Você está na página 1de 7

Chinese Journal of Physiology 53(6): 465-471, 2010 DOI: 10.4077/CJP.2010.

AMM034

465

Responses of Midbrain Auditory Neurons in Rats to FM and AM Tones Presented Simultaneously


M.F. Lee 1, Edmund C. So 2, 3, and Paul W.F. Poon 1
1

Department of Physiology, Medical College, National Cheng Kung University 2 Department of Anesthesia, Chi-Mei Medical Center (Yu Kang & Lui Yi) and 3 Institute of Basic Medical Sciences, National Cheng Kung University Tainan, Taiwan, Republic of China

Abstract
Speech and other communication signals contain components of frequency and amplitude modulations (FM, AM) that often occur together. Auditory midbrain (or inferior colliculus, IC) is an important center for coding time-varying features of sounds. It remains unclear how IC neurons respond when FM and AM stimuli are both presented. Here we studied IC neurons in the urethaneanesthetized rats when animals were simultaneously stimulated with FM and AM tones. Of 122 units that were sensitive to the dual stimuli, the responses could be grossly divided into two types: one that resembled the respective responses to FM or AM stimuli presented separately (simple sensitivity, 45% of units), and another that appeared markedly different from their respective responses to FM or AM tones (complex sensitivity, 55%). These types of combinational sensitivities were further correlated with individual cells frequency tuning pattern (response area) and with their common response pattern to FM and AM sounds. Results suggested that such combinational sensitivity could reflect local synaptic interactions on IC neurons and that the neural mechanisms could underlie more developed sensitivities to acoustic combinations found at the auditory cortex. Key Words: combinational sensitivity, frequency modulation, amplitude modulation, inferior colliculus, speech recognition

Introduction
Most natural sounds have components that are modulated in frequency and amplitude (frequency modulation, FM and amplitude modulation, AM). FM and AM sounds are considered building blocks of communication signals and their coding mechanisms are linked to speech recognition (12). Neural sensitivities to FM or AM sounds have been reported at various levels of the auditory pathways, and the responses typically change with stimulus parameters altered in a variety of ways (7, 11, 21-23, 28). FM sensitivity, when defined as the selective response to frequency sweeps of one direction but not

both (upward and downward changes in the tonal frequency), involves a central mechanism as it is not found at the auditory nerves (24). In an attempt to understand the basic response determinants and to unify the various stimulus parameters affecting FM responses, a simple classification of FM sensitivities was proposed for auditory midbrain neurons based on differential responses to FM and steady tones (19), namely: (a) FM specialized that respond to FM but not steady tone, (b) FM insensitive that respond to steady but not FM tone, and (c) FM mixed that respond to both FM and steady tones. FM specialized sensitivity emerges at the midbrain (or inferior colliculus, IC), as its incidence is low at the cochlear nucleus (24). A

Corresponding author: Dr. Paul W.F. Poon, Department of Physiology, Medical College, National Cheng Kung University, 1 University Road, Tainan 70101, Taiwan, Republic of China. Fax: +886-6-2362780, E-mail: ppoon@mail.ncku.edu.tw Received: December 28, 2009; Revised: February 11, 2010; Accepted: February 14, 2010. 2010 by The Chinese Physiological Society and Airiti Press Inc. ISSN : 0304-4920. http://www.cps.org.tw

466

Lee, So and Poon

parallel classification for AM responses was also proposed (2, 20): (a) AM specialized with response during the rising phase, but not the steady phase of the AM envelope; (b) AM insensitive with response during the steady phase but not the rising phase; (c) AM mixed with response during both the steady and the rising phases. Evoked potential studies in human showed that when FM and AM sounds are presented together, responses are larger in amplitude as compared to presentation of the sounds individually (10). Results of neuroimaging also support the presence of combined sensitivity to FM and AM as cortical fields activated by FM or AM stimuli appear overlapped (8). In auditory specialized animals like bats and songbirds, combinational sensitivities to multiple FM sounds are reported at single cell level in the cortex (14, 26). Similar sensitivity is also found at their midbrain (15). But combinational sensitivities to FM and AM sounds in the less specialized animals (like rodents) have not been studied. Furthermore, to what extent cortical responses may reflect what have been accomplished at the subcortical centers is unclear. Here we investigated at single cell level at the rats auditory midbrain, how responses change when FM and AM tones are both present. Results supported the presence of combination sensitive cells at the midbrain of rodents.

An exponential signal (0.5 cycle/sec) was digitally synthesized (TDT system II) and fed to the voltagecontrol-frequency input of a function generator (Tektronix FG280 Manufactory, Beaverton, OR, USA) to generate a glide tone of slowly-varying frequency to induce spikes from the cell. The range of frequency sweep was first adjusted to cover the cells best frequency and minimum threshold, and tone levels dropped systematically across 60 repeated trials (0.5-1.0 dB steps from high to low; 2-sec/trial; details see 3). FM Stimuli Two types of FM tones were generated by feeding digitally synthesized modulating waveforms to the function generator: (a) Exponential FM: They were tones modulated by a family of exponential ramps (n = 3 to 6), each of different FM velocities, with ramps appearing intermittently in each stimulus trial. Tone levels dropped systematically across repeated trials in a session. To best activate the cells, three different ranges of ramp rates were used: (a) 2 to 5, (b) 4 to 20, and (c) 6.6 to 100 cycle/sec. (b) Pseudo-random FM: The modulating envelope was generated digitally by low-pass filtering a white noise signal (at 50 or 100 Hz). The peaks and valleys in the filtered signal were reconnected digitally with an exponential envelope before input to the function generator. AM Stimuli Two types of AM tones were generated by feeding a digitally synthesized modulating waveform to the function generator through its voltage-controlamplitude input (at 40 dB dynamic range): (a) Exponential AM: They were tones modulated by a family of exponential ramps (n = 3 to 6), each of different AM velocities, with ramps appearing intermittently in each stimulus trial. To best activate the cell, three different ranges of ramp rates similar to FM were used. In a given stimulus session, unlike FM, AM stimulus trials were repeated mostly at a fixed level (instead of being systematically varied). (b) Pseudo-Random AM: The envelope was generated in a way similar to FM stimuli. Dual FM-and-AM Stimuli

Materials and Methods


Animal Preparation A total of 107 adult rats (Sprague-Dawley, 150350 gm) were studied under urethane anesthesia (Sigma, 2 gm/kg i.p.). Detailed procedures were reported earlier (2). In brief, skull was opened over the occipital cortex for electrode entry to the underlying midbrain and cerebrospinal fluid drained at the foramen magnum to reduce brain pulsation. Skin incision was sutured to restore pinnae to the resting position for free-field sound stimulation. A screw was cemented onto the cranium for head fixation to a modified stereotaxic frame in a sound-attenuated room (IAC). Rectal temperature was controlled at 38C with a thermal regulator. Sounds were delivered free-field from a loud speaker (Pioneer SP77, QUAD 306) placed 70 cm away, 30 in azimuth contralateral to the surgical side. Stimulus sound levels were calibrated at the site of the animal (B&K microphone 4191). The procedure was approved by the Animal Ethics Committee, NCKU. Acoustic Stimuli Slow Frequency Sweep for Characterizing Response Area (RA)

Two types of envelopes similar to the description above for FM and AM stimuli were used: (a) exponential and (B) pseudo-random. An audio-mixer com-

Combined FM-AM Sensitivity

467

bined the FM and AM signals before input to the speaker. The direction of sweeps was opposite between AM and FM sweeps as that was empirically found to be more effective in exciting IC cells when presented together. Electrophysiological Recording Single-units were recorded extracellularly from a microelectrode (3M KCl-filled glass micropipette, 30-60 M ; Dagan) advanced into the IC using a motorized microdrive (Narishige 5113). Amplified and band-pass filtered spike waveforms (Axonprobe1A, PARC; 0.3-3.0 kHz) were conditioned into rectangular pulses (duration 0.5 ms) with a level discriminator and spike times stored in a computer for off-line analysis. An intensive click (0.1 msec pulse, ~95 dB SPL) was first used to search for sound responsive units. After identifying an auditory unit, steady tones were then presented, followed by the random FM stimulations to determine audio-visually the cells response characteristics, i.e., best frequency (BF) and minimum threshold (MT). Responses to the battery of timevarying sounds were determined through the computer controlled system. Carrier frequency of the stimulus tone was adjusted to cover the units BF and the stimulus level set at 30 to 50 dB above MT. Statistical Analyses

Instant. tone frequency

Hi Lo

dB SPL dB SPL

A Simple
[Lo: 4.5, Hi: 40 kHz)

U74-10

73 13 38 0 83 29 21 0 73

B Asymmetrical
[Lo: 0.2, Hi: 12 kHz)

U96-8

dB SPL

C Multi-band
[Lo: 2.4, Hi: 20 kHz)

U94-4 43 45 0 73 13 1 0 1.0 1.5 2.0

dB SPL

Response trial 60 PSTH Hi


(no. spikes Lo /ms/60 trials) 0

D No response
[Lo: 14, Hi: 32 kHz) 0.5

U109-4

Time (sec)

Fig. 1. Types of response areas (A: simple, B: asymmetrical, C: multi-band, and D: no response RAs) as showed by 4 representative neurons in the IC. Top panel: modulating waveform (or instantaneous frequency) of the 2-sec long glide tone. Panels below: top showing spike responses of the cell in dot raster (each plot represents a stimulus session in which stimulus level is changed systematically from high to low over repeated trials), bottom: corresponding peri-stimulus time histogram (PSTH). Similar captions apply to this and other figures.

with previous reports (2, 19, 20). Response Areas (RAs) To determine dependence of the categorical variables (response area, FM or AM sensitivity, and combinational sensitivity), the chi-square test or Fishers exact test was used wherever appropriate. To confirm the categorical classification of response types, results from another observer were compared using Cohens kappa coefficient. In all statistical tests, significance level was set at P < 0.01. Frequency tuning characterized as response area (RA) was determined in 683 units (Fig. 1). The majority (73%, n = 497) responded to the slow glide tone and displayed dot-raster patterns that were classified as: simple, multi-band or asymmetric. A minority (27%, n = 186) of them responded to other sounds in the stimulus battery but not to this slow glide tone, and the cells were classified to have no response RA. Results are in general comparable with literature (3). FM and AM Responses A total of 261 units were fully characterized in terms of their response to the FM or AM stimuli that were presented separately (Fig. 2). For the same cell, the FM and AM sensitivities are highly correlated (P < 0.001, chi-square test; Table 1). Similarly a high correlation was found between RA type and the cells sensitivity to the modulated sounds ( P < 0.001, chisquare test; Table 2). For example, simple RA is associated more often with insensitive types (93%); no response RA exclusively with specialized types (100%); and multi-band RA with both specialized

Results
A total of 939 click-sensitive single units were recorded from 107 rats. Distributions of BFs and MTs were consistent with earlier reports from this laboratory (BF range: 0.5 to 50 kHz; lowest MT: -10 dB SPL at 11 kHz), suggesting a comprehensive sampling of the midbrain and the relative stability of response within the range of body weight. FM responses were tested on 837 units (31% insensitive, 44% specialized, 25% mixed), and AM responses of 286 were further characterized (38% insensitive, 30% specialized, 22% mixed). The breakdowns of both FM and AM sensitivities were moderately consistent across two observers (Cohens kappa coefficient = 0.60, P < 0.001), and results are comparable

468

Lee, So and Poon

Table 1. FM and AM sensitivity types are highly correlated (P < 0.001, chi-square test) Response type (total n = 261 units) AM insensitive (n = 91) AM mixed (n = 61) AM specialized (n = 109) FM insensitive (n = 102) 80 (31% of 261) 1 15 7 FM mixed (n = 66) 1 33 (13% of 261) 22 FM specialized (n = 93) 0 13 80 (31% of 261)

Table 2. Response area type and the FM-and-AM sensitivity type are highly correlated (P < 0.001, chi-square test) Response Area type (total n = 191 units) Simple RA (n = 69) Asymmetrical RA ( n = 20) Multi-band RA (n = 66) No response RA (n = 40) Insensitive to both FM, AM 64 (93%) 5 (25%) 11 (16%) 0 (0%) Specialized to both FM, AM 2 (3%) 7 (35%) 32 (49%) 40 (100%) Mixed to both FM, AM 3 (4%) 8 (40%) 23 (35%) 0 (0%)

Frequency Modulation

Amplitude Modulation
+ Volt.

Instant. tone freq.

Hi Lo 73

Insensitive

A
[0.9-2.0 kHz] U143-4

D
[4.0 kHz] U139-21

73

dB SPL

13 17 0 63

13 21 0 63

B
IMixed
[1.9-5.6 kHz] U143-6 0 29 0 73

E
[1.5 kHz] U144-7 60 18 0 60

dB SPL

Specialized

C
[1.4-5.2 kHz] U139-20 13 32 0 0 1 2 0

F
[9.0 kHz] U130-1 0 21 0 1 2

dB SPL Spk ont

Time (sec)

Fig. 2. Types of responses to pseudo-random FM or AM tone (A, D: insensitive, B, E: mixed and C, F: specialized) as showed by 6 units. The modulating waveform for AM represents the instantaneous level of the sound. Note the striking similarity in selective responses of the same type to the changing or steady phases to FM and AM.

or mixed types. Responses to Dual FM-and-AM Stimuli A total of 122 units were characterized in terms of their response to the dual FM-and-AM stimuli. For the sake of description, the combinational sensitivity is classified into two gross categories: (a) simple

sensitivity: one that resembled the cells respective AM or FM responses (45% of 122 units); and (b) complex sensitivity: one that did not (the remaining 55%). Figs. 3 to 5 show some representative examples of the two types of responses. The category of combinational sensitivity is also highly correlated with the cells RA type ( P < 0.001, Fishers exact test; Table 3). Specifically simple combinational sensi-

Combined FM-AM Sensitivity

469

A
Freq. [kHz]
15.6 4.4

U107-4 7

B
2 43

U96-8

FM

FM
73

dB SPL

[4.4-15.6 kHz] 0

[2-7 kHz] 43

FM + AM

FM + AM
73

dB SPL
43

Level [dB SPL] Response trial

60

90

AM
20 1 50

AM

[10.0 kHz] 0 1 2 0

[4.5 kHz] 1 2

60

Time (sec)

Fig. 3. Simple combinational responses to FM and AM sounds as showed by 2 units (A, B).

C
9.3

U75-15 9.0

D
FM
63 2.2

U93-7

kHz
3.0

FM

78

dB SPL

[3-9.3 kHz] 3

[2.2-9 kHz] 48

FM + AM
63

FM + AM
78

dB SPL

dB SPL

80

95

48

AM
40 1 55

AM

Response trial

[6.1 kHz] 0 1 2 0

[5.6 kHz] 1 2

60

Time (sec)

Fig. 4. Complex combinational responses to FM and AM sounds as showed by 2 units (C, D).

tivity is associated more often with simple RA (92% out of 24 units), and asymmetrical RA (89% out of 9 units), and complex combinational sensi-

tivity on the other hand is associated more with no response RA (87% out of 37 units). Multi-band RA is associated with both combinational sensi-

470

Lee, So and Poon

Table 3. Combinational sensitivity and response area type are highly correlated (P < 0.001, Fishers exact test) Response Area type (n = 122 units) Simple RA (n = 24) Asymmetrical RA (n = 9) Multi-band RA (n = 52) No response RA (n = 37) Combinational FM-and-AM sensitivity Simple (45%, n = 55) 22 (91% of 24) 8 (89% of 9) 20 (39% of 52) 5 (13% of 37) Complex (55%, n = 67) 2 (9% of 24) 1 (11% of 9) 32 (61% of 52) 32 (87% of 37)

U109-4 35

kHz

11

FM
74

[11-35 kHz]

48

FM + AM
74

48 90

dB SPL

AM

Response trial

50 1

[22.9 kHz]

60

Fig. 5. Example of a unit with complex combinational sensitivity, with its response strongly depended on the simultaneous presentation of FM and AM sounds.

tivities (61% for complex, and 39% for simple out of 52 units).

Discussion
We reported for the first time that combination sensitive cells to FM and AM sounds are present at the level of midbrain of the rat, a species that is considered not as specialized in hearing as the bat or songbird. Our division of the combinational sensitivities into simple and complex patterns is likely more of operational than functional value at this point. The most interesting finding is that some cells (in particular those with complex combinational sensitivity) need the presence of both FM and AM to generate strong spike responses ( e.g. , Fig. 5). It is likely that the proportion of IC cells with complex combinational sensitivity could be even higher than the current percentage of 55%, should the hunting procedure be more efficient, i.e., a random combination of the dual sounds without requiring response of the cell to one

or both stimuli in the first instance. One primary difficulty of studying combinational sensitivity is the provision of a complete dual stimulus set. The number of ways FM and AM stimuli can be combined for presentation is larger than what had been tested in the present study based on stereotyped modulations. For technical reasons, we were only able to vary a limited number of stimulus parameters. Hence, responses to dual FM-and-AM stimuli were mostly studied with one stimulus parameter systematically varied while keeping the other constant. Similar findings of simple and complex combinational sensitivities using the more extensive pseudo-random dual stimuli supported the validity of the results. More advanced stimulus strategies for hunting these units are required for a deeper understanding of the effects of other stimulus parameters on combinational responses ( e.g. , intensity level, modulation rate, and modulating envelope). The simple combinational sensitivity, by virtue of its resemblance to FM and AM respective responses suggested minimal or simple interactions between the two inputs, like summation of stimulus energy from FM and AM inputs, post-excitatory inhibition or lateral inhibition that are known to take place in the IC (1, 21, 25). The complex combinational sensitivity on the other hand suggested non-linear interactions between the two inputs. Furthermore, the complex combinational sensitivity associated more often with the specialized responses types. Since FM-specialized cells in the IC have been found to be larger neurons (18), the interactions underlying the complex combinational sensitivity could potentially take place over their extensive dendritic fields. The complex inhibitory and excitatory synaptic neurotransmitter systems found in the IC is also in support of complex interactions in the midbrain (6, 9, 16) in addition to other non-linear response properties of neurons that exist as low as the level of cochlear nucleus (17). There has been growing evidence that some of the non-linear response properties found at the cortex are built upon mechanisms lower in the brainstem (4, 5, 13, 27). Our present finding of complex combinational sensitivity as low as the midbrain is in line of this view.

dB SPL dB SPL

Combined FM-AM Sensitivity

471

The highly correlated FM and AM response types for the same cell suggested strongly that timevariance is a common denominator of response for neurons in the auditory midbrain. Consistent with this finding, similar co-variance of FM and AM response types has been reported at auditory cortex (7). The high correlation we found between FM and AM sensitivities and RA type and combinational sensitivities further supported the usefulness of our three-type classification on time-varying sensitivity of midbrain auditory neurons.

Acknowledgments
Supported in part by grants from National Science Council, Taiwan, and CMFHR 9559 of Dr. Teng-Kuei Hsieh.

References
1. Chiu, T.W., Liu, Y.D. and Poon, P.W.F. Transient frequency and intensity sensitivities of central auditory neurons estimated with sweep tone - a new approach. Chinese J. Physiol. 41: 133-138, 1998. 2. Chiu, T.W. and Poon, P.W.F. Basic response determinants for amplitude modulation of single neurons in the auditory midbrain Exp. Brain Res. 134: 237-245, 2000. 3. Chiu, T.W. and Poon, P.W. Multiple-band trigger features of midbrain auditory neurons revealed in composite spectro-temporal receptive fields. Chinese J. Physiol. 50: 105-112, 2007. 4. Deutscher, A., Kurt, S., Scheich, H. and Schulze, H. Cortical and subcortical sides of auditory rhythms and pitches. NeuroReport 17: 853-856, 2006. 5. Eggermont, J.J. Temporal modulation transfer functions for AM and FM stimuli in cat auditory cortex. Effects of carrier type, modulating waveform and intensity. Hearing Res. 74: 51-66, 1994. 6. Escabi, M.A. and Schreiner, C.E. Nonlinear spectrotemporal sound analysis by neurons in the auditory midbrain. J. Neurosci. 22: 41144131, 2002. 7. Gaese, B.H. and Ostwald, J. Temporal coding of amplitude and frequency modulation in the rat auditory cortex. Eur. J. Neurosci. 7: 438-450, 1995. 8. Hart, H.C., Palmer A.R. and Hall, D.A. Amplitude and frequencymodulated stimuli activate common regions of human auditory cortex. Cereb. Cortex 13: 773-781, 2003. 9. Ito, T., Bishop, D.C. and Oliver, D.L. Two classes of GABAergic neurons in the inferior colliculus. J. Neurosci. 29: 13860-13869, 2009. 10. John, M.S., Dimitrijevi, A., van Roon, P. and Picton, T.W. Multiple auditory steady-state responses to AM and FM stimuli. Audiol. Neuro-Otol. 6: 12-27, 2001.

11. Joris, P.X., Schreiner, C.E. and Rees, A. Neural processing of amplitude-modulated sounds. Physiol. Rev. 84: 541-577, 2004. 12. Kay, R.H. Hearing of modulation in sounds. Physiol. Rev. 62: 894-969, 1982. 13. Malmierca, M.S., Cristaudo, S., Perez-Gonzalez, D. and Covey, E. Stimulus-specific adaptation in the inferior colliculus of the anesthetized rat. J. Neurosci. 29: 5483-5493, 2009. 14. Margoliash, D. and Fortune, E.C. Temporal and harmonic combination-sensitive neurons in the zebra finchs HVc. J. Neurosci. 12: 4309-4326, 1992. 15. Mittmann, D.H. and Wenstrup, J.J. Combination-sensitive neurons in the inferior colliculus. Hearing Res. 90: 185-191, 1995. 16. Nataraj, K. and Wenstrup, J.J. Roles of inhibition in creating complex auditory responses in the inferior colliculus: facilitated combination-sensitive neurons. J. Neurophysiol. 93: 3294-3312, 2005. 17. Nelken, I., Kim, P.J. and Young, E.D. Linear and nonlinear spectral integration in type IV neurons of the dorsal cochlear nucleus. II. Predicting responses with the use of nonlinear models. J. Neurophysiol. 78: 800-811, 1997. 18. Poon, P.W.F., Chen, X. and Cheung, Y.M. Differences in FM responses correlate with morphology of neurons in the rat inferior colliculus. Exp. Brain Res. 91: 94-104, 1992. 19. Poon, P.W.F., Chen, X.Y. and Hwang, J.C. Basic determinants for FM responses in the inferior colliculus of rats. Exp. Brain Res. 83: 598-606, 1991. 20. Poon, P.W.F. and Chiu, T.W. Similarities of FM and AM receptive space of single units at the auditory midbrain. Biosystems 59: 229237, 2000. 21. Rees, A. and Malmierca, M.S. Processing of dynamic properties of sounds. In International Review of Neurobiology Vol. 70: Auditory Spectral Processing, Malmierca, M.S. and Irvine, D.R.F. (eds), New York, Academic Press, 2005, pp. 299-330. 22. Rees, A. and Moller, A.R. Responses of neurons in the inferior colliculus of the rat to AM and FM tones. Hearing Res. 10: 301330, 1983. 23. Ricketts, C., Mendelson, J.R., Anand, B. and English, R. Responses to time-varying stimuli in rat auditory cortex. Hearing Res. 123: 27-30, 1998. 24. Sinex, D.G. and Geisler, C.D. Auditory-nerve fiber responses to frequency-modulated tones. Hearing Res. 4: 127-148, 1981. 25. Sivaramakrishnan, S., Sterbing-DAngelo, S.J., Filipovic, B., DAngelo, W.R., Oliver, D.L. and Kuwada, S. GABA(A) synapses shape neuronal responses to sound intensity in the inferior colliculus. J. Neurosci. 24: 5031-5043, 2004. 26. Suga, N., ONeill, W.E., Kujirai, K. and Manabe, T. Specificity of combination-sensitive neurons for processing of complex biosonar signals in auditory cortex of the mustached bat. J. Neurophysiol. 49: 1573-1626, 1983. 27. Theunissen, F.E., Sen, K. and Doupe, A.J. Spectral-temporal receptive fields of nonlinear auditory neurons obtained using natural sounds. J. Neurosci. 20: 2315-2331, 2000. 28. Whitfield, I.C. and Evans, E.F. Responses of auditory cortical neurons to stimuli of changing frequency. J. Neurophysiol. 28: 655-672, 1965.

Você também pode gostar