| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Department of Neurobiology, Northeastern Ohio Universities College of Medicine, Rootstown; and 2Department of Biomedical Engineering, The University of Akron, Akron, Ohio
Submitted 8 November 2004; accepted in final form 29 January 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
-aminobutyric acid (GABA) (bicuculline, BIC) showed that STRY abolished facilitation in 96% of tested units, but BIC eliminated facilitation in only 28%. This suggests that facilitatory interactions are created in IC and reveals a differential role for these neurotransmitters. The facilitation may be created by coincidence of a postinhibitory rebound excitation activated by the low-frequency signal with the BF-evoked excitation. Unlike facilitation, inhibition at earlier delays was not eliminated by application of antagonists, suggesting an origin in lower brain stem nuclei. However, inhibition at delays later than facilitation, like facilitation itself, appears to originate within IC and to be more dependent on glycinergic than GABAergic mechanisms. Facilitatory and inhibitory interactions displayed by these combination-sensitive neurons encode information within sonar echoes and social vocalizations. The results indicate that these complex response properties arise through a series of neural interactions in the auditory brain stem and midbrain. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
, 1990; Saint Marie et al. 1989; Winer et al. 1995, 1996). Inhibition determines basic auditory response properties such as evoked discharge rates ( Backoff et al. 1999; Foote et al. 1975; Wang et al. 2000), temporal response patterns ( Caspary et al. 1979; Chen and Jen 2000; LeBeau et al. 1996; Pollak and Park 1993), responsiveness as a function of sound level ( Chen and Jen 2000; Vater et al. 1992; Yang et al. 1992), and frequency response areas ( Fuzessery and Hall 1996; LeBeau et al. 2001; Yang et al.1992). Inhibition also plays major roles in the creation of more complex responses. In the inferior colliculus (IC), for example, many complex response properties arise de novo through specific inhibitory input that creates or enhances selectivity for certain stimulus features. GABAergic inhibition contributes to directional selectivity to frequency modulated sweeps ( Fuzessery and Hall 1996) and to selectivity for binaural cues ( Faingold et al. 1991; Fujita and Konishi 1991; Park and Pollak 1993). Both GABAergic and glycinergic mechanisms mediate duration tuning ( Casseday et al. 1994). In each of these examples, inhibition appears to have frequency tuning, aurality, or timing that is different from the dominant excitatory input. Understanding the inhibition is essential for understanding the mechanisms that create the response selectivity.
This study analyzes the role of inhibition among neurons of the IC that show temporally sensitive facilitation to the combination of acoustic elements in vocalizations. In the mustached bat, these combination-sensitive neurons are abundant in the auditory forebrain ( Fitzpatrick et al. 1998; Olhemiller et al. 1996; Suga et al. 1978, 1983) but also in the IC where they appear to originate ( Mittmann and Wenstrup 1995; Portfors and Wenstrup 1999; Wenstrup and Leroy 2001). This integration of distinct spectral sensitivities appears to depend on specific low-frequency-tuned inputs to high-frequency regions of the IC ( Wenstrup et al. 1999). Inhibition is associated with these facilitated responses, both in terms of the response to the relative timing of different signal elements ( Edamatsu and Suga 1993; Olsen and Suga 1991b; Portfors and Wenstrup 1999) and in terms of the underlying mechanisms ( Wenstrup and Leroy 2001). Although it is clear that inhibitory interactions accompany facilitation in the IC, medial geniculate body, and auditory cortex, it is not clear how the inhibition and facilitation are related in IC neurons, nor is it clear where the inhibitory interactions originate. In the present study, we related inhibitory and facilitatory interactions across a sample of facilitated combination-sensitive IC neurons, and we blocked receptors to the inhibitory neurotransmitters glycine and -aminobutyric acid (GABA). The results show that glycinergic input to the IC is essential for the facilitatory interactions, but GABAergic input is not required. Moreover, while inhibitory interactions appear to affect the timing of the facilitatory interaction, blockade of the receptors indicates that some of the inhibitory combination-sensitive interactions originate at levels below the IC.
Work by K. Nataraj was in partial fulfillment of a Ph.D. in Biomedical Engineering at the University of Akron.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Surgical procedures
Twelve bats provided data used in this study. Some bats were anesthetized with methoxyflurane as described previously ( Portfors and Wenstrup 1999), but most were anesthetized with isoflurane (1.5–2.0%; Abbott Laboratories, North Chicago, IL). Ten minutes prior to isoflurane application, the animal received intraperitoneal injections of butorphanol (5 mg/kg, Fort Dodge Animal Health, Fort Dodge, IA) and atropine (0.06 mg/kg, Phoenix Scientific, St Joseph, MO). After the anesthetic abolished nociceptive reflexes, the dorsal surface of the head was shaved and cleaned. A midline incision was made in the skin, and the underlying muscles were reflected laterally to expose the skull. The skull surface was cleaned, and a tungsten ground electrode was cemented into the lateral surface of the skull overlying the cerebral cortex. A metal pin was cemented to the skull to secure the head in the stereotaxic apparatus during physiological experiments. A hole (usually <0.5 mm in diameter) was placed in the skull above the IC. After these procedures were completed, a local anesthetic (Lidocaine) was applied to the surgical areas, and the animal was returned to the holding cage. The bat recovered for 2–3 days before starting physiological experiments.
Acoustic stimulation and data acquisition
Acoustic stimulation was computer-controlled. Single and multiple tone bursts (4- to 31-ms duration, 0.5-ms rise/fall time, 4 presentations/s) were computer synthesized, downloaded to a digital signal processor (Tucker-Davis Technologies, Model AP2), then converted to analog signals at a sampling rate of 500 kHz (Tucker-Davis Technologies, Model DA3-2). The analog signal was filtered (Tucker-Davis Technologies, Model FT6-2), attenuated (Tucker-Davis Technologies, Model PA4), and then sent to a power amplifier (Parasound Model HCA-800II) and the loudspeaker (Technics Leaf Tweeter, Model EAS 400B). The speaker was placed 10 cm in front of the bat and 25° into the sound field contralateral to the IC under study.
Extracellular action potentials were amplified, filtered (band-pass, 500–6,000 Hz), and sent through a spike signal enhancer (Fredrick Haer, Model 40-46-1) before being digitized at a sampling rate of 40 kHz (Tucker-Davis Technologies, Model AD2). The digitized signal was uploaded to the computer via a second AP2 digital signal processing card. Custom software calculated the time of occurrence of the spikes and displayed peristimulus time histograms (PSTHs), raster plots, and statistics of the neural responses in real time.
Recording procedures
Physiological experiments were conducted inside a single-walled acoustic chamber lined with polyurethane foam to reduce echoes. On experimental days, the bat was placed in a stereotaxic apparatus inside the heated and humidified acoustic chamber. If the bat showed signs of discomfort or distress, it was lightly sedated with acepromazine (1 mg/kg, subcutaneous) or butorphanol (0.05 mg/kg, subcutaneous) or removed from the stereotaxic apparatus for the day. The bat was offered water from a medicine dropper between electrode penetrations. Recording sessions did not exceed 4–6 h and were limited to one per day.
Single-unit recordings were obtained using a glass micropipette electrode mounted on a five-barreled pipette ( Havey and Caspary 1980). The multibarreled pipette (A-M Systems) was pulled to a tip and broken to a total tip diameter of 15–30 µm. The single-barrel recording electrode (A-M Systems) was glued onto the tip of the multibarrel pipette at a 20° angle and protruded 10–25 µm beyond it. The recording electrode was filled with 1 M NaCl and had resistances in the 10- to 30-M
range. One barrel of the pipette, filled with 1 M NaCl, balanced currents of other channels that delivered or retained drugs. The remaining barrels were filled with strychnine-HCl (STR; (10 mM in distilled water, 3.0 pH, Fluka, Milwaukee, WI), and bicuculline-methiodide (BIC; 10 mM in 0.9% NaCl, 3.0 pH, Sigma, St. Louis, MO). All drugs and recording solutions were prepared on the day of the experiment. Both strychnine and bicuculline-methiodide were retained with negative current (–15 nA each) and ejected using positive currents (15–40 nA for strychnine and 10–50 nA for bicuculline). Each barrel of the multibarreled pipette was connected via a silver wire to a channel of a microiontophoresis current generator (Dagan Programmable current generator, Model No. 6400). The current generator controlled retention and ejection currents for each barrel separately. Drug ejection times and ejection currents varied depending on the drug and the effect of the drug as evaluated audiovisually and by quantitative tests.
Electrodes were angled 10–20° (dorsolateral to ventromedial) to record from tonotopic regions of the central nucleus of the inferior colliculus (ICC) representing higher frequencies (56–120 kHz), and advanced by a hydraulic micropositioner (David Kopf Instruments, Model 650).
Data were obtained only from well isolated single units. When a single unit was isolated, the best frequency (BF; defined as that frequency requiring the lowest sound level to elicit consistent, stimulus-locked action potentials) and minimum threshold (MT) at BF (defined as the lowest sound level at the BF to elicit consistent, stimulus-locked action potentials) were obtained with single tone burst stimuli. This use of the term "best frequency," common in the neuroethological literature, is equivalent to "characteristic frequency." Because many IC neurons respond best to the combination of distinct spectral elements in sounds (combination sensitivity), we used a two-tone stimulus paradigm to evaluate the underlying neural interactions. One of the tones was at the unit's BF (high-frequency signal), 10 dB above threshold. The second signal (low-frequency signal), tuned to lower frequencies within (23–30 kHz) or adjacent to the first harmonic of the biosonar call, was varied over a range of frequencies, sound intensities, and timing (delays) relative to the BF signal. When a combination-sensitive response was suspected, quantitative measures of delay-sensitive facilitation or inhibition were obtained and compared with single-tone responses. Audiovisual examination established the range of delays over which quantitative data were obtained. Delay sensitivity was evaluated in 2-ms steps. Typically, very short signals (4 ms) were used to reveal maximum temporal sensitivity; longer-duration stimuli were only used if units were unresponsive to 4-ms signals. Responses to 32 presentations of each stimulus were collected.
Units were considered to show facilitatory combination-sensitive interactions if the response to the combined low- and high-frequency signals, separated by the appropriate delay, was 
20% higher than the sum of the responses to the two signals presented separately (20% facilitation). Similarly, units were considered to show inhibitory combination-sensitive interactions if the response to the combined low- and high-frequency signals, separated by the appropriate delay, was 20% lower than the sum of the responses to the two signals presented separately (20% inhibition). The strength of combination-sensitive interactions was quantified by an interaction index, where index = (Rc – Rl – Rh)/(Rc + Rl + Rh). Rc is the unit's response to the combination of low- and high-frequency signals, Rl is the unit's response to the low-frequency signal alone, and Rh is the unit's response to the high-frequency signal alone. Interaction index values of +1 and –1 correspond to the strongest facilitation and inhibition, respectively, and interaction index values of 0.09 and –0.11 correspond to thresholds for facilitation and inhibition, respectively. In some units, high background discharge affected calculation of the interaction index. This effect was minimized by analyzing responses within a 50- to 60-ms window.
Delay sensitivity tests and rate-level functions were obtained prior, during, and, in some cases, after application of drugs. During drug application, rate-level or delay sensitivity tests were continuously obtained. Changes in response magnitude or the shape of response functions served to indicate the presence of a drug near the recording site. Low ejection currents, e.g., 10 nA, were used initially. If no effect was observed, the ejection current was gradually increased. For every current setting, delay curves were obtained until no further changes in response magnitudes or response functions were observed. With currents used in this study, effects of BIC and STRY could be observed as early as 2 and 3 min, respectively, while steady-state effects of BIC and STRY could be observed after 4 and 8 min, respectively. If there was any indication that drugs could not be ejected, the pipette/electrode assembly was retracted and discarded. New piggy-backed electrodes were used for every penetration.
Identical tests were performed before applying a drug (control), during drug application (drug), and after drug application stopped (recovery). A unit was considered to show complete recovery when response functions and response magnitude returned to control values. Recovery could be observed as early as 15 min or as long as 45 min after drug application was stopped and was dependent on the particular drug, ejection currents, and the duration of drug application. To save time, data were collected in the following sequence: control, drug-1, drug-1+drug-2, recovery, drug-2, and recovery. The two drugs were STRY and BIC and were, interchangeably, applied as drug-1 or drug-2.
Data analysis
Combination-sensitive interactions, facilitatory or inhibitory, were characterized by three features: the delay at which the interaction was at its maximum (best delay), the range of delays over which the interaction occurred (delay width), and the maximum strength of the interaction (interaction index). Figure 1 illustrates these features for a unit that was both facilitated and inhibited.
|
) level of 0.05. Analyses of variance tested for differences in various features of combination-sensitive interactions (best delay, delay width, interaction index) across the groups of facilitatory units. Within each group, paired t-test were used to find significant differences among the features of the combination-sensitive interactions and regression analyses were performed to find significant correlations between these features. Mean values are reported with the corresponding SE. |
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Most of these facilitated units (76%, n = 58) also showed combination-sensitive inhibition at delays earlier and/or later than the delays causing facilitation. RESULTS describes the contributions of inhibition to the facilitatory interaction in two ways. Across the population of single units, we studied the relationships between features of inhibition and features of facilitation. In a subset of the units (n = 47), we further examined the effect of blocking glycinergic and/or GABAergic inhibition by local iontophoretic application of STRY or BIC, respectively.
Features of facilitatory and inhibitory interactions
Two-tone delay curves revealed the time-dependent facilitatory and inhibitory influences of low-frequency signals on the responses to the higher, BF signals (Fig. 2). Figure 2A shows the response of a facilitated unit with no apparent inhibition at earlier or later delays. We observed a strong response to the BF tone (86.3 kHz) but a weak response to a low-frequency tone (29.2 kHz). The strongest facilitation occurred when the two tones were presented simultaneously at 0-ms delay. The absence of inhibition in the delay function suggests that the low-frequency signal evoked only excitation. Figure 2B shows the response of a facilitated combination-sensitive unit with inhibition only at early delays. There was a moderate response to the BF tone (72.6 kHz) but a weak response to a low-frequency tone (24.3 kHz). The strongest facilitation occurred when the BF tone was presented 12 ms after the low-frequency tone. Strong inhibition occurred when the BF tone was presented 0–4 ms after the low-frequency tone, with complete elimination of the response at 0- and 4-ms delay. This pattern suggests that the low-frequency signal activated inhibition followed by excitation. Figure 2C shows the response of a facilitated combination-sensitive unit with inhibition only at later delays. We observed a moderate response to the BF tone (72.1 kHz) but no response to the low-frequency tone (26.8 kHz). The strongest facilitation occurred when the BF tone was presented 2 ms after the low-frequency tone. Inhibition was evident when the BF tone was presented 8–10 ms after the low-frequency tone (Fig. 2C). The pattern displayed by this unit suggests that the low-frequency signal activated excitation followed by inhibition. Figure 2D shows the response of a facilitated combination-sensitive unit with both inhibition at early delays and inhibition at later delays. There was a moderate response to the BF tone (81.6 kHz) but no response to the low-frequency tone (23.6 kHz). The strongest facilitation occurred when the BF tone was presented 4 ms after the low-frequency tone. Suppression at early delays occurred when the BF tone was presented 0–2 ms before the low-frequency tone. Suppression at later delays occurred when the BF tone was presented 12–16 ms after the low-frequency tone. For these units, the delay curve suggests that the low-frequency signal activated, in turn, inhibition, excitation, and inhibition.
|
|
|
Twenty-three units showed inhibition at later delays. The best inhibitory delay ranged from 6 to 28 ms (Fig. 4C). The width of inhibition at later delays, measured in 19 units, ranged from 2 to 15 ms (Fig. 4C). The strength of inhibition at later delays ranged from –0.17 to –0.75. There were no significant correlations among the features of inhibition at later delays.
As expected, the best later inhibitory delays were longer than the best early inhibitory delays (F[2,144] = 53.2, P < 0.001). Inhibition at early delays was stronger than at later delays (F[2,143] = 323.1, P < 0.001), but delay width, a measure of the duration of the inhibition, was the same (F[2, 127] =5.9, P > 0.1). The width of inhibition at early or later delays was greater than the width of facilitation (F[2,127] = 5.9, P < 0.01).
Relationship between inhibition and facilitation
How do inhibitory interactions influence the features of combination-sensitive facilitation? In this section, we examine whether features of facilitation differ among combination-sensitive units based on the presence and timing of inhibition and whether facilitatory features are correlated with the timing, duration, or strength of inhibitory interactions among units showing these interactions.
There was a clear relationship between early inhibition and the best facilitatory delay. Most units without early inhibition (83%, n = 26) had best facilitatory delays of 
4 ms (range: –3–10 ms; average: 2.4 ± 0.5 ms). In contrast, nearly all units (98%, n = 44) with early inhibition had best facilitatory delays of 4 ms (range: 0–30 ms; average: 9.1 ± 0.9 ms). This difference is highly significant (F[1.74] = 31.4, P < 0.001), suggesting that inhibition at early delays alters the best facilitatory delay of units. Moreover, among the 44 units showing inhibition at early delays, there was a positive correlation between the best early inhibitory delay and the best facilitatory delay (r = 0.63, P < 0.01, Fig. 5A). An even stronger correlation existed between the width of inhibition at early delays and the best facilitatory delay (r = 0.87, P < 0.01, Fig. 5B). The strength of inhibition at early delays was not significantly correlated with any measure of delay-tuned facilitation. This suggests that timing of inhibition at early delays, but not its strength, can affect the best delay of facilitation.
|
Surprisingly, there was little evidence that the width of the facilitated delay curve was influenced by early or later inhibition. Thus there was no significant difference in facilitatory delay width among units without inhibition, units with inhibition at early delays, and units with inhibition at later delays (F[2,86] = 0.6, P > 0.5). No difference existed even between units lacking inhibition and units having both early and later inhibition. Further, there was no significant correlation between any measure of early inhibition and the delay width of facilitation. These analyses suggest that the duration of delay-tuned facilitation, expressed as the delay width of facilitation, is not limited by early or later inhibition. However, it is noteworthy that the range of facilitatory delay widths for units without inhibition at later delays was broader (0.4–19.7 ms) than the range of facilitatory delay widths for units with inhibition at later delays (0.6–11.8 ms). This suggests that inhibition occurring at delays greater than the best facilitatory delay could potentially limit the width of long-duration facilitatory interactions but did not have this effect on most units.
There were no significant differences in the strength of facilitatory interactions among units without inhibition, units with inhibition at early delays, and units with inhibition at later delays (F[2,88] = 0.1, P > 0.5). Moreover, there was no significant correlation between the strength of facilitation and any feature of early inhibition. Thus early inhibition had no measurable effect on the strength of the facilitatory interaction.
Because inhibition at later delays, by definition, occurs at delays later than both the best facilitatory delay and the best early inhibitory delay, we found significant correlations between: the best facilitatory delay and best later inhibitory delay (r = 0.84, P < 0.01), the best early inhibitory delay and best later inhibitory delay (r = 0.72, P < 0.05), and the width of inhibition at early delays and best later inhibitory delay (r = 0.8, P < 0.05). The strength of inhibition at later delays was not significantly correlated with either the best facilitatory delay (r = 0.10, P > 0.1) or the strength of facilitation (r = 0.15, P > 0.1). However, the width of inhibition at later delays was significantly correlated with the delay width of facilitation (r = –0.53, P < 0.05, Fig. 5C) and strength of facilitation (r = –0.50, P < 0.05, Fig. 5D). This suggests that the inhibition at later delays depends to some extent on the strength and timing of the facilitatory interaction.
Effect of drugs on delay-tuned combination-sensitive responses
We tested the role of glycine and GABA in facilitatory combination-sensitive units by examining their response properties before and during the application of STRY (n = 23), BIC (n = 25), or both (BIC+STRY: n = 20).
Application of STRY eliminated some but not all features of the delay-sensitive response to combinations of tones. For the unit in Fig. 6A, the facilitatory peak, occurring in control tests at delays centered at 10 ms, was completely eliminated. However, the inhibition at 0 ms delay was unaffected (control interaction index: –0.65, STRY: –0.63). Figure 6B shows the effect of STRY in a facilitated unit in which the lower-frequency facilitating sound was tuned to frequencies outside the sonar range. The unit, recorded in the 80–90 kHz tonotopic representation of the IC, displayed two excitatory responses tuned to 20.3 and 85.6 kHz. When the two signals were presented simultaneously (delay = 0 ms), the response of the unit was facilitated. Application of STRY eliminated the facilitation (Fig. 6B).
|
|
Table 1 provides a general summary of the effects of STRY and BIC on facilitation and the accompanying inhibition at early and later delays. Facilitation was strongly dependent on glycinergic inhibition but much less dependent on GABAergic inhibition. STRY eliminated facilitation in 96% of the units when applied alone and in 100% of the units when applied in combination with BIC. The loss of facilitatory interactions with application of STRY is consistent with earlier results ( Wenstrup and Leroy 2001). In contrast, BIC eliminated facilitation in 28% of the units. In most units, inhibition at early delays was not eliminated by STRY, BIC, or BIC+STRY. Inhibition at later delays was largely dependent on glycinergic inhibition. Thus STRY eliminated later inhibition in 78% of tested units and BIC+STRY eliminated inhibition in 83% of tested units, whereas BIC eliminated later inhibition in only 20% of the units. In the following text, we quantitatively assess the effects of the drugs on facilitation and the accompanying inhibition.
|
Application of STRY eliminated facilitation in 22 of 23 units tested (Table 1) and severely reduced the strength of facilitation in the remaining unit (change in interaction index from 0.76 to 0.12, Fig. 8A). Across the sample, the average strength of facilitation was significantly reduced (control mean: 0.27, STRY mean: –0.03; P < 0.001, paired t-test). In five units, the elimination of facilitation revealed an underlying inhibitory interaction at the best delay.
|
Of 25 units tested with BIC application, facilitation was eliminated in seven units (Table 1). Across the sample, the strength of facilitation was significantly reduced (control mean: 0.32, BIC mean: 0.16; P < 0.01, paired t-test), but the mean for BIC application nonetheless corresponded to facilitation. For the 18 units in which BIC did not eliminate facilitation, the mean change in interaction index was only 0.06, corresponding to a change in facilitatory strength of 
13% (control mean: 0.33, BIC mean: 0.27; P < 0.01, paired t-test). Application of BIC significantly increased the response magnitudes of units, both to the combination stimulus at best delay (average increase, 82%) and to the BF signal alone (181%; P < 0.05, paired t-test, Fig. 8, E and H). For units tested with BIC, there was no significant correlation between the change in facilitation strength and changes in either the response to the BF stimulus (r = –0.38; P > 0.05) or the response to the combination stimulus at best facilitatory delay (r = –0.13; P > 0.05). Of the seven units in which BIC eliminated facilitation, rate-level functions were obtained in four units with BIC. In three of the four units, a comparison of the rate-level and delay functions suggested that the loss of facilitation was not due to saturation effects. In the rest of the units, we could not determine whether the loss of delay-tuned facilitation is due to application of BIC or due to saturation of the unit's response. Despite the changes in response rates, other measures of combination-sensitive facilitation were not affected. For the 18 units in which BIC did not eliminate facilitation, BIC did not alter the best facilitatory delay (P > 0.1, paired t-test, Fig. 9A) or the range of delays over which facilitation occurred (delay width: control mean: 6.9 ms, BIC: 6.9 ms, P > 0.1, paired t-test, Fig. 9B). These data suggest that the features of facilitatory combination-sensitive interactions are generally independent of GABAergic inhibition.
|
Quantitative effects of drugs on inhibitory interactions
INHIBITION AT EARLY DELAYS. The effect of drug application on the strength and the duration of early inhibition was tested in 30 units. In all 15 units tested with STRY, inhibition at early delays was not eliminated (Table 1). However, the average strength of inhibition, expressed as the interaction index, was significantly reduced from –0.62 to –0.41 (P < 0.01, paired t-test, Fig. 10A). In somewhat over half of the units (60%, n = 9), the strength of inhibition decreased by >20%, but inhibitory strength was relatively unchanged in the remainder. In 10 of the units tested with STRY, the delay width of inhibition at early delays could be measured. There was no significant reduction in this inhibitory delay width due to application of STRY (control mean: 10.9 ms, STRY mean: 7.7 ms, P > 0.05, paired t-test, Fig. 10G). There was also no significant correlation between the change in strength of inhibition and change in duration of inhibition due to application of STRY (r = –0.1; P > 0.05).
|
After combined application of BIC and STRY, inhibition at early delays persisted in 9 of 11 units (Table 1). However, the average strength of inhibition was significantly reduced, from –0.65 to –0.33 (P < 0.05, paired t-test, Fig. 10C). As with separate BIC and STRY application, BIC+STRY had varied effects on the strength of this inhibition across the sample. For six units in which the delay width of inhibition at early delays could be measured, there was no significant reduction due to BIC+STRY (control: 9.2 ms, BIC+STRY: 4.7 ms, P > 0.05, paired t-test, Fig. 10I). There was no significant correlation between the change in strength of inhibition and change in the duration of inhibition due to application of BIC and STRY (r = –0.03; P > 0.05).
The results from application of the drugs suggest that the rapid inhibition at early delays depends, in part, on inhibitory mechanisms that are independent of glycine or GABAA receptors in the IC. Nonetheless, in most units, the action of glycine or GABAA receptors in the IC may contribute to this inhibition.
INHIBITION AT LATER DELAYS. Inhibition at delays later than the best facilitatory delay was tested in 19 units. For the unit shown in Fig. 11, application of STRY eliminated the inhibition at later delays (delays between 14 and 16 ms). Further, addition of BIC did not change the effect of STRY. STRY application eliminated inhibition at later delays in six of nine units tested (Table 1) and reduced the inhibition in the remaining three units. Across the sample, the average strength of inhibition was significantly reduced from –0.36 to –0.11 (P < 0.01, paired t-test, Fig. 10D).
|
Combined application of BIC and STRY showed effects similar to STRY application. In 10 of 12 units (Table 1) tested, BIC+STRY application eliminated the inhibition at later delays. Across the sample, the average strength of inhibition was significantly reduced from –0.36 to –0.09 (P < 0.01, paired t-test, Fig. 10F). These results suggest that the inhibition at later delays is more dependent on glycinergic inhibition than on GABAergic inhibition.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Inhibitory interactions in facilitatory combination-sensitive IC neurons
Both BF and low-frequency inputs to combination-sensitive neurons may activate excitation and inhibition. Nonetheless, the interactions revealed in the delay sensitivity functions (e.g., Fig. 2) are largely the result of the excitation and inhibition evoked by the lower-frequency signal because in most neurons the response to the higher-frequency signal is significantly larger than the response to the lower-frequency signal. Three-quarters of the neurons (76%) displayed inhibition in addition to the response facilitation evoked by a low-frequency signal. Inhibition at early delays was more common, occurring in 59% of all neurons, whereas inhibition at later delays occurred in 37% of the neurons. One-fifth of neurons showed inhibition at both early and later delays.
There are some clear relationships between inhibitory properties and facilitation. The strongest relates the inhibition at early delays and the timing of the facilitation. First, as has been reported previously ( Olsen and Suga 1991b; Portfors and Wenstrup 1999), neurons with inhibition at early delays have on average significantly longer best facilitatory delays than do neurons without inhibition at early delays. Second, there is a strong positive correlation between the duration of the inhibition at early delays and the best facilitatory delay. These results indicate that the temporal positioning of the response peak (in terms of delay) is dependant on the inhibition, as has been observed in neurons sensitive to interaural time differences (ITDs) in gerbil medial superior olive ( Brand et al. 2002) and barn owl external nucleus of IC ( Zheng and Knudsen 1999). Further, these results suggest that some component of the inhibition at early delays determines the timing of the low-frequency evoked excitation, a suggestion in accord with models of postinhibitory rebound used to explain response facilitation ( Casseday et al. 1994; Wenstrup and Leroy 2001).
A second clear relationship exists, although not as strong, between the strength and timing of the facilitation and the subsequent inhibition at later delays. As expected, the best longer inhibitory delay is well correlated with the best facilitatory delay, but the results suggest that the inhibition is also influenced by the duration (e.g., delay width) and strength of the facilitatory interaction. The later inhibition, or suppression, could thus be activated by the facilitation within the IC neuron. This possibility is consistent with our observation that the inhibition at later delays has little influence on the best facilitatory delay of these units. It is also consistent with our iontophoretic results showing that strychnine usually eliminated this inhibition. We believe this occurred because strychnine removed the facilitation that may evoke this later inhibition.
Some expected relationships did not materialize. For example, there was no relationship between the strength of inhibition at early delays and any measure of the subsequent facilitation. We had hypothesized that the strength of early inhibition may alter the facilitation because glycinergic input appears to be an essential component of facilitatory interactions. However, we detected no effect. One possible explanation is that the early inhibition is the result of multiple interactions at different levels of the auditory brain stem and midbrain. The glycinergic inhibition acting on IC neurons may form only one component of this early inhibition, with the result that its correlation with delay-tuned facilitation was obscured.
We also hypothesized that the low-frequency inhibition would restrict the duration of delay-tuned facilitation (delay width). This corresponds to a presumed role of lateral inhibition that is characteristic of many frequency tuning curves in the auditory system or the center-surround organization of retinal ganglion cells and lateral geniculate neurons in the visual system. Although neurons with the longest best delays (15 ms) all lacked low-frequency inhibition at later delays, there was no relationship between the duration of the facilitatory interaction and the presence or absence of inhibition. In terms of encoding properties, these observations suggest that low-frequency inhibition does not function to restrict the range of delays to which the neuron responds but instead acts to enhance the contrast between preferred delays and nonpreferred delays. By comparison, inhibition acting on IC neurons in the barn owl that are sensitive to interaural timing differences (ITD) both enhances contrast and sharpens delay selectivity ( Fujita and Konishi 1991).
Origins of facilitatory and inhibitory interactions
The present results support the hypothesis that combination-sensitive facilitation originates within the mustached bat's inferior colliculus. Wenstrup et al. (1999) proposed that the facilitatory interactions are created in the IC by the convergence of high- and low-frequency brain stem inputs. The elimination of facilitatory interactions by strychnine application in this and a previous study ( Wenstrup and Leroy 2001) provide strong support for that proposal. Moreover, there are very few facilitatory interactions in nuclei of the lateral lemniscus ( Portfors and Wenstrup 2001) and none in the cochlear nucleus ( Marsh and Wenstrup 2002), both major sources of input to IC combination-sensitive neurons ( Wenstrup et al. 1999). Although combination-sensitive facilitatory interactions were first reported in the auditory cortex of this animal ( Suga et al. 1978) and subsequently in the MGB ( Olsen and Suga 1991a, b), physiological, anatomical, and pharmacological evidence all support the conclusion that these facilitatory interactions originate in the IC.
In contrast to facilitatory interactions, the inhibition at early delays observed in facilitated combination-sensitive neurons appears to be created at levels below the IC. In nearly all units, application of strychnine, bicuculline, or both together failed to eliminate the inhibition at early delays. These results are consistent with physiological studies showing the presence of combination-sensitive inhibition in both the cochlear nucleus ( Marsh and Wenstrup 2002) and lateral lemniscal nuclei ( Portfors and Wenstrup 2001). In these brain stem neurons, responses to BF sounds are suppressed by simultaneous lower-frequency signals with temporal sensitivity that matches the early inhibition described in this study. We believe that such neurons send excitatory projections to the IC, as described later in DISCUSSION, and dominate the excitatory response of facilitatory neurons that show early inhibition. However, while the strength of early inhibition in the IC facilitatory neurons was not eliminated, it was significantly reduced in the population by the application of the drugs. There are two possible explanations of this result. First, low-frequency-tuned glycinergic and GABAergic input from brain stem nuclei terminate onto the high-frequency-tuned neurons of the IC to create a purely inhibitory effect. If true, such inhibitory interactions within the IC repeat the interactions occurring in the auditory brain stem. A second explanation is that high-frequency-tuned glycinergic and GABAergic inputs preferentially suppress high-frequency excitatory inputs that do not show combination-sensitive inhibition. Blockade of the inhibitory inputs would result in less inhibition at early delays. Neither possibility can be ruled out.
Like facilitation, inhibition at later delays was eliminated in the majority of neurons tested with STRY (67%), and BIC+STRY (83%). In contrast, BIC eliminated inhibition at later delays in only 20% of the neurons tested. These data suggest that inhibition at later delays originates in the IC and may be related to mechanisms creating the facilitation.
Special contribution of glycinergic input to combination-sensitive facilitation in the IC
The majority of studies addressing the physiological response properties of auditory neurons in the IC show similar roles for glycine and GABAA receptors. Antagonists of both receptors result in changes in temporal response patterns, evoked discharge rate, duration tuning, and latencies ( Burger and Pollak 1998; Casseday et al. 1994, 2000; Fuzessery et al. 2003; LeBeau et al. 1996, 2001; Vater et al. 1992). In the present study, we also observed similar effects on the BF response when we blocked glycine or GABAA receptors, particularly with respect to temporal response patterns and discharge rate.
In contrast, this study demonstrates a much more substantial role for glycinergic input in combination-sensitive, facilitatory interactions in the auditory midbrain. In this and a previous study ( Wenstrup and Leroy 2001), blockade of glycine receptors by strychnine always eliminated or drastically reduced facilitation. Blocking GABAA receptors by bicuculline eliminated facilitation in about one-quarter of the neurons tested, and in some of these, there were strong indications that the loss of facilitation was due to rate saturation effects. Moreover, in neurons where bicuculline did not eliminate facilitation, the addition of strychnine did. This demonstrates that facilitation in most neurons requires glycinergic input and does not require GABAergic input. The minor role of GABAA-mediated mechanisms is reinforced by our observations that BIC had little effect on response latency, best facilitatory delay, facilitatory delay width, or strength of facilitation. Such specific effects of glycinergic inputs to the IC have not been reported previously. A subsequent section discusses possible mechanisms for this specificity.
An alternative to the conclusion that glycine contributes to response facilitation in the IC is that the observed effects of strychnine to eliminate facilitation can be explained by its action to increase neuronal discharge rate. This saturation hypothesis thus considers that the facilitatory interactions are still present during strychnine application but views the loss of measurable facilitation as resulting from the neuron's inability to increase its discharge rate beyond the response to the BF tone. Although we cannot discount this possibility for each recorded unit, the behavior of the population is not consistent with it. First, the saturation hypothesis would predict that the increase in discharge rate to BF or combination tones would be strongly correlated with the reduction in measurable facilitation. There was no such correlation in any drug-application regime. Moreover, application of BIC raised response rates much more than did STRY but eliminated facilitation in many fewer units. Finally, for all units in which facilitation was removed by STRY or BIC+STRY, the rate-level functions suggested that the units were able to respond at higher discharge rates but did not. This occurred, we believe, because the facilitatory mechanism was blocked by STRY.
For inhibition at early delays, a differential role of glycinergic and GABAergic mechanisms in the IC was not observed. Both STRY and BIC failed to eliminate inhibition at early delays but caused a similar reduction in the strength of inhibition. For whatever mechanism underlies this reduction in inhibition (see previous section), the results suggest that glycinergic and GABAergic mechanisms contribute similarly.
Inhibition at later delays appears to be substantially more sensitive to the application of STRY than to BIC. This suggests that inhibition at later delays, like facilitatory interactions, is mediated by glycinergic mechanisms in the IC and may be mechanistically related to facilitation.
Neural mechanisms underlying facilitatory interactions
Suga and colleagues ( Olsen and Suga 1991b; Suga et al. 1990) proposed that combination-sensitive facilitatory interactions occur when excitations in response to lower- and higher-frequency sounds coincide. This coincidence model predicts that the difference in latencies of excitation to the two signals will match the best facilitatory delay, a prediction confirmed in both the IC ( Portfors and Wenstrup 1999) and MGB ( Olsen and Suga 1991b). Because many neurons show facilitatory peaks at positive delays (i.e., when the BF sound follows the low-frequency sound), the excitation due to the low-frequency signal must be delayed to coincide with the excitation due to the BF signal. The delay at which facilitation occurs is a key feature of combination-sensitive neurons, encoding the distance of sonar targets or the timing of spectral elements in social vocalizations. Mechanisms that seek to explain these facilitatory interactions should also explain how a range of low-frequency latencies is created. In the following text, we discuss possible mechanisms by which glycinergic input may contribute to these features of delay-tuned facilitation.
DISINHIBITION HYPOTHESIS—GLYCINE IS INHIBITORY. A potential explanation of the experimental results is that both the low- and high-frequency excitations are created by excitatory neurotransmitters and that glycinergic input has a purely inhibitory effect within a disinhibitory network. As illustrated in Fig. 12A, facilitation is created due to the coincidence of excitations due to high- and low-frequency inputs on the combination-sensitive neuron. However, the GABAergic inhibition (timed to the low-frequency input) suppresses the facilitation at all delays except at the best delay. At the best delay, the glycinergic inhibition (timed to the high-frequency input and acting on GABAergic neuron), suppresses the GABAergic inhibition on the combination-sensitive neuron, thus allowing facilitation to be expressed. While the predictions of the model for the application of STRY and BIC agree with experimental data, the model fails to explain the effect of BIC+STRY. We have also considered other disinhibition models based on the removal of GABAergic inhibition (in this case a constant inhibition or one timed to the high-frequency response) by appropriately timed glycinergic inhibition. However, the predictions of these models fail to explain the results observed with the application of BIC. Thus we believe that facilitatory interactions are unlikely to be created by disinhibition mechanisms.
|
). The enhancement of the EPSC is thought to result from a depolarizing effect of glycine on the presynaptic terminal due to a more positive chloride equilibrium potential, and the depolarization increases the release of glutamate at the synaptic terminal. A similar mechanism could explain the role of glycine in combination-sensitive facilitation. In the model illustrated in Fig. 12B, an appropriately timed, low-frequency-tuned glycinergic input, located on the
and 
, respectively. Numbers to the right of each PSTH indicate the spike count. Dashed vertical lines in some PSTHs indicate the time window within which response magnitude was evaluated. In these cases, spike counts and delay curves are based on time-windowed responses.
and error bars indicate the mean and SE for the sample. Asterisks indicate significant differences between the control and drug tests using paired t-test (* P < 0.05, ** P < 0.01). D–F: response magnitude to combination tones at the best facilitatory delay before and during application of STRY (D), BIC (E), and BIC+STRY (F). For each group, drug application caused a significant increase in the response magnitude. G–I: response magnitude to high-frequency tone before and during application of STRY(G), BIC (H), and BIC+STRY (I). For each group, drug application caused a significant increase in the response magnitude. J–L: response magnitude to low-frequency tone before and during application of STRY (J), BIC (K), and BIC+STRY (L). Drug application failed to significantly alter the low-frequency response in any group.
