Role of the Dorsal Cortex of the Inferior Colliculus in the Precedence Effect

Introduction

In a reverberating environment, a direct wavefront generated from a sound source reaches the listener (leading sound), and other parts bounce off obstacles (eg, tables and walls), forming countless reflected sounds (lagging sounds). Although the listener receives a considerable level of lagging sounds carrying different directional cues, the auditory system enables the listener to locate the sound source accurately by suppressing lagging sound interferences. This physiological phenomenon is called the precedence effect (PE) [1,2]. When the inter-stimulus delay (ISD) between leading and lagging sounds is brief, the listener hears only 1 fusion sound, and when the leading and lagging sounds are just distinguishable, the corresponding ISD is called the echo threshold [3].

Electrophysiological studies in rabbits, cats, and rats have demonstrated that neurons in each nucleus of the auditory pathway, particularly in the inferior colliculus (IC), exhibit a phenomenon called lag suppression [4–7], which is the neural basis of the PE [8]. Scholars proposed that inhibitory projections contribute a significant role in mediating lag suppression, by studies in bats and cats [9–11]. The IC receives ascending projections, descending projections, and projections from the subdivisions of the IC and contralateral IC [8,12–14]. The ascending inhibitory projections play important roles in PE formation in the IC, for example, the ascending gamma-aminobutyric acid (GABA)ergic (ie, mediating the effects of GABA) projections from the dorsal nucleus of the lateral lemniscus, ventral nucleus of the lateral lemniscus, and superior paraolivary nucleus [1,4,12,15,16] and glycinergic (ie, mediating the effects of glycine) projections from the lateral superior olive and ventral nucleus of the lateral lemniscus [4,17]. Similarly, descending projections from the auditory cortex to the shell of the IC play a certain role in regulating the acoustic processing of the central nucleus of the IC (CNIC) [18,19] and may indirectly regulate PE encoding in the CNIC [8]. Although the descending projections from the auditory cortex terminate to the shell of the IC and the CNIC [20,21], the shell of the IC is the main region for the descending feedback information from the auditory cortex [22–25]. The shell of the IC is divided into the lateral cortex of the IC and the dorsal cortex of the IC (DCIC) [26]. The lateral cortex of the IC receives auditory and multisensory inputs, such as visual and somatosensory, while the DCIC predominantly receives auditory input from the auditory cortex [27–29]. Furthermore, projections from the contralateral IC pass through the DCIC [14]; therefore, the DCIC is a very critical site in the auditory pathway. However, the role of the DCIC in PE formation remains unclear.

Therefore, the objective of this study was to understand the role, if any, of the DCIC on PE formation in the CNIC of rats. An in vivo 16-channel electrophysiological recording technique was performed to investigate the neuronal responses to paired sounds at different ISDs, and Western blot analysis was used to obverse the expression of inhibitory receptors in the CNIC, after inducing electrolytic lesions in the DCIC. We propose that the DCIC potentially modulates PE encoding and formation in the CNIC, and clarifying this will lay a foundation for further exploring the mechanism of the PE in the IC and enhancing our understanding of the mechanism underlying sound localization.

Material and Methods

ETHICS STATEMENT:

All animal experiments were approved by the Animal Experiments and Experimental Animal Welfare Committee of Capital Medical University (AEEI-2021-015) and complied with the Animal Research Reporting of In Vivo Experiments guidelines.

ANIMAL COLLECTION AND REARING:

Healthy male Sprague Dawley rats (8 weeks; 290–340 g) from Capital Medical University were used. Each rat was randomly assigned to either the model or control group. For electrophysiological experiments, there were 6 rats in the model group and 5 rats in the control group. The sample size referred to previous similar studies [7,30]. For Western blot analysis, there were 6 rats in the model and control groups, separately. All rats were housed in a temperature-constant room (24°C, 12-h light/dark cycle), with food and water provided ad libitum.

SURGICAL PROCEDURE AND ELECTRODE IMPLANTATION:

The rats were anesthetized with the induction dose of sodium pentobarbital (50 mg/kg), and their temperature was maintained at 37°C, using a feedback-controlled blanket. The rats were fixed in an RWD stereotaxic apparatus (RWD Life Science, Shenzhen, China), using zygomatic arch bars (patent application no. 201020600962.8) [6]. We located the DCIC and CNIC based on the stereotaxic coordinates [31]: with bregma serving as the reference for each plane, (1) DCIC: 8.4 to 9.12 mm anteroposterior; 0.7 mm mediolateral, and 3.5 mm dorsoventral; and (2) CNIC: 8.4 to 9.12 mm anteroposterior, 2 mm mediolateral, and 4.5 mm dorsoventral. In the model group, a tungsten wire electrode was inserted into the right DCIC, and a continuous 3-mA current was applied for 10 s (Isolated Pulse Stimulator Model 2100, A-M Systems, USA) [32,33]. Then, a 16-channel recording electrode was embedded in the right CNIC. The same procedure was performed in the control group, but without electricity. Animals were allowed to recover for 2 weeks.

STIMULI AND APPARATUS:

The sound stimuli included a single white noise burst of 4 ms and paired white noise bursts of 4 ms at different ISDs (2, 4, 6, 8, 10, 20, 50 ms), with equal levels and coherent phases. The paired sound stimuli were repeated 50 times with 500-ms intervals separating each pair, which totally took 25 s. Then, the successive 25-s bursts were repeated with a 30-s interval to prevent IC neural fatigue. The sound stimuli were generated digitally at a sampling rate of 192 kHz, using Adobe Audition 2019 (Adobe Systems Inc, CA, USA). The electrostatic speakers (DrumBass III BT, Lifetrons, St. Gallen, Switzerland) were located approximately 25 cm from the center of the rat’s head in the soundproof chamber (1.2×1.0×0.9 m). One rat was kept inside the soundproof chamber for recording at a time. The single 4-ms stimulus was played from 1 of the 2 loudspeakers at different azimuths. The leading and lagging sounds were played through 2 electrostatic speakers at different azimuths (Table 1). The speaker outputs were calibrated using a sound level meter (Type HS5670A, Hongsheng, Zhejiang, China), and sound stimuli were presented at a 70-dB sound pressure level.

DATA ACQUISITION:

Recording started 1 h after the induction dose of sodium pentobarbital (50 mg/kg) and 20 min after the maintenance dose (20 mg/kg), as in our previous study [6]. The rats were still under anesthesia at the time of recording. Thereafter, the auditory brain response was tested in each animal. Finally, the brains were removed by rapid decapitation, when the rats were under anesthesia, and flash frozen in liquid nitrogen. Cytochrome oxidase staining was performed on frozen coronal brain tissue sections (thickness, 35 μm) to distinguish the DCIC and CNIC [34]. If the electrolytic lesion site or recorded location were found to be incorrect, the rat would not be included in the experiment. The data from the extracellular recordings were collected at a sampling rate of 40 kHz. Continuous data were acquired with a bandpass filter (300–6000 Hz), to extract neural spike signals. Spike signals were further sorted using the Valley Seeking algorithm and were sent to NeuroExplorer and MATLAB for quantifying neuronal firings. We used manual thresholding to detect spikes at least 2 standard deviations over the baseline hash [35]. The normalized neural responses to the lagging sound were calculated as the ratio of the spike numbers evoked by the lagging sound to the spike numbers evoked by a single stimulus. When the ratio was less than 1.0, it indicated that responses to the lagging sound were suppressed [36]. Moreover, the half-maximal ISD were calculated as neuronal responses reaching 50% recovery to the lagging sound [36].

WESTERN BLOT:

The CNIC tissue of each rat [37] was homogenized in enhanced radio immunoprecipitation assay lysis buffer (Applygen, Beijing, China). Protein concentrations were determined using a bicinchoninic acid kit (Thermo Fisher Scientific, MA, USA). A protein sample (45 μg) was loaded onto 8% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto polyvinylidene difluoride membranes (Millipore, Burlington, MA, USA). After washing with Tris-buffered saline containing 0.05% Tween 20, the membranes were blocked in NcmBlot Blocking Buffer (NCM Biotech, Suzhou, China) for 15 min at approximately 25°C. Then, the membranes were cut and incubated overnight at 4°C with the following primary antibodies: GABA A receptor α1 (GABAARα1) (1: 400, 12410-1-AP, Proteintech, Chicago, IL, USA); GABA B receptor 2 (GABABR2) (1: 500, ab75838, Abcam, Cambridge, UK); and β-actin (1: 10 000, 66009-1-Ig, Proteintech). Horseradish peroxidase-conjugated goat anti-rabbit secondary antibodies (1: 2000, SA00001-2, Proteintech) were used for 2 h at room temperature. The membranes were developed using an enhanced chemiluminescence kit (Applygen), and images were acquired using a gel imaging system (ProteinSimple, San Jose, CA, USA). The relative expression levels of different proteins were normalized to that of β-actin.

STATISTICAL ANALYSIS:

The results were analyzed using NeuroExplorer (Nex Technologies, Colorado Springs, CO, USA), Offline Sorter (Plexon Inc, TX, USA), and MATLAB (MathWorks, Natick, MA, USA). Graphs were constructed using Prism 8 (GraphPad, San Diego, CA, USA) and SPSS 24.0 software (IBM, Armonk, NY, USA). All values are expressed as the mean ± standard error of the mean. The normality of the data was assessed using the Shapiro-Wilk test. The t test was used to compare 2 groups at different ISDs. The data from multiple groups were compared using one-way analysis of variance, followed by the Bonferroni post hoc test. A statistical significance level of P<0.05 was set for all analyses.

Results

HISTOLOGICAL EXAMINATION OF THE ELECTRODE PROBE SITES:

Histological evaluation showed that electrodes were implanted correctly in the target structures. A 16-channel multielectrode probe was inserted into the CNIC in the control and model groups (Figure 1), and a tungsten wire electrode was used to induce electrolytic lesions into the DCIC (Figure 1B).

DEGREE OF LAG SUPPRESSION DEPENDED ON THE LOCATION OF LEADING SOUND:

We investigated the electrical activity exhibited by the CNIC neurons in response to different paired sound stimuli. Small spontaneous firing was observed when the neurons were at rest, while regular firings were observed after the acoustic stimuli were delivered (Figure 2A). Figure 2B shows the leading and lagging sounds evoked potentials. At 2 ms ISD, the leading and lagging sounds made neurons fire once, while neuron firings were induced twice at 50 ms ISD. However, the amplitude of the evoked potential by the leading sound at ipsilateral 18° was lower than that of the lagging sound at contralateral 18°.

The neuronal responses to paired stimuli at different ISDs revealed 2 distinct spike clusters (Figure 2C). One cluster was evoked by the leading sound and the other by the lagging sound. At a 2 ms ISD, the lagging sound was primarily suppressed. However, as the ISDs increased, the lagging sound evoked spike numbers gradually increased. The spike numbers evoked by the lagging sound at ipsilateral 18° were lower than those evoked by the lagging sound at contralateral 18°. This was more pronounced at ipsilateral 45° and contralateral 45°. When the leading-lagging sounds were located at contralateral-ipsilateral 18°, compared with ipsilateral-contralateral 18°, the normalized neuronal response of CNIC neurons decreased (Figure 2D). Statistical differences were found when the ISD was 4 ms (P=0.049), 6 ms (P=0.0028), and 8 ms (P=0.043). The half-maximal ISD was shorter when the leading sound was located at the ipsilateral 18° (3.966±1.353 ms) than that at the contralateral 18° (9.474±2.361 ms, P=0.022). Similarly, when the leading-lagging sounds were located at the contralateral-ipsilateral 45°, compared with ipsilateral-contralateral 45°, the normalized neuronal response of CNIC neurons decreased (Figure 2D). Significant differences were found when ISDs were 2 ms (P<0.001), 4 ms (P<0.001), 6 ms (P<0.001), 8 ms (P<0.001), 10 ms (P<0.001), and 20 ms (P=0.039). The half-maximal ISD was significantly shorter when the leading sound was located at the ipsilateral 45° (1.104±0.841 ms) than at the contralateral 45° (12.999±1.444 ms, P<0.001).

ELECTROLYTIC LESIONS IN THE DCIC RESULTED IN WEAKENED LAG SUPPRESSION:

After electrolytic lesions were induced in the DCIC, lag suppression persisted. Spike trials were evoked by the different paired sound stimuli in the CNIC neurons (Figure 3A). The evoked potential was shown when the leading-lagging sounds were located at contralateral 45°-contralateral 18° and contralateral 18°-contralateral 45° (Figure 3B). However, the amplitude of the evoked potential by the lagging sound was lower than that evoked by the leading sound both at contralateral 45° and contralateral 18°.

There were 2 clear spike clusters when leading-lagging sounds were at either contralateral 45°-contralateral 18° or contralateral 18°-contralateral 45°, and the number of the lagging sound evoked spikes increased after electrolytic lesions were induced in the DCIC (Figure 3C). When the leading-lagging sounds were located at the contralateral 45°-contralateral 18°, the normalized neuronal response in the model group significantly increased compared with that in the control group (Figure 3D). Significant differences were found when the ISD was 4 ms (P=0.005) and 6 ms (P<0.001). The half-maximal ISD in the model group (4.184±0.523 ms) was shorter than that in the control group (5.448±0.299 ms, P=0.002). Similarly, when the leading-lagging sounds were located at the contralateral 18°-contralateral 45°, the normalized neuronal response in the model group increased, compared with that in the control group (Figure 3D). A significant difference was found when the ISD was 4 ms (P=0.009). The half-maximal ISD in the model group (3.642±0.450 ms) was shorter than that in the control group (4.439±0.239 ms, P=0.008).

:

We examined the GABAARα1 and GABABR2 expression levels in the CNIC for the control and model groups (Figure 4A). There was no significant difference between the right and left CNIC in the control group. However, GABAARα1 expression (Figure 4B) in the right CNIC of the model group was significantly reduced, compared with that in the left CNIC of the model group (P=0.017) and right CNIC in the control group (P=0.040). Moreover, GABABR2 expression (Figure 4C) in the right CNIC of the model group was significantly reduced, compared with that in the left CNIC of the model group (P=0.004) and right CNIC in the control group (P=0.026).

Discussion

Owing to its complexity, the neuronal mechanism underlying the PE in the IC remains poorly understood. Several studies have investigated the role of the ascending projections in PE formation in the CNIC [1,4,8,12]. To date, this study is the first to provide the role of the shell of the IC on PE formation in the CNIC of rats. We found that the DCIC potentially modulates PE formation in the CNIC, and it may be an important site to study the mechanism of PE in the IC.

In the first part, we presented leading-lagging sounds at different azimuths: contralateral-ipsilateral 18° and contralateral-ipsilateral 45°, as in previous studies on the PE in free fields [6,38]. Irrespective of whether the leading-lagging sounds were located at contralateral-ipsilateral 18° or contralateral-ipsilateral 45°, we observed lag suppression in the CNIC neurons. However, when the leading and lagging sounds were presented at different azimuths, the lag suppression was significantly different, especially at contralateral or ipsilateral 45°, which is consistent with previous findings [1,6]. This phenomenon indicated that CNIC neurons were sensitive to the contralateral sound but exhibited a poor response to the ipsilateral sound.

The reason for this phenomenon may be the monaural response properties of CNIC neurons. The CNIC neurons are sensitive to contralateral sound and insensitive to ipsilateral sound [39]. Early studies focused on the differences in neuron type in the CNIC and found most of the neurons in the CNIC were excitatory-inhibitory neurons, which become excited by contralateral sound and inhibited by ipsilateral sound [40,41]. Therefore, when the leading sound was on the ipsilateral side and the lagging sound on the contralateral side, the amplitude of the leading sound evoked spike was small. In this case, for the lagging sound, it was easy to induce neuronal firing, and the amplitude of the evoked spike was high. Furthermore, the electrical activity evoked by the leading sound was weak, leading to the lag suppression reduction.

Moreover, when the leading-lagging sounds were located at contralateral-ipsilateral 45° or ipsilateral-contralateral 45°, the difference in lag suppression was considerably obvious; for example, the spike amplitude of the contralateral sound evoked was higher than that evoked by the ipsilateral sound. Therefore, to explain the complex mechanisms underlying the PE in the IC and to avoid the monaural response properties of CNIC neurons, we placed the leading and lagging sounds both at the contralateral side of the recording site in free fields. When the leading-lagging sounds were located at either contralateral 45°-contralateral 18° or contralateral 18°-contralateral 45°, we observed lag suppression in the CNIC neurons. However, there was no significant difference between the model and control groups when the ISD was 2 ms. This finding was likely attributable to the neuronal refractory period, besides lag suppression. Rare or no firing was induced when the sound stimulus appeared during the refractory period [6]. Therefore, although the DCIC lesion weakened lag suppression, it did not have a significant effect on the firing at 2 ms ISD.

However, as the ISD extended beyond the refractory period, lag suppression played a significant role. When the ISD was 8 ms, the echo threshold was approached, and the spikes evoked by the lagging sound only increased minimally. Additionally, the half-maximal ISD of the CNIC neurons in the model group was significantly shortened, compared with that in the control group, which is consistent with a previous finding [36]. Although the leading and lagging sounds were located at the contralateral side, the degree of lag suppression was still different, possibly because of the optimal azimuth [1]. Our results showed that lag suppression was significantly weakened because the inhibitory projection from the DCIC to the CNIC was disrupted. This finding is similar to the result reported by Tollin et al [3], as the dorsal nucleus of the lateral lemniscus lesion reduced lag suppression in the IC. Furthermore, the inhibitory neurons of the IC are mainly GABAergic neurons [42,43], and through in vitro patch-clamp experiments, we found that DCIC exhibited inhibitory GABAergic projections to the CNIC (data not shown). Therefore, our findings proposed the regulatory effect of the DCIC on PE formation in the CNIC, indicating that this structure can be an important candidate for PE study in the IC.

GABA receptors are the main receptors that mediate synaptic inhibition in the IC, including ionotropic GABAAR and metabotropic GABABR [37]. GABAARs mediate fast synaptic inhibition and play a critical role in regulating neuronal excitability and information processing in the auditory system [44]. GABABRs are distributed in presynaptic and postsynaptic neurons, with different functions. Presynaptic receptor excitation causes decreased Ca2+ influx, followed by decreased glutamate and GABA release [37]. Postsynaptic receptor excitation causes increased K+ channel opening, inhibiting the excitatory postsynaptic potential and resulting in delayed neural pulse conduction [37]. Therefore, GABA receptors play an important role in regulating the synaptic inhibition of the CNIC. In the present study, we found that the expression of GABAARα1 and GABABR2 in the CNIC was reduced after the DCIC lesion. This indicates that inducing electrolytic lesions in the DCIC affected the expression of inhibitory receptors in the CNIC, probably because the GABAergic inhibitory projection from the DCIC to the CNIC was disrupted. Once the expression of inhibitory receptors decreased, the balance of excitation and inhibition in the CNIC might be broken, resulting in a change in the processing and conduction of acoustic information. Results from Wang et al [36] demonstrated that the local application of GABA to the IC of rats decreased neural responses to lagging stimuli, raised the half-maximal ISD, and prolonged the recovery period for responses to lagging stimuli. Moreover, antagonizing these inhibitory GABA receptors in the IC leads to enhanced auditory activity, while stimulating them decreases it [45]. These findings are consistent with our Western blot analysis results, which showed that the expression of inhibitory GABA receptors in the CNIC after inducing electrolytic lesions in the DCIC decreased. This suggested that the inhibitory GABAergic projections might have been involved in the process.

This study was mainly to investigate the role of the DCIC in PE formation. The DCIC not only receives the descending projections from the auditory cortex [28] but also accepts projections from the contralateral IC through the commissure of the IC [46]. Commissural projections can affect the responses of IC neurons to sound frequency and intensity, and it plays a certain role in binaural sound processing and sound localization [47,48]. Due to the commissural fibers passing bilateral DCICs, electrical lesions into the DCIC must damage the commissural inputs from the contralateral IC. That also means the DCIC is actually a critical site for sound processing.

Notwithstanding, the present study has certain limitations. Previous research found that pentobarbital can prolong the recovery time of lagging stimuli to paired sources, producing PE illusions, while it was gradually attenuated during recovery from anesthesia. Thus, our recording started 1 h after the induction dose of sodium pentobarbital and 20 min after the maintenance dose, to avoid the effect of the sodium pentobarbital. However, in the absence of effective electrophysiological research models on the PE in the awake rat, the role of the DCIC in PE formation in this study was verified only from anesthetized rats. In addition, considering the fact that the transverse and sagittal sinuses of rats are almost directly above the DCIC, it was impossible to remove the dura for the injection of drugs or viruses; thus, we used a tungsten wire electrode to induce electrolytic lesions in the DCIC [32,33,49]. Nevertheless, results in this study suggest the role of the DCIC in studying the mechanisms of the PE in the CNIC of rats.

Conclusions

Currently, the mechanisms underlying the PE in the IC have not been fully elucidated. Our study suggests that the DCIC can modulate the formation of the PE in the CNIC, potentially through inhibitory GABAergic mechanisms. These findings position the DCIC as a key candidate in elucidating the neural mechanisms underlying the PE within the IC.