Logo der Universität Wien

RESEARCH Harald Tichy

ON and OFF responses to changes in food odor concentration

 In our lab, we are interested in the temporal features of olfactory stimuli that are encoded in the discharge rates of insect olfactory receptor neurons (ORNs). One of the greatest challenges that insects have overcome is localizing distant sources of important odors, like food or mate. Unlike visual and auditory stimuli, olfactory stimuli are not inherently directional. In order to achieve spatial orientation, insects use the flow of air as directional information. However, plume tracking cockroaches, which are faced with an abrupt stop of the wind flow, continue to track successfully the now stationary odor plume “hanging in a zero wind environment” (Willis et al 2008). How cockroaches orient to an odor source under windless conditions remains a largely unexplored question. Cockroaches seem to perform true chemotaxis by exploiting the “odor landscape” in order to localize the odor source. In an “odor landscape”, the average peak concentration of the odor pulse and the strongly correlated steepness of the pulse onset slope, or equivalently, the rate at which the pulse concentration increases, tend to rise with decreasing distance to the source (Moore and Atema 1991; Atema 1996). This concentration patterns contain directional information that is used by marine crustaceans for orientation and distance estimates.


In our studies we are not concerned with the combinatorial code of odor identity. We are interested in the problem of encoding odor pulse steepness or the rate of pulse concentration as it relates to the spatial-temporal odor distribution. In the field of insect olfaction, variations in the pulse onset slope have not received the same attention as variations in the pulse repetition rates, the latter commonly tested as trains of transient on–off pulses. Going beyond, coding of pulse slopes requires varying the rate of concentration change in a controlled manner and recording the ORN’s activity. To this end, we developed a computerized air dilution flow olfactometer capable of delivering both discrete pulsed concentration changes as well as slow and continuous concentration changes at various rates (Burgstaller and Tichy 2011, 2012; Tichy and Hellwig 2018). This form of stimulation was critical for the discovery of the ON and OFF ORNs on the cockroach’s antennae which provide excitatory responses for both increments and decrements in food odor concentration (Fig. 1A,B). This arrangement suggests that, if the cockroach moves to the food odor source, it uses the ON ORNs to detect the odor signal and the OFF ORNs to recognize that it gets lost from the odor signal.


The parallel ON and OFF ORNs perform the olfactory equivalent of temporal contrast enhancement and promote the detection of slight changes in food odor concentration. During slowly oscillating changes in the concentration of the odor of lemon oil, the ON and OFF ORNs adapt to the actual odor concentration and the rate at which concentration changes. When odor concentration oscillates rapidly with brief periods, adaptation improves the gain for instantaneous odor concentration and reduces the gain for the rate of concentration change. Conversely, when odor concentration oscillates slowly with long periods, adaptation increases the gain for the rate of change at the expense of instantaneous concentration. Without this gain control the ON and OFF olfactory cells would, at brief oscillation periods, soon reach their saturation level and become insensitive to further concentration increments and decrements. At long oscillation periods, on the other hand, the cue would simply be that the discharge begins to change. Because of the high gain for the rate of change, the cockroach will receive creeping changes in odor concentration, even if they persist in one direction. Gain control permits a high degree of precision at small rates when it counts most, without sacrificing the range of detection and without extending the measuring scale (Burgstaller and Tichy 2012; Tichy and Hellwig 2018).


The ON and OFF odor detection is the most significant division among olfactory features extracted by the peripheral olfactory system. Still, separate ON and OFF information channels do not help with temporal contrast enhancement because each channel signals exactly what it detects without caring about what is going on in its antagonistic pair. To accomplish such a phenomenon, one has to postulate that the ON and OFF channels talk to each other – opposite signals may enhance each other whereas similar or equal signals may weaken each other. It is also conceivable that at some stage the ON and OFF channels are combined thereby providing ON/OFF responses to both the onset and the termination of the odor stimulus. Since odor increments and odor decrements do not occur physically at the same time in the same place, the convergence of the ON and OFF channels may not loose information about the rate of concentration change. Some ON/OFF neurons may be sensitive to rapid concentration changes but insensitive to slow concentration changes, others may be sensitive to slow concentration changes but insensitive to rapid concentration changes. Thus it is tempting to reject the hypotheses that olfactory performance is limited to sampling ON responses alone. The aim of our present experiments is to simultaneously examine the activity from multiple neurons in the cockroach’s antennal lobe to food odor stimulation by means of 16-channel silicon based microprobe arrays.

Fig. 1A,B. Antagonistic responses of the ON and OFF olfactory cells to oscillating changes in the concentration of the odour of lemon oil; oscillation periods are 6 s in A and 60 s in B: a, time course of odour concentration; b, simultaneously recorded impulses of an ON and OFF olfactory cell. The OFF olfactory cell displays larger impulse amplitudes than the ON olfactory cell. c and d, responses of the ON and OFF olfactory cells represented in raster plots

Accurate orientation movements require that the cockroach detects odor concentration and its rate of change independently of the air flow velocity. Rising air flow velocity will increase the volume of air flowing in a given period of time across the antenna. An increase in the volume flow rate of an odor pulse with constant onset slope and constant peak concentration leads to an increase in the rate of odor molecules arriving at the sensillum. One would expect that an ORNs acting as “pulse slope detectors” should not confuse (1) an increase in the molecule arrival rate due to increasing the air flow rate at constant concentration with (2) an increase in the molecule concentration due to diluting the odor a little less with pure air that in the actual ratio, no matter how fast the air is flowing. In the first case, the number of odor molecules per unit volume of air is constant but in the second it is raised. Therefore, the ORNs response magnitude should be the same if a constant dilution ratio is delivered at various flow rates, but should be changed in a predictable manner if the dilution ratio of the odor sample in pure air is changed, even if the flow rate is varied. This applies exactly to the cockroach’s ON and OFF ORNs. Our study showed that increasing the flow rate of a constant odor concentration air stream has no effect on the response of the ON and OFF ORNs. By contrast, changing the odor concentration (in the amount of odor molecules per unit air volume) affects in a reproducible manner the responses of both types of ORNs, independently of the flow rate of the odor carrying air stream, the air volume passing over the sensillum per unit time or the number of molecules arriving at the sensillum (Fig. 2A-D). Thus, the ORNs act as “pulse slope detectors” with the ability to extract directional and source distance information from the odor plume (Hellwig et al 2019). Much work lies still ahead to refine our understanding of how ORNs satisfy the ability to discriminate differences in the rate of concentration changes despite flow velocity-dependent variations in the number of molecules.



Fig. 2A–D. Impulse frequency of an ON ORN (A) stimulated with three constant–amplitude, oscillating changes in the concentration of the lemon oil odor (D). Air flow ve-locity is increased in a step-wise manner (0.5, 1 and 1.5 m/s) by increasing the volume of the odor-carrying air passing during the oscillation period over the antenna (C). The rate of arrival of odor molecules increases with rising air flow velocity (B). Nevertheless, the ORN’s re-sponse is the same during each oscillation. The ORNs assess the rate of concentration change of the odor carrying air, independently of the absolute number of molecules involved in the change, the air volume or its flow rate.



Atema J (1996) Eddy chemotaxis and odor landscapes: exploration of nature with animal sensors. Biol Bull 191:129–138

Burgstaller M, Tichy H (2011) Functional asymmetries in cockroach ON and OFF olfactory receptor neurons. J Neurophysiol 105:834–845

Burgstaller M, Tichy H (2012) Adaptation as a mechanism for gain control in cockroach ON and OFF olfactory receptor neurons. Eur J Neurosci 35:519–525

Hellwig M, Martzok A, Tichy H (2019) Encoding of slowly fluctuating concentration changes by cockroach olfactory receptor reurons is invariant to air flow velocity. Front. Physiol. 10:943. doi: 10.3389/fphys.2019.00943

Moore PA, Atema J (1991) Spatial information in the three-dimensional fine structure of an aquatic odor plume. Biol Bull 181:404–418

Tichy H, Hellwig M (2018) Independent processing of increments and decrements in odorant concentration by ON and OFF olfactory receptor neurons. J Comp Physiol A 204(11): 873–891. doi: 10.1007/s00359-018-1289-6

Willis MA, Avondet JL, Finnell AS (2008) Effects of altering flow and odor information on plume tracking behavior in walking cockroaches, Periplaneta americana (L.). J Exp Biol 211: 2317–2326


Funding Supported by a Grant from the Austrian Science Fund (Project P30594-B17).


Humidity transduction models


Insect hygroreceptors are associated in antagonistic pairs of a moist and a dry cell in the same sensillum with a thermoreceptive cold cell. The mechanism by which humidity stimulates the moist and dry cells remains controversial. Three main models of humidity transduction have been proposed, namely in which hygroreceptors operate as 1) mechanical hygrometers in which activity is initiated by swelling and shrinking of hygroscopic sensillum structures, 2) evaporimeters in which the rate of evaporation of water due to the dryness of the air leads to quantitative changes in the lymph concentration, and 3) psychrometers in which the degree of cooling during evaporation of water is used to measure the humidity (or the dryness) of the air (Tichy and Loftus 1996; Steinbrecht 1999; Tichy and Gingl 2001; Tichy et al 2017). These models imply very different stimulus-response relationships, which give rise to some confusion concerning the definition of the adequate humidity stimulus. The moist cell and the dry cell appear to be bimodal in that their responses to humidity strongly depend on temperature. Either modality can be changed independently of the other, but both are related in some way to the amount of moisture in the air and to its influence upon evaporation. The assessment below discusses the validity of the three models of humidity transduction (mechanical hygrometer, psychrometer and evaporation rate detector) by determining whether specific predictions based upon them are indeed observed in electrophysiological responses. The hygro-mechanical function, most favored for humidity transduction, has been challenged and arguments for an evaporative function will be presented. Firstly, some background information regarding the ways of expressing and measuring humidity will be provided (Tichy et al 2017).

Humidity refers to the amount of evaporated water in the air and is defined as the partial pressure exerted by the evaporated water vapor on the total pressure of the air. The water vapor pressure (Pw) is not affected by air temperature (Fig. 3A). Increasing air temperature increases the kinetic energy of the molecules in the air, but not the concentration of the vapor molecules. The more kinetic energy in the air, the more water can be evaporated, and the more water vapor is required for saturation of the air. The saturation vapor pressure (Ps) increases as air temperature increases (Fig. 3B). The relative humidity (rH) is the ratio between Pw actually in the air and Ps (Fig. 3C), indicating how close the air is to saturation. Therefore, rH is not a direct measure of the amount of water vapor in the air. As a ratio, rH is dimensionless; if it is used as a humidity parameter, then – for completeness – the air temperature must be supplied with it. The saturation deficit (SD), in contrast, is a measure of humidity that is expressed in vapor pressure units. It is the difference between the theoretical water pressure at saturation and actual vapor pressure in the air being measured at the same temperature (Fig. 3D). While an increase in rH at constant Pw is correlated with decreasing air temperature (Fig. 3C), SD increases with increasing air temperature (Fig. 3D). Thus, rH is inversely related to air temperature (Fig. 3C) and SD is directly related to air temperature (Fig. 3D). In contrast to rH, SD integrates in a single value the effects of both temperature and humidity (or dryness) of the air on the evaporation rate.

Evaporation of water can also be measured psychrometrically by the degree of cooling at the evaporating surface. As the water molecules escape, they take kinetic energy with them, leaving the surface with a diminished total kinetic energy. If all the latent heat of vaporization has been supplied to the air, then this temperature is known as the wet-bulb temperature: the lowest temperature to which the surface can be cooled by evaporation of water (Fig. 3E). It is the temperature that would be taken up by a thermometer bulb kept moist by a thin wet covering. In order to determine the temperature depression due to the cooling effect of the evaporating water, a second temperature reading is needed from a thermometer with a dry surface (dry-bulb temperature) indicating air temperature. Raising air temperature by adding (sensible) heat increases both dry-bulb (Fig. 3E) and wet-bulb temperature (Fig. 3F), but there is no change of the evaporated water in the air. A system functioning as a psychrometer would require two temperature -sensitive hygroreceptors, one beneath a dry surface and unaffected by cooling and another with a wet surface cooled by evaporation.

Fig. 3A-H. 3D-mesh plots of the effects of air temperature on different ways of expressing humidity and the responses of a moist cell and a dry cell during two periods of slowly oscillating changes in water vapor pressure. (A) Humidity stimulation consisted of two consecutive periods of constant-amplitude oscillations in vapor pressure (0-15 mbar) at four different temperature levels of 21.0, 22.8, 24.7 and 26.8°C. (B) Saturation vapor pressure at the same temperature levels. Constant-amplitude oscillating change in vapor pressure in A produces, with rising temperature, continuously deceasing oscillations in relative humidity (C), but continuously increasing oscillations in both saturation deficit (D) and wet-bulb temperature. (E) Dry-bulb temperature as a function of air temperature. (G,H) Impulse frequency of a moist cell and a dry cell recorded from the same sensillum on the cockroach’s antenna during oscillations in vapor pressure at the different temperature levels (21.0, 22.8, 24.7, 26.8°C) illustrated in A. With rising temperature, the oscillations in impulse frequency of the moist and dry cells shift upwards on the frequency scale. dry T dry-bulb temperature, Ps saturation water vapor, Pw vapor pressure, rH relative humidity, SD saturation deficit, wet T wet-bulb temperature.

Electrophysiological studies on the cockroach revealed a dependence of the responses of the moist and the dry cell to slowly fluctuating changes in vapor pressure on the temperature level at which the humidity changes are carried out (Fig. 3G,H). The positive temperature coefficient has implications for the validity of the humidity transduction models. Expressing the fluctuation in vapor pressure (Fig. 3A) as fluctuations in the relative humidity (Fig. 3C), the resulting negative temperature coefficient excludes a mechanical hygrometer function. The positive temperature coefficient of the wet-bulb temperature (Fig. 3E) supports a psychrometric measurement. However, the close proximity of a “wet-surface” and a “dry-surface” on the same sensillum make it difficult to understand how the dry air temperature could be taken without being affected by evaporative cooling. The positive temperature coefficient of the saturation deficit (Fig. 3D) highlights an evaporative function. A key experiment for testing a specific prediction drawn from the evaporation model would be to alter the flow rate of the stimulating air stream. Flow rate drastically affects evaporation power. Thus, experiments with slow and continuous changes in the saturation deficit at different flow rates and temperatures would assess the validity of an evaporative function. From the small amount of information currently available it is not possible to relate the type of response of the hygroreceptive sensory cells with any locomotor reactions which result in the insect aggregation in preferred humidity zones. Further detailed studies of these aspects in insect species known to inhabit different humidity environments would be highly desirable.




Steinbrecht RA (1999) “Bimodal thermo- and hygrosensitive sensilla” in Microscopic Anatomy of Invertebrates, Vol. 11B, eds Harrison FW, Locke M (New York, Wiley-Liss). 405–422

Tichy H, Gingl E (2001) “Problems in hygro- and thermoreception” in Ecology of Sensing, eds Barth FG and Schmid A (Berlin, Heidelberg, New York: Springer). 271–287. doi: 10.1007/978-3-662-226

Tichy H, Hellwig M, Kallina W (2017) Revisiting theories of humidity transduction: a focus on electrophysiological data. Front. Physiol. 8:650. doi: 10.3389/fphys.2017.00650

Tichy H, Loftus R (1996) Hygroreceptors in insects and a spider: humidity transduction models. Naturwissenschaften 83, 255–263. doi: 10.1007/BF01149598


Funding Supported by a Grant from the Austrian Science Fund (Project P20197-B17).


Department für Neurobiologie
Universität Wien

Althanstrasse 14
A-1090 Wien
T: +43 1 4277 56501
Universität Wien | Universitätsring 1 | 1010 Wien | T +43-1-4277-0