each concentration needed to produce four fixed levels of perceived intensity. The resulting (iso-intensity) curves, in log-log coordinates, appear in Figure 4. Linear functions described the curves quite well. Slopes ranged from –1.47 to –1.81 for the range of perceived intensities that the three concentrations had in common. Slopes appeared to increase with perceived intensity. Average (across the common intensity-range) slope was –1.64., indicating that a 3.1-fold increase in duration was required to compensate for a 2-fold decrease in concentration to maintain a fixed level of perceived irritation.
Discussion
The finding that ratings of irritation increased with stimulus duration for a fixed concentration demonstrates that integration occurred (Fig. 3). The finding that fixed-ratio increases in duration led to fixed-ratio increases in ratings suggests that rated intensity is related in some simple way to total mass delivered to the nose, i.e., that some form of simple mass-integrator model can describe the data. The finding that iso-intensity curves in Figure 4 were also approximately linear supports the same conclusion. Finally, the finding that slopes of intensity versus duration functions were less than 1 (Fig. 3), and the finding that slopes of iso-intensity curves (Fig. 4) were less than –1, suggests that an imperfect mass-integrator model is appropriate. All three findings are in good qualitative agreement with lateralization of carbon dioxide (Wise et al., 2004), lateralization of NH3 (Experiment 1), and an older study of supra-threshold ratings of irritation from NH3 (Cometto-Muiz and Cain, 1984).
Inspection of Figure 3 reveals another point of agreement between the results of Experiment 2 and those of Cometto-Muiz and Cain (1984). In both experiments, slopes of intensity versus stimulus-duration functions went up as concentration increased. All in all, qualitative agreement among the studies is striking.
The results of Experiment 2 disagree with those of Cometto-Muiz and Cain (1984) in degree of integration. Experiment 2 suggests that a 3.1-fold increase in duration is required to compensate for a 2-fold decrease in concentration to maintain a fixed level of perceived intensity, whereas the older study suggested that a 2.1-fold increase in duration would suffice. Accordingly, the two supra-threshold studies of irritation from NH3 disagree with each other about as strongly as the findings of Cometto-Muiz and Cain (1984) disagree with the Wise et al. (2004) results on lateralization of CO2 (which suggested that a 3.4-fold increase in concentration is required to compensate for a 2-fold decrease in concentration). This finding strongly suggests that differences between the two older studies are not exclusively stimulus driven. If the differences between the older studies were instead driven largely by differences between threshold and supra-threshold integration, then one would expect better integration in Experiment 2 than in Experiment 1. In other words, given the same stimulus and similar methods, we would expect supra-threshold integration to be more perfect than threshold-level integration. Results of Experiment 1, which suggested that a 2.5-fold increase in duration is needed to compensate for a 2-fold decrease in concentration, actually came closer to perfect integration than did those of Experiment 2. In short, the results suggest that neither differences between NH3 and CO2 nor differences between threshold-level and supra-threshold level integration can completely account for differences in degree of integration seen in previous studies.
Some disagreement probably comes from methodological differences. In the present experiments, subjects
GENERAL DISCUSSION
Research on temporal integration in detection of nasal irritation has been scant. Wise and colleagues showed that an imperfect, mass-integrator model provides a good description of short-term integration in lateralization of CO2 (Wise et al., 2004). Carbon dioxide probably stimulates nerve endings through tissue acidification (Hummel, 2000; Shusterman and Avila, 2003). The current report extends the previous finding by demonstrating simple but imperfect integration in lateralization of the base, NH3. Furthermore, in agreement with a previous report (Cometto-Muiz and Cain, 1984), the current experiments suggest that an imperfect, mass-integrator model can also describe short-term integration for supra-threshold irritation from NH3. Studies of other acids and bases can determine how well the findings generalize. Studies of still other compounds, e.g., nicotine and menthol, can determine whether an imperfect, mass-integrator model can describe irritation from compounds that do not stimulate through changes in pH.
Limitations
The simple integration model is of the black-box type, and it includes all events from entry of the stimulus into the nostril to execution of the response. Dynamics of the generated stimulus will strongly influence dynamics of concentration in the peri-receptor environment, i.e., in epithelial or sub-epithelial layers of the mucosa (Finger et al., 1990). However, patterns of flow and diffusion through the nasal cavity and subsequent diffusion into the tissue will also influence dynamics of peri-receptor concentration. Online tracking of peri-receptor concentration, perhaps through a pH meter (Shusterman and Avila, 2003) or pH-sensitive dye, may help elucidate this component of the black box (see Wise et al., 2004, for more discussion). Psychophysiology, e.g., combinations of psychophysics and measurement of mucosal or cortical evoked potentials (Hummel, 2000; Hummel et al., 2003), may help elucidate other components of the black box.
Another limitation is instrumental, because the olfactometer could not reliably produce pulses briefer than about 100 ms. Perhaps integration would behave differently for very brief stimuli. The visual system, for example, displays perfect integration up to about 100 ms (200 ms for rod vision), and imperfect integration up to as long as 3 s (Baumgardt, 1972).
Finally, the stimulation technique, i.e., passive injection of vapor into the nose during velopharyngeal closure, is un-physiological. To elucidate basic response properties of the sensory system, the rigorous experimental control of the method has advantages. However, experimental conditions differed from natural breathing. For example, experimental flow rates fell below those of normal breathing. Further, stimulus flow was probably confined mostly to the nose itself, rather than being distributed throughout the nasal cavity. Even within the nose, patterns of flow almost certainly differed from those associated with natural breathing. Given the considerable body of literature which suggests that rates and patterns of flow strongly influence deposition and absorption of volatile compounds in the nasal cavity (e.g., Frederick et al., 1994, 1998; Kurtz et al., 2004; Morris, 2001), one should exercise caution in generalizing the results of the current studies to more natural conditions. Additional studies using natural breathing techniques would be useful in this regard.
Basic Significance
Simple but imperfect integration is potentially consistent with a variety of mechanisms (Cain, 1990; Wise et al., 2004). For example, integration may come from a build-up of stimulus molecules in the mucosa over time, but the build-up could be undermined by continual clearance or breakdown of molecules. Adaptation could also undermine perfect integration. Further, both unmyelinated C-fibers and thinly myelinated A-fibers in the nose can respond to chemical irritants (see Bryant and Silver, 2000). The two types of fibers have different temporal response properties and may give rise to different sensations, i.e., burning versus stinging (see Hummel, 2000). Some effects of varying stimulus duration could come from sequential stimulation of the two populations of fibers.
As suggested above under Limitations, it will require more than psychophysics to relate patterns of perception to specific mechanisms. However, to understand the relationship between physiology and sensation, it is necessary to measure sensation. The psychophysical models described above, or extensions thereof, can guide biophysicists, molecular biologists, and physiologists in the quest to elucidate the mechanisms of perceived irritation.
Practical Significance
Short-term integration, i.e., integration that might occur within the duration of a single, natural inspiration, may have limited applicability to integration in natural settings. However, increased understanding of short-term integration can have immediate implications for how data are collected and interpreted in the laboratory. In the laboratory, investigations of sensitivity to irritants, both for patients and for normal controls, are often based on presentations that last between.25 and 3 s (Hummel, 2000; Hummel et al., 2003; Shusterman, 2002). Often, subjects simply sniff from bottles, and duration is uncontrolled. As a result of integration, individual differences in sniff duration can become confounded with differences in sensitivity. One can control stimulus duration through olfactometery. Still, because of individual differences in slopes of integration functions (see Wise et al., 2004, and Figure 1 above), stimuli of different lengths may provide very different pictures of individual differences. Furthermore, much of the growing body of literature on the how irritant potency is related to molecular parameters is based on brief exposures in the laboratory (e.g., Abraham et al., 1998, 2003). If compounds differ in the slopes of their integration functions, sampling at a fixed duration will provide at best an incomplete picture of differences among compounds. Simple models that allow researchers to characterize entire time-concentration trading functions by measuring a few carefully selected points can be powerful tools that give researchers a better understanding of their data.
NOTES
The authors certify that all research involving human subjects was done under full compliance with all government policies and the Helsinki Declaration.
ACKNOWLEDGMENTS
Financial support came in part from grants P50 DC00214 and T32 DC00014 from the National Institute on Deafness and other Communication Disorders of the National Institutes of Health. We thank three anonymous reviewers for helpful comments. Conflict of interest: none declared.
REFERENCES
Abraham, M. H., Hassanisadi, M., Jalali-Heravi, M., Ghafourian, T., Cain, W. S., and Cometto-Muniz, J. E. (2003). Draize rabbit eye test compatibility with eye irritation thresholds in humans: A quantitative structure–activity relationship analysis. Toxicol. Sci., 76, 384–391.
Abraham, M. H., Kumarsingh, R., Cometto-Muiz, J. E., and Cain, W. S. (1998). An algorithm for nasal pungency thresholds in man. Arch. Toxicol. 72, 227–232.
Anton, F., Euchner, I., and Handwerker, H. O., (1992). Psychophysical examination of pain induced by defined CO2 pulses applied to the nasal mucosa. Pain 49, 53–60.
Baumgardt, E. (1972). Threshold quantal problems. In Visual Psychophysics (D. Jameson and J. Hurvich, Eds.), pp 29–55. Springer-Verlag, New York.
Bliss, C. I. (1940). The relation between exposure time, concentration and toxicity in experiments on insecticides. Ann. Entomol. Soc. Am. 33, 721–766.
Bryant, B. P., and Silver, W. L. (2000). Chemesthesis: The common chemical sense. In The Neurobiology of Taste and Smell (T. E. Finger, W. L. Silver, and D. Restrepo, Eds.), 2nd ed., pp. 73–100. Wiley-Liss, New York.
Cain, W. S. (1996). Overview: Odors and irritation in indoor pollution. In Indoor Air and Human Health (R. B. Gammage and B. A. Berven, Eds.), 2nd ed., pp. 23–30. Lewis, Boca Raton, FL.
Cain, W. S. (1990). Perceptual characteristics of nasal irritation. In Chemical Senses (Vol. 2): Irritation (B.G. Green, J. R. Mason, and M. R. Kare, Eds.), pp. 43–58. Marcel Dekker, New York.
Cain, W. S., See, L. C., and Tosun, T. (1986). Irritation and odor from formaldehyde: Chamber studies. In IAQ '86: Managing the Indoor Air for Health and Energy Conservation, pp. 126–137. American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE), Atlanta.
Caterina, M. J., and Julius, D. (2001). The vanilloid receptor: A molecular gateway to the pain pathway. Annu. Rev. Neurosci. 24, 487–517.
Cometto-Muiz, J. E., and Cain, W. S. (1984). Temporal integration of pungency. Chem. Senses 8, 315–327.
Doty, R. L., and Cometto-Muiz, J. E. (2003). Trigeminal chemosensation. In Handbook of Olfaction and Gustation (R. L. Doty, Ed.), 2nd Ed., pp. 981–999. Marcel Dekker, New York.
Doty, R. L., Cometto-Muniz, J. E., Jalowayski, A. A., Dalton, P., Kendal-Reed, M., and Hodgson, M. (2004). Assessment of upper respiratory tract and ocular irritative effects of volatile chemicals in humans. Crit. Rev. Toxicol. 34, 85–142.
Finger, T. E., Getchell, M. L., Getchell, T. V., and Kinnamon, J. C. (1990). Affector and effector functions of peptidergic innervation of the nasal cavity. In Chemical Senses (Vol. 2): Irritation (B. G. Green, J. R. Mason, and M. R. Kare, Eds.), pp. 1–20. Marcel Dekker, New York.
Frasnelli, J., Ltsch, J., and Hummel, T. (2003). Event-related potentials to intranasal trigeminal stimuli change in relation to stimulus concentration and stimulus duration. J. Clin. Neurophysiol. 20, 80–86.
Frederick, C. B., Bush, M. L., Lomax, L. G., Black, K. A., Finch, L., Kimbell, J. S., Morgan, K. T., Subramaniam, R. P., Morris, J. B., and Ultman, J. S. (1998). Application of a hybrid computational fluid dynamics and physiologically based inhalation model for interspecies dosimetry extrapolation of acidic vapors in the upper airways. Toxicol. Appl. Pharmacol. 152, 211–231.
Frederick, C. B., Morris, J. B., Kimbell, J. S., Morgan, K. T., and Scherer, P. T. (1994). Comparison of four biologically based dosimetry models for the deposition of rapidly metabolized vapors in the rodent nasal cavity. Inhal. Toxicol. 6(Suppl.), 135–157.
Garcia-Medina, M. R., and Cain, W. S. (1982). Bilateral integration in the common chemical sense. Physiol. Behav. 29, 349–353.
Garner, W. R., and Miller, G. A. (1947). The masked threshold of pure tones as a function of duration. J. Exp. Psychol. 37, 293–305.
Hempel-Jorgensen, A., Kjaergaard, S. K., Molhave, L., and Hudnell, H. K. (1999). Time course of sensory eye irritation in humans exposed to N-butanol and 1-octene. Arch. Environ. Health 54, 86–94.
Hummel, T. (2000). Assessment of intranasal trigeminal function. Int. J. Psychophysiol. 36, 147–155.
Hummel, T., Mohammadian, P., Marchl, R., Kobal, G., and Ltsch, J. (2003). Irritation of the nasal mucosa using short- and long-lasting painful stimuli. Int. J. Psychophysiol. 47, 147–158.
Kobal, G., and Hummel, T. (1991). Olfactory evoked potentials in humans. In Smell and Taste in Health and Disease (T. V. Getchell, R. L. Doty, L. M. Bartoshuck, and J. B. Snow, Eds.), pp. 255–275. Raven Press, New York.
Kobal, G., Van Toller, S., and Hummel, T. (1989). Is there directional smelling. Experientia 45, 130–132.
Korhonen, K., Liukkonen, T., Ahrens, W., Astrakianakis, G., Boffetta, P., Burdorf, A., Heederik, D., Kauppinen, T., Kogevinas, M., Osvoll, P., et al. (2004). Occupational exposure to chemical agents in the paper industry. Int. Arch. Occup. Environ. Health 77, 451–460.
Kurtz, D. B., Zhao, K., Hornung, D. E., and Scherer, P. (2004). Experimental and numerical determination of odorant solubility in nasal and olfactory mucosa. Chem. Senses 29, 763–773.
Miller, F. J., Schlosser, P. M., and Janszen, D. B. (2000). Haber's rule: A special case in a family of curves relating concentration and duration of exposure to a fixed level of response for a given endpoint. Toxicology 149, 21–34.
Morris, J. B. (2001). Overview of upper respiratory tract vapor uptake studies. Inhal. Toxicol. 13, 335–345.
NIOSH (National Institute for Occupational Safety and Health). (1994). NIOSH Pocket Guide to Chemical Hazards (Publication No. 94–116). National Institute for Occupational Safety and Health, Cincinnati.
Omland, O. (2002). Exposure and respiratory health in farming in temperate zones—A review of the literature. Ann. Agr. Environ. Med. 9, 119–136.
Proctor, N. H., Hughes, J. P., and Fischman, M. L. (1988). Chemical Hazards of the Workplace. J. B. Lippincott, Philadelphia.
Shusterman, D. (2002). Individual factors in nasal chemesthesis. Chem. Senses 27, 551–564.
Shusterman, D., and Avila, P. C. (2003). Real-time monitoring of nasal mucosal pH during carbon dioxide stimulation: Implications for stimulus dynamics. Chem. Senses 28, 595–601.
Stevens, S. S. (1975). Psychophysics: Introduction to Its Perceptual, Neural, and Social Prospects. Wiley, New York.
Tarun, A., Bryant, B., Zhai, W., Soloman, C., and Shusterman, D. (2003). Gene expression for carbonic anhydrase isoenzymes in human nasal mucosa. Chem. Senses 28, 621–629.
Tominaga, M., Caterina, M. J., Malmberg, A. B., Rosen T. A., Gilbert, H., Skinner, K., Raumann, B. E., Basbaum, A. I., and Julius, D. (1998). The cloned capsaicin receptor integrates multiple pain-producing stimuli. Neuron 21, 531–543.
Wetherill, G., and Levitt, H. (1965). Sequential estimation of points on a psychometric function. Br. J. Math. Stat. Psychol. 18, 1–10.
Wise, P. M., Radil., T., and Wysocki, C. J. (2004). Temporal integration in nasal lateralization and nasal detection of carbon dioxide. Chem. Senses 29, 137–142.
Wise, P. M., Wysocki, C. J., and Radil T. (2003). Time–intensity ratings of nasal irritation from carbon dioxide. Chem. Senses 28, 751–760.
Wysocki, C. J., and Wise, P. (2003). Methods, approaches, and caveats for functionally evaluating olfaction and chemesthesis. In Handbook of Flavor Characterization: Sensory, Chemical and Physiological (K. Deibler and J. F. Delwiche, Eds.), pp 1–40. Marcel Dekker, New York.(Paul M. Wise, Thomas M. C)