David E. Hornung

A quantitative analysis of sniffing strategies in rats performing odor detection tasks

The sniffing strategies of rats performing two learned odor detection tasks were monitored with a pneumotachograph and quantitatively analyzed with respect to fifty-two characteristics. The results of this study demonstrated that the rats sniffing varied for different odorants, different concentrations of the same odorant, and between air and odor trials. The variations resulted from changes in such descriptors as volume, duration, average flow rate, peak flow rate and sniff number. In general, a sniffing pattern began with one or two inspirations followed by alternating inspirations and expirations. Comparison of earlier and later sniffs in a bout demonstrated a growth towards both a maximum inspiratory and expiratory sniff which had the largest duration, volume, average flow rate and peak flow rate. These maximum sniffs occurred at or near the end of a bout. Although analysis of the fifty-two characteristics was quantitatively useful in determining the physiologic values and airflow patterns generated by sniffing, a single univariate response measure incorporating twelve characteristics was the best descriptor of how sniffing patterns varied with odorant stimuli.

The olfactory loss that accompanies an HIV infection

A number of studies have shown that HIV infection is associated with decreased olfactory ability. Additionally, it has been hypothesized that a reduced odorant identification may precede the advent of AIDS Dementia Complex (ADC). However, it is not known whether changes in olfactory ability are a manifestation of neurocognitive decline which may precede the appearance of AIDS Dementia Complex, damage to the peripheral olfactory system from opportunistic infection, or whether olfactory structures have a particular sensitivity to HIV. These issues were addressed in a cross-sectional study examining variability in the neuropsychological, neurological, otolaryngological, auditory, and olfactory status in HIV-positive subjects. A stepwise regression provided evidence that the ability to identify odorants was influenced by age, nasal structure and pathology, neurocognitive ability, and level of AIDS Dementia Complex. On the other hand, only nasal pathologies accounted for the variability in olfactory thresholds. These data suggest that identification and thresholds tests may reflect different olfactory pathologies. Additionally, these data suggest at least part of the decline in olfactory ability accompanying an HIV infection may be secondary to nasal pathologies. Because of their rapidly changing neurocognitive status, HIV-positive patients represent an excellent group in which to study the determinants of olfactory ability.

Congenital Lack of Olfactory Ability

Twenty-two patients, all of whom reported never having been able to smell anything, were studied to determine the particular features that distinguish individuals with congenital anosmia. The clinical evaluation on these patients included a thorough medical and chemosensory history, physical examination, nasal endoscopy, chemosensory testing, olfactory biopsies, and imaging studies. There was no evidence to indicate that these patients ever had a sense of smell. The results of olfactory testing suggested that these patients had an inability to detect both olfactory and trigeminal odorants; however, many of the patients in the group seemed to have a slight ability to perceive at least some component of trigeminal odorants. The olfactory epithelium, if it was present at all on biopsy, was abnormal in appearance.

A method for establishing a five odorant identification confusion matrix task in rats

Using a cross-modal association paradigm, rats were trained to associate a particular tunnel and response location with one of five different odorants (isoamyl acetate, propyl acetate, acetic acid, phenethyl alcohol, and anethole). Each of the five tunnels differed with respect to: 1) the illuminated pattern on the response key; 2) the brightness of the illuminated pattern; and 3) the somesthetic quality of the tunnel floor. Standard operant techniques were used to train trial initiating and sampling behavior at a central odorant presentation point. Following acquisition training, the animals were tested using a standard 5 X 5 confusion matrix design. The results showed for the first time that rats are capable of performing, with a high degree of accuracy, an odorant identification confusion matrix task analogous to humans. Furthermore, using multidimensional scaling techniques, these data represent the first instance in which the perceptual odor space of an animal can be determined. With the animal model in hand, we can begin to examine how, in the presence of neural dysfunction, one odorant may be correctly identified as another.

Determination of carbon dioxide detection thresholds in trained rats

Concentration-response functions for the detection of CO2 were established for six rats. The animals were tested in a wind tunnel apparatus and trained using standard operant techniques and a discrete trials, go, no-go successive discrimination paradigm. The primary conclusion to be drawn from the performance measurements is that, at least under carefully controlled conditions, rats can detect physiologic concentrations of CO2 (0-4%). Minimum detectable concentrations fell within the range of 0.04-1.7% CO2. The concentration-response function describing the detectability of CO2 for the six rats was divided into an upper and lower limb at a concentration (5%) that was approximately equal to the end expiratory CO2 levels for the rat (4.88%). High levels of performance were observed for concentrations above this point, while those below it (0.02-2.5%) represented the dynamic range of detectability. Based on a 65% performance criterion, the average threshold performance for six rats was 0.52%. Possible interpretations of these data are discussed.

Distribution of butanol molecules along bullfrog olfactory mucosa.

ALTHOUGH electrophysiological1 and gas chromatographic2 studies have indicated that odorant molecules drawn into the intact frog olfactory sac establish a concentration gradient along the olfactory mucosa, conclusive evidence for such gradients requires a direct determination of how the molecules are actually distributed. This requirement is met here by labelling butanol molecules with tritium and directly mapping their sorption along the mucosa.

Phonological and perceptual components of short-term memory for odors.

Just as a written word can be encoded and retained in memory verbally or visually, an odor might be retained as a verbal description or perceptual (olfactory) code. However, one view holds that olfactory memory in the short term does not exist as a separate perceptual code. This was examined in an experiment in which memory errors could be seen as deriving from the substitution of similar verbal or olfactory codes. The odorants presented for recall were divided into three groups: base odorants (which might be replaced in memory by similar verbal or olfactory representations), verbal foils (stimuli dissimilar to the base stimuli in odor but similar in name), and odor foils (the reverse). The substitution errors made in attempting to recall test odorants were classified as verbal or olfactory. A substantial proportion of the errors were olfactory, but verbal errors also occurred. These results support the presence of short-term perceptual olfactory memory rather than simply verbal encoding of olfactory perceptions.

Initial mechanisms basic to olfactory perception

Animal experimentation has proposed three mechanisms at the olfactory mucosa that may underlie olfactory discrimination. First, the olfactory receptor cells appear selectively tuned to different odorants. Second, in a gas chromatographic-like process, the molecules of different odorants appear to be distributed in different sorption patterns across the mucosal surface. Third, different regions of the mucosa appear to have different selective sensitivities. These three mechanisms could complement each other by together generating a greater number of neural discharge patterns to encode the odorants passing over the mucosal surface. In this interplay, the mucosal distribution patterns could differentially limit the receptor cells and mucosal regions to which different odorants have access. The mucosal distribution pattern could thereby affect the odorant analyses made by these other mechanisms as well as contribute its own analysis. The mucosal distribution patterns appear fairly stable in the face of rather wide variations in the pertinent variables characterizing the nasal airflow (namely, odorant concentration, flow rate, volume, and duration). There are, however, limits to these variables beyond which significant shifts in the molecular distributions and neural discharge patterns can be produced. Thus, in humans any naturally occurring or surgically induced alteration in the nasal airflow which appreciably alters these variables may affect olfactory perception. Olfaction in a laryngectomized patient is discussed as an example.

A changing density technique to measure nasal airflow patterns

To study the relationship between nasal airflow and olfactory function, a radiation detection technique was developed to quantify the airflow patterns through a uninasal model of the human air passageways. The theoretical basis of this technique is that for any point in the nasal model, the rate of displacement of one gas by another is proportional to the flow rate at that point. In these studies, the model was first filled with xenon gas, and then, as room air was drawn through the model, the xenon was displaced. To measure the changing xenon density, a collimated gamma ray source (200 millicuries of lz5I) was directed through the model toward a specially designed sodium iodide detector (FIG. 1). Because of the collimation, only a small cross section of the model (approx. 4 mm*) was studied at any one time. This technique of measuring the changing xenon density takes advantage of the fact that the number of gamma rays received per second by the detector is dependent upon the density of gas inside the nasal model. Since xenon gas is much denser than air, the number of gamma rays received by the detector increased as the xenon gas was removed from the model. The detector output was fed into a multichannel scaler such that each succeeding channel contains the output for each succeeding 0.4 msec interval. To analyze the data, the counts versus time least-squares regression equation was calculated. The slope of this equation is the rate of displacement of xenon by the air. As described above, the slope is, therefore, also directly proportional to the flow rate. To determine the reproducibility of the data, the slope of the counts versus time curve for each of four flow rates (2, 6, 15, 45 I/min) was determined for a single nasal position. The four randomly ordered flow rates defined a block of trials, and four such blocks were run. The two-way analysis of variance showed there was a significant effect of flow rate ( p < 0.01), but not blocks ( p < 0.05). Thus, it appears that this technique can give reproducible results. In the second study three blocks at the above flow rates were run at each of two nasal positions. Position A was in the middle meatus near the center of the nasal fossa. Position B was just posterior to the nasal valve, midway between the nasal floor and dorsum. The results of this study (TABLE 1) show that at position A there was

Anosmic Patients Can Separate Trigeminal and Nontrigeminal Stimulants

Anosmia is the inability of the first cranial nerve pathway to detect airborne odorant molecules. In humans, however, information about airborne odorant molecules comes from a variety of sensory systems in addition to the first cranial nerve. As a first step in determining the degree of involvement of other sensory systems in the smell deficits of anosmic patients, a retrospective study was undertaken to determine if these patients could separate trigeminal and nontrigeminal odorants.