Doklady Biological Sciences, Vol. 382, 2002, pp. 5–7. Translated from Doklady Akademii Nauk, Vol. 382, No. 1, 2002, pp. 131–133. Original Russian Text Copyright © 2002 by Egorova, Vartanyan, Ehret. PHYSIOLOGY Neural Critical Bands and Inhibition in the Auditory Midbrain of the House Mouse (Mus domesticus) M. A. Egorova*, I. A. Vartanyan*, and G. Ehret** Presented by Academician V.L. Sviderskii August 1, 2001 Received August 1, 2001 Critical bands is a well-known psychophysical phenomenon to describe the mechanisms of sound frequency analysis [1–3]. Studies on the neurophysiological basis of perceptual critical bands (CBs) showed that the main properties of CBs are coded by the neural activity of the auditory midbrain [4–7]. Analysis of the shape of the receptive fields of inferior colliculus neurons in relation to neural critical bandwidths raised the hypothesis that inhibition is a main neurophysiological mechanism of critical bandwidth regulation [5, 8]. In the previous study on the spatial distribution and interaction of excitatory and inhibitory receptive fields of neurons in the central nucleus of the mouse inferior colliculus, we divided the neurons into three main classes: “primary-like,” mostly reproducing the frequency tuning of auditory nerve afferents; “inhibitiondominated” with strong inhibitory side-bands invading deeply into narrow areas of excitation; and “V-shaped,” characterized by weak inhibitory inputs and broad excitatory tuning [8]. To clarify the role of inhibitory processes in forming neural CBs, the parameters of neural CBs and inhibition have to be statistically analyzed and compared in the classes of mouse inferior colliculus neurons. Here, we report the results of this analysis. Extracellular single-unit recordings from the central nucleus of the inferior colliculus were made in 40 anesthetized (35 mg/kg of ketavet and 0.1 mg/kg of rompun) female house mice, Mus domesticus (hybrids of the outbred strain NMRI and feral mice aged 8– 15 weeks). Excitatory and inhibitory frequency response areas were mapped for 122 neurons by conventional single-tone and two-tone interaction paradigms [9]. Single-tone excitatory tuning curves and two-tone response areas were measured automatically with the use of computer controlled programs. Tone stimuli (50 ms in duration; with 5-ms trapezoidal rise * Sechenov Institute of Evolutionary Physiology and Biochemistry, Russian Academy of Sciences, pr. Morisa Toreza 44, St. Petersburg, 194223 Russia ** University of Ulm, D-89069 Ulm, Germany and fall times) were delivered at 300 ms intervals. In the two-tone paradigm, one tone was presented at the neuron’s excitatory characteristic frequency (CFE) 10 dB above the threshold, while the second tone varied across a broad frequency and intensity range, corresponding to both the excitatory and inhibitory response areas. Measurements were taken within the whole frequency range of mouse hearing (3–80 kHz) [6] and sound levels from the unit’s threshold up to 80 dB above it (from –10 dB SPL to 90 dB SPL). Of the 122 units recorded, 54 were classified as primary-like, 36 as inhibition-dominated, and 32 as V-shaped. The estimation of the neural critical bandwidths was made by means of simultaneous narrowband noise masking with a constant noise spectral level and a variable bandwidth. The neurons’ responses to a characteristic-frequency tonal signal 10–70 dB above the response threshold were masked by a continuous broadband noise. As soon as the response was masked, i.e., the neuron did not respond to the tone bursts any more, the noise bandwidth (steepness of the slope, 96 dB/oct) was narrowed gradually from the high-frequency and low-frequency sides until the tone response appeared again [5]. The noise bandwidth that was just effective in masking of the tonal response directly determined the lower and the upper borders of the critical bands. Measurements of the critical bandwidths were made for 98 units (43 primary-like, 28 inhibitiondominated, and 27 V-shaped ones). The distribution of the values of critical band’s lowand high-frequency borders as a function of the neurons’ CFE is shown in Fig. 1 for all units recorded. The values of critical-band borders increase with increasing CFE. This relationship could be approximated by linear regressions with a high statistical significance (p < 0.001 in each case). The regression lines follow the equations (frequencies in kHz): The critical band’s low-frequency border = 0.85 CFE + 0.5; r = 0.972, n = 346. (1) The critical band’s high-frequency boarder = (2) 1.06 CFE + 0.2; r = 0.987, n = 330. 0012-4966/02/0102-0005$27.00 © 2002 MAIK “Nauka /Interperiodica” 6 EGOROVA et al. kHz 70 kHz 70 (a) 60 60 50 50 40 40 30 30 20 20 10 10 0 70 60 0 10 20 30 40 50 60 kHz Fig. 1. The relationship between excitatory characteristic frequencies of neurons and their critical band borders. Linear regression lines show statistically significant relationships (Eqs. (1), (2)). The low-frequency critical band border, solid circles; the high-frequency critical band border, open circles. Abscissa: excitatory characteristic frequency, kHz; ordinate: border frequencies of critical bands, kHz. The slopes of the regression lines in both equations are close to 1.0 but differ significantly (p < 0.005) from each other. The regression coefficients indicate that the average critical bands are not centered around the unit’s CFE in logarithmic coordinates. Ignoring the constant values in Eqs. (1) and (2) (0.5 and 0.2, respectively) the average distance between CFE and the low-frequency border of the critical band is about 1/4 octave, whereas the average distance between CFE and the high-frequency border of the critical band is only about 1/12 octave. The relationship between the neurons’ CFE and characteristic frequencies of low-frequency and highfrequency inhibition (i.e., the frequencies corresponding to the lowest thresholds of inhibition of the inhibitory response areas below (lower) and above (higher) the CFE, respectively) are shown for the three classes of neurons in Fig. 2 (primary-like, Fig. 2a; inhibitiondominated, Fig. 2b; V-shaped, Fig. 2c). The correlation between the neurons’ CFE and characteristic frequencies of low-frequency and high-frequency inhibition show a high statistical significance (linear regression, p < 0.001 in each case). Primary-like neurons: low-frequency inhibition (CFIL) = 0.83 CFE – 1.2, r = 0.944, n = 48, (3) high-frequency inhibition (CFIH) = 1.12 CFE + 2.4, r = 0.986, n = 51, (4) Inhibition-dominated neurons: CFIL = 0.92 CFE – 1.3, r = 0.987, n = 36, (5) (b) 50 40 30 20 10 0 70 (c) 60 50 40 30 20 10 0 10 20 30 40 50 60 70 kHz Fig. 2. The relationship between excitatory and inhibitory characteristic frequencies in the classes of (a) primary-like, (b) inhibition-dominated, and (c) V-shaped neurons. Linear regression lines (solid lines) show statistically significant relationships (Eqs. (3)–(8)). Low-frequency inhibitory CFs (CFIL), solid circles; high-frequency inhibitory CFs (CFIH), open circles. Dotted lines show the linear regressions for the relationship between neurons’ excitatory characteristic frequencies and critical band borders (Eqs. (1), (2)). Abscissa: excitatory characteristic frequency, kHz; ordinate: inhibitory characteristic frequencies and border frequencies of critical bands, kHz. CFIH = 1.03 CFE + 2.6, r = 0.990, n = 36, (6) V-shaped neurons: CFIL = 0.95 CFE – 3.7, r = 0.960, n = 27, CFIH = 0.98 CFE + 8.6, r = 0.962, n = 22. DOKLADY BIOLOGICAL SCIENCES Vol. 382 (7) (8) 2002 NEURAL CRITICAL BANDS AND INHIBITION IN THE AUDITORY MIDBRAIN All slopes of the regression lines (Eqs. (3)–(8)) are close to 1.0, as in Eqs. (1) and (2). Nevertheless, the slopes of the regression lines for CFIL and CFIH differ significantly in primary-like (p < 0.001) and inhibitiondominated neurons (p < 0.05). The distances between the neurons’ CFE and characteristic frequencies of lowfrequency and high-frequency inhibition are proportional to 1/4–1/12 octave. The values of the constants of the CFIL equations are about half of those of the CFIH equations. The values of the constants in the CFIL and CFIH equations for primary-like and inhibition-dominated neurons are only about one third of the corresponding constants for V-shaped neurons. The dotted lines in Fig. 2 are the regression lines fitting the Eqs. (1) and (2) (see also Fig. 1). Comparing both relationships shown in Fig. 2, i.e., the dependencies of critical band borders and inhibitory characteristic frequencies on the neurons' CFE, demonstrates their strikingly parallel courses in all classes of neurons. In inhibition-dominated neurons, the regression lines described by Eqs. (5) and (6) are practically identical to the respective regression lines described by Eqs. (1) and (2) (no significant differences between the slopes). Thus, the highly significant correlations between the characteristic frequencies of inhibition and excitation closely reflecting the correlations between the characteristic frequencies of excitation and borders of critical bands, especially in inhibition-dominated neurons, support the hypothesis that inhibition plays a fundamental role in shaping the neural critical bands [5, 8, 10]. The relations between the constants in the regression equations shown above could reflect specific features of organization of the neurons in the three classes, e.g., periodicity in the organization of inhibitory inputs, DOKLADY BIOLOGICAL SCIENCES Vol. 382 2002 7 and thus express an ordered functional pattern [10, 11] of the inferior colliculus structure. ACKNOWLEDGMENTS This study was supported by the Volkswagen Foundation (project no. 1/69-589) and the Russian Foundation for Basic Research (project nos. 96-04-122 and 0004-48516). REFERENCES 1. Fletcher, H., Rev. Mod. Phys., 1940, vol. 12, pp. 47–65. 2. Scharf, B., in Foundations of Modern Auditory Theory, vol. 1, New York: Academic Press, 1970, vol. 1, pp. 159– 202. 3. Moore, B.C.H., in Human Psychophysics, Berlin: Springer, 1993, pp. 56–115. 4. Ehret, G. and Merzenich, M.M., Science, 1985, vol. 227, pp. 1245–1247. 5. Ehret, G. and Merzenich, M.M., Brain Res. Rev., 1988, vol. 13, pp. 139–163. 6. Ehret, G., in Advances in Hearing Research, Singapore: World Scientific, 1995, pp. 387–397. 7. Vartanyan, I.A., Egorova, M.A., and Ehret, G., Dokl. Akad. Nauk, 1999, vol. 368, no. 2, pp. 270–272. 8. Vartanyan, I.A., Egorova, M.A., and Ehret, G., Dokl. Akad. Nauk, 2000, vol. 373, no. 5, pp. 701–703. 9. Egorova, M.A., Ehret, G., Vartanian, I.A., and Esser, K.H., Exp. Brain Res., 2001, vol. 140, pp. 145– 161. 10. Ehret, G., in The Central Auditory System, New York: Oxford University Press, 1997, pp. 259–316. 11. Schreiner, C.E. and Langner, G., Nature, 1997, vol. 388, pp. 383–386.