Spike Analysis of the Neural Activities Across the Rats' Auditory Brain Structures

Abstract

Tinnitus is a health condition that affects a large population. Clinical diagnosis and treatment have been developed for treating tinnitus for years. However, there are still limitations because researchers have yet to elucidate the mechanisms underlying how tinnitus neural signals develop in brain structures. Abnormal neural interactions among the brain areas are considered to play an important role in tinnitus generation. Researchers have been studying neural activities in the auditory brain structures, including the dorsal cochlear nucleus (DCN), inferior colliculus (IC), and auditory cortex (AC), to seek a better understanding of the information flow among these brain regions, especially in comparison with both health and tinnitus conditions. In this project, neural activities from the DCN, IC, and AC were collected and analyzed before and after the animals were noise-exposed and before and after their auditory cortices were electrically stimulated. These conditions in rats were used to estimate healthy animals, noise-trauma-induced tinnitus, and after auditory cortex electrical stimulation (ACES) treatment. The signal processing algorithms started with the raw measurement data and focused on the local field potentials (LFPs) and spikes in the time domain. The firing rate, shape of spikes, and time differences among channels were analyzed in the time domain, and phase–phase correlation was used to test the phase-frequency information. All the analysis results were summarized in plots and color-heat maps and also used to identify if any neural signal differs and cross-channel relation changes at various animal conditions and discussed.

Introduction

Tinnitus is a medical condition with the perception of a sound (often ringing) that does not come from an external stimulus that others could hear [1,2]. While the sound is commonly described as “ringing,” there have been a plethora of other presentations with descriptors such as roaring, buzzing, humming, and squealing. Continuing the substantial heterogeneity of tinnitus symptoms, the noise may be loud or soft, low or high pitched, and it may be heard in one ear, both, or centralized in one's head [3]. It may even be pulsatile, meaning it may become synchronous or asynchronous with one's own heartbeat [4]. The manifestations of the conditions are vast, and these presentations can have varying impacts on an affected individual's daily life. A large-scale retrospective cohort study found numerous studies that show a range of 15.75% and others reporting a high of 21.6% of tinnitus participants to have suicidal idealizations and 1.2% who reported suicide attempts compared to just 0.6% attempted from those without tinnitus [5].

There is also a broad spectrum of research regarding the causes of tinnitus. As a whole, tinnitus generally develops from pathological changes along the auditory pathway; however, there is not one specific cause. Common risk factors documented include hearing loss and aging [6]. The most common cause stems from initial acoustic lesions of the cochlea as a result of hearing loss, noise trauma, or ototoxic drug use [7]. Other possibilities include but are not limited to, abnormal changes to the auditory nerve, head, or neck injuries [8], emotional distress, or an internal source of the noise, like in the case of heartbeat-synchronous pulsatile tinnitus.

It is additionally difficult to diagnose in some cases because hearing loss does not always lead to tinnitus and those with tinnitus may not have an abnormal audiogram to compound the difficulty of diagnosing tinnitus is the larger challenge of treating it. The starting point is counseling, which is used to explain the symptoms and develop coping skills to deal with the symptoms. This is the common treatment method because there is nothing that cures the ringing noise, so the only option is to develop ways to manage the symptoms. The purpose of the therapy is to avoid negative, self-reinforcing loops of self-talk, which could worsen the tinnitus via hyperfixation [9]. However, as the definite causation and location of the source of tinnitus remain unknown, these harmful habits can be difficult to prevent.

Our study looked at the neural activity in the auditory brain structures, specifically the dorsal cochlear nucleus (DCN), the inferior colliculus (IC), and the auditory cortex (AC). With auditory information starting at the cochlear nucleus, the dorsal aspect has the ability to detect minute frequency differences, providing the brain with auditory information regarding pitch and analyzing the type and quality of the external sound [10]. The axons from the DCN neurons travel both ipsilateral and contralateral, crossing the midline of the brain to form the lateral lemnisci (LL), which then ascend to the IC. Each IC receives auditory input from both ears via a relay pathway, meaning unilateral damage doesn't result in total loss of hearing. Still, there may be an inability to localize a contralateral sound [11]. Deep brain stimulation of the IC has been found to be effective in reducing behavioral evidence of tinnitus in rats [12]. From the IC, the auditory pathway continues to the geniculate nucleus and then arrives at the primary AC located in the temporal lobe, where information such as sound frequency is processed at a higher level.

Acoustic overexposure is known to cause damage to the auditory brain structures, which then leads to increased firing rates or hyperactivity in the inflicted brain areas. Noise-induced hyperactivity related to tinnitus has been observed in research of other species, such as chinchillas, specifically located in the DCN [13]. This hyperactivity can present as phantom tinnitus, and studies have shown that medications, such as lidocaine, can be injected into the auditory cortex to enhance or suppress the tinnitus symptoms. A study by Reyes cited a correlation in that if the lidocaine increased activity in the auditory cortex, the tinnitus symptom (loudness) increased. Those who experienced a decrease in the symptom showed a decrease in the activity of the auditory cortex [14].

As a near-defined mandatory relay structure of the auditory ascending pathway, the IC is a point where almost all lemniscal ascending inputs converge [15]. Therefore, changes and disruptions associated with this pathway could alter sound perception. A study of the use of furosemide, a loop diuretic, in guinea pigs with cochlear trauma showed a significant reduction in spontaneous firing rates and no impact on thresholds in the IC, highlighting potential beneficial pathways to therapeutic treatments of tinnitus [16]. Reduced functional thalamic gaiting has also been observed in human studies of tinnitus with reduced functional connectivity between IC and AC. This dysfunction may be linked to the thalamocortical dysrhythmia model of tinnitus, arguing the causation of tinnitus is related to the disinhibition of AC as a result of the dysfunctional thalamic input [17]. In a review analysis of tinnitus, Berger proposes that the IC may be a prerequisite or contributing factor to the thalamocortical dysrhythmia model explaining tinnitus [18].

In the current study, we characterized the neural activities in the DNC, IC, and AC of rats before and after noise-induced tinnitus, and before and after ACES treatment. The use of electrical brain stimulation (EBS) has shown success as a treatment for movement disorders, like Parkinson's, as well as to reduce seizure frequency [19]. EBS can also be used to assist in the resection of certain brain tumors. This has been done in other clinical settings using EBS as a tool to “map” the brain via direct observations of the functional/cognitive effects of the stimulated area [20]. In a study by Ridder et al., they proposed the use of electrodes to induce stimulation of the auditory cortex in patients with tinnitus in relation to those who suffer from neuropathic pain. Pointing out the similarities in symptoms may suggest a similar pathophysiology, with the results concluding that all the participants who were experiencing them experienced a significant reduction after the stimulations [21].

There is a broad range of study on tinnitus analysis and researches on the underlying mechanism of tinnitus-related hyper neural activities. Yet in our study, we focused on signal analysis and aimed to find any detectable changes in neural signals recorded during different statuses of animals. Particularly, besides the frequency analysis and correlations, which are commonly used in neural computation, we have added the evaluations of signals in the time domain and discussed the time instant and shape pattern of hyper neural activities. A preliminary study on two animals has been conducted, analyzing electrode-collected neural activities and to find how they behaved differently before and after noise exposure and ACES. Meanwhile, the signals were analyzed across the areas of DCN, IC, and AC, which could lead to a future analysis of information transfer in the auditory brain structure for tinnitus diagnosis and treatment. The numerical analysis methods were introduced in section of Materials and Methods, followed by sections of Results and Discussion, and Conclusion.

Materials and Methods

In the experimental test, multichannel electrodes were used to collect the neural activities in the DCN (Channels 1–32), IC (Channels 33–64), and AC (Channels 65–80), each measured local field potentials (LFPs) for 5 min. Two animals have received complete data for analysis. One rat was measured before and after noise exposure. The other rat has noise-induced tinnitus and data were collected before and after ACES. The success of tinnitus induce was evaluated with behavior tests [22].

Spike Activity Extraction.

In the 5-min measurement, the time-domain LFPs were employed in the study in this paper. Neural spikes were identified by finding peaks with amplitude equal to or higher than three times of standard deviation above the mean value and a peak-to-peak interval of 50 samples (0.13 s) to remove any fluctuations around peaks. The firing rate was therefore counted by the number of peaks per minute. Figure 1(a) has shown an example of the raw data, from the animal before noise exposure, the LFPs signals from each auditory brain area, DCN (Channel #1), IC (Channel #33), and AC (Channel #65), respectively. And the triangle marks the identified spike time instants. Figure 1(b) has zoomed in and shown the activities around one of the spikes. It can be seen that the three areas, DCN, IC, and AC, behaved similar with a spike activity, but the amplitude and location of spike peaks vary. And also for this case when fluctuations happened, the highest peak was chosen for the spike location. The time-domain signals are synchronized among channels, and the different time instants of spikes are further used to calculate the time delay among channels.

Fig. 1.

Time-domain signals and spike extraction: (a) 5-min data from DCN, IC, and AC and (b) one spike at DCN, IC, and AC

Phase-Phase Correlation.

Phase–phase correlation algorithm [23] was applied to the raw LFPs data. The time-domain signals were converted to the frequency domain with the Hilbert transform and the correlations in phase on mth and nth channels were calculated in the following equation:

$$R_{m,n}=\frac{1}{K}\sum _{k=1}^K\left|e^{i\mathrm{\Delta }\left({\varphi }_m-{\varphi }_n\right)}\right|$$ (1)

where K segments were randomly selected through the phase array to calculate the angle differences and the averaged number was used to present R_m,n. The result had values between 0 and 1, with 0 meaning no correlation exists between the two arrays and 1 for 100% correlation.

Spike Shape Analysis.

All spikes in each channel were extracted with a rectangular window on the time domain, synchronized, and averaged to investigate the spike shape. The amplitudes of the averaged spike were collected. The widths of the peak were also calculated at half prominence of each peak to further describe the shape of the spikes.

Time Delay Estimation.

Besides the correlations among channels interpreting the information translation in the auditory brain structure, the time delay [24] among channels was also tested, aiming to identify how much time may take to transfer the information between areas, and if the time delay varies on different animal conditions. The amount of time needed to transfer information from the mth to nth channels was calculated in time delay at peaks (TDAP)

$${\mathrm{TDAP}}_{m,n}=\frac{1}{J}\sum _{j=1}^J\left|t_n-t_m\right|$$ (2)

where t_m and t_n were the time instant of the peak location of the spikes, and J indicated the number of total spikes found in the mth channel. The TDAP results are in terms of numbers with unit of second. Positive numbers indicated the peak at nth channel happened after the mth channel, negative vice versa.

Results and Discussion

A preliminary study of two animals was conducted using the numerical methods mentioned in the previous section. In the first case, one animal was used to collect 5-min spontaneous activity in the three auditory brain structures. After the measurement, this animal was then exposed to a tuning curve of 10–18 kHz frequency to introduce tinnitus, followed by another 5-min measurement. The two sets of data were processed and used to compare and investigate the difference before and after noise exposure.

In the second case, another animal was used whose tinnitus was identified with behavior tests in advance. 5-min spontaneous activity was collected, followed by ACES, and then another 5-min measurement on post-ACES condition. This was used to compare the neural activities before and after ACES. The results and discussion are as follows.

Case 1: Comparison of the Animal Before and After Noise Exposure to Induce Tinnitus.

To compare the firing rate at DCN, IC, and AC areas with all available channels data averaged, the animal showed 21.6, 22.4, and 20.2 spikes per minute before noise exposure, and 19.2, 20.0, and 19.0 after noise exposure, respectively. No significant differences were observed for this animal.

The direct measured LFPs were first compared in terms of phase-phase correlation among the 80 channels. Note that two channels (32 and 40) had cable connection problems during the experiment; thus, their data were removed. Color maps were generated to show the correlation in numbers across channels, with 0 meaning no correlation between channels, and 1 meaning 100% correlation. As shown in Fig. 2, strong self-coherence within areas was found in the phase–phase correlation, especially within IC and AC areas. For correlations between different areas, a slightly higher correlation was observed between DCN and IC. No significant change before and after noise exposure.

Fig. 2.

Phase-phase correlation before/after noise exposure: (a) before noise exposure phase–phase correlation and (b) after noise exposure phase–phase correlation

The other analysis conducted was to compare the spike shapes in the three areas before and after the noise exposure. For all available channels, the averaged amplitude values at DCN, IC, and AC were 1.53 × 10⁻⁴, 1.94 × 10⁻⁴, and 4.09 × 10⁻⁴ before noise exposure, and increased to 2.30 × 10⁻⁴, 2.83 × 10⁻⁴, and 4.58 × 10⁻⁴ respectively after noise exposure. For the width of the peak in the three areas, before noise exposure, it was 2.04 × 10⁻², 4.04 × 10⁻², and 2.88 × 10⁻², and after noise exposure, it was 2.33 × 10⁻², 4.56 × 10⁻², and 3.60 × 10⁻², respectively. Figure 3 showed the comparison of one channel selected from each area. The solid lines indicated all spikes averaged in the channel before noise exposure and the dashed lines after noise exposure. It can be seen that amplitude of the spike increased when tinnitus was induced; this was understandable as hyperactivity was induced. Meanwhile, no significant change was observed in viewing the width of the peak.

Fig. 3.

Averaged spike shape in the auditory brain structure before/after noise exposure: (a) averaged spike shape in the DCN before (solid line) and after (dashed line) noise exposure, (b) averaged spike shape in the IC before (solid line) and after (dashed line) noise exposure, and (c) averaged spike shape in the AC before (solid line) and after (dashed line) noise exposure

Lastly, the time sequence of spikes in the three auditory brain structures was evaluated. The average time delay showed the time difference when the spike happened, which could possibly indicate how soon the information could be transferred forward. The results were shown in Fig. 4, where the color bar shows the time delay in seconds. Among the time of signal traveling to different areas, DCN to IC almost happened at the same time, while AC has shown a response time of about 0.01–0.02 s. For before and after noise exposure, not much difference existed. This agreed with our expectation as the measurements were for spontaneous activities.

Fig. 4.

Time relations in the auditory brain structure before/after noise exposure: (a) time relations in the auditory brain structure before noise exposure and (b) time relations in the auditory brain structure after noise exposure

Case 2: Comparison Before and After Auditory Cortex Electrical Stimulation.

In the second case, another animal was used to investigate the effect of ACES. 5-min spontaneous activities were collected before and after ACES. One channel (65) had a hardware connection issue and thus was removed from the data.

For this animal before ACES, the average firing rates at DCN, IC, and AC were 37.4, 32.4, and 60.4 spikes per minute, and the numbers reduced to 30.6, 30.8, and 34.6 after ACES. The reduction of spike rate has shown the effectiveness of ACES as a tinnitus treatment on this animal.

In the phase–phase correlation analysis, as shown in Fig. 5, strong self-coherence was observed within areas, especially within IC and AC areas. For cross-area correlation, there was a relatively stronger coherence between DCN and IC. It is also noticed that the coherence between IC and AC was slightly increased after ACES.

Fig. 5.

Phase–phase correlation before/after ACES: (a) before ACES treatment phase–phase correlation and (b) after ACES treatment phase–phase correlation

For the shapes of the spikes at DCN, IC, and AC areas, the overall averaged amplitude changed from 6.48 × 10⁻⁵, 1.28 × 10⁻⁴, 2.94 × 10⁻⁴ to 7.84 × 10⁻⁵, 1.23 × 10⁻⁴, 3.38 × 10⁻⁴, and width of the peak from 1.56 × 10⁻², 5.41 × 10⁻², 3.46 × 10⁻² to 2.69 × 10⁻², 3.34 × 10⁻², 2.97 × 10⁻² after ACES. Figure 6 showed the average shape pattern of the spikes for one channel from each area, in comparison before and after ACES. The solid lines are before ACES, and the dashed lines are after ACES. No significant change was found. However, note that the firing rate has decreased significantly as discussed above, which indicated the effectiveness of ACES as a tinnitus treatment.

Fig. 6.

Averaged spike shape in the auditory brain structure before/after ACES: (a) averaged spike shape in the DCN before (solid line) and after (dashed line) ACES treatment, (b) averaged spike shape in the IC before (solid line) and after (dashed line) ACES treatment, and (c) averaged spike shape in the AC before (solid line) and after (dashed line) ACES treatment

The time sequence of spikes in the three auditory brain structures was shown in Fig. 7. It can be seen that before ACES, IC had some delay to DCN. In comparison, the delay disappeared after ACES. The other found is that the AC signal doesn't respond in the sequence of the DCN-IC-AC as expected; this could be caused by hyper-activities of AC without external stimulus, say, tinnitus-related activities.

Fig. 7.

Time relations in the auditory brain structure before/after ACES: (a) time relations in the auditory brain structure before ACES treatment and (b) time relations in the auditory brain structure after ACES treatment

Another issue we found is that the post-ACES measure was right after ACES; therefore, it could include the direct measure of ACES, not the outcome of treatment.

In summary, our test results have shown that spikes played an important role and showed detectable changes among different animal conditions: spike amplitude has increased significantly after noise exposure on one animal, and spike rate reduced after ACES on the other animal. The width of the spikes didn't show noticeable changes though. Phase–phase correlation results showed areas in the auditory brain structure had strong self-correlation, and some areas had more information transfer than others, yet little change between different animal conditions. Time relations results represented how soon the neural signals travel across auditory brain structure and it was found for the animal with tinnitus before/after ACES, it didn't respond in the sequence of the DCN-IC-AC as expected. This could be caused by tinnitus-related activities, which are worth further investigation.

Conclusion

The paper proposed using multiple signal analysis algorithms in analyzing the spike changes at DCN, IC, and AC areas before and after noise exposure, and before and after ACES. Signal analysis results in animal data could be explained by tinnitus-related hyper-neural activities induced by noise exposure and the effect of ACES. Some of the signal patterns, such as firing rate, phase-phase correlation, spike amplitude, and time delays, have shown differences in various animal conditions and are thus worth further investigation. The proposed numerical methods were applied to the data of two animals, one for before/after noise exposure-induced tinnitus and the other for before/after ACES cases; therefore, the processed results and the detection of signal changes could be limited. In the future, more animal tests will be conducted to evaluate the changes and therefore help in the diagnosis and treatment of tinnitus with electrode signal measurement.

Funding Data

• National Institute on Deafness and Other Communication Disorders, NIH (Grand No. 1R21 DC010059-01; Funder ID: 10.13039/100000055).

Data Availability Statement

The datasets generated and supporting the findings of this article are obtainable from the corresponding author upon reasonable request.

Nomenclature

R =

phase–phase correlation

TDAP =

time delay at peaks