Fifteen guinea pigs (Cavia porcellus, Hartley Tricolor) of either sex, weighing 200 to 300 g, were obtained from the University of Western Australia’s Preclinical Facility (PCF). Guinea pigs were kept under controlled conditions and were provided with appropriate access to food, water and shelter throughout the duration of the experiment. They were divided into an experimental group (n = 5) which underwent an acoustic trauma procedure and a control group (n = 10) that received a sham procedure. Ethics regarding experimental procedures were approved by the Animal Ethics Committee of the University of Western Australia.

Animals received a subcutaneous (s.c.) injection of atropine sulphate (0.05 mg/kg; 0.1 mL), followed by an intraperitoneal (i.p.) injection of Diazepam (Pamlin 5 mg/kg, 5 mg/mL diazepam). Then, guinea pigs received an intramuscular (i.m.) injection of Hypnorm (1 mL/kg; 0.315 mg/mL fentanyl citrate + 10 mg/mL fluanisone) to induce full surgical anaesthesia. Animals were shaved at the incision site and received a s.c. injection of Lignocaine (0.1 mL 1% solution). Once full depth of anaesthesia was obtained, animals were placed on a heating blanket in a soundproof room and mounted into hollow ear bars that allowed controlled acoustic stimuli to be delivered to the animal. The cochlea was exposed by a skin incision followed by a small opening in the bulla. An insulated silver wire recording electrode was positioned on the round window. Compound action potential (CAP) recordings were made to assess peripheral auditory thresholds between 4 and 24 kHz (2 kHz steps) (Johnstone et al., 1979). CAP recordings (32 averages per recording) were made in response to 10 ms tone bursts with repetition rate of 4/s.

When normal hearing was confirmed (Johnstone et al., 1979), a unilateral hearing loss was induced in the left ear in the experimental group by an acoustic trauma (continuous pure tone at 10 kHz and 124 dB SPL for 120 min). Plasticine was used to block the right ear. A half-inch condenser microphone driven in reverse was used as a speaker (Bruel and Kjaer, type 4134). A DIGI 96 soundcard connected to a digital/analog interface (ADI-9 DS, RME Intelligent Audio Solution, Haimhausen, Germany) and a custom-built software program (Neurosound MI Lloyd) was used to synthesise acoustic stimuli. This method of acoustic trauma is used routinely in our laboratory and causes a small, frequency restricted permanent hearing loss (Mulders and Robertson, 2009; Mulders et al., 2011). Sham animals received the same surgery and measurement of peripheral thresholds by CAP, but with no acoustic trauma. Animals received a top up of Hypnorm (one third of original dose) halfway the acoustic trauma. After the acoustic trauma, another CAP audiogram was recorded to determine the extent of immediate hearing loss, the incision was sutured, and the animals recovered for 2 weeks before the final experiment.

Two weeks after the recovery experiment, guinea pigs were administered a s.c. injection of atropine sulphate (0.05 mg/kg; 0.1 mL), followed by an i.p. injection of sodium pentobarbitone (30 mg/kg). Full surgical anaesthesia was then achieved by an i.m. injection of Hypnorm (initial dose 0.15 mL). Lignocaine (0.1 mL 1% solution) was injected s.c. at the incision site. Once full depth of anaesthesia was obtained, guinea pigs were placed on a heating blanket in the soundproof room. An electrocardiogram (ECG) was used to monitor the depth of anaesthesia throughout the experiment. Anaesthesia was maintained by half of the initial dose of sodium pentobarbitone every 2 h and a full dose of Hypnorm every hour. Guinea pigs were artificially ventilated by carbogen (95% oxygen and 5% carbon dioxide) through a tracheotomy. Animals were then positioned into hollow ear bars and as described in the recovery experiment (see section “Recovery Experiment for Acoustic Trauma or Sham Procedure”), a CAP audiogram was recorded from both the left and right ear to determine peripheral auditory thresholds.

Then the skull was exposed and using coordinates from a guinea pig atlas (Rapisarda and Bacchelli, 1977), position of the PFC and MGN were determined. A dental drill was used to perform two small craniotomies. A custom made (tungsten in glass) bipolar stimulating electrode was positioned in the PFC and a recording electrode [impedance < 1 MΩ at 1 kHz (Merrill and Ainsworth, 1972)] was lowered into the MGN. To minimise dehydration and limit movement of the brain, 5% agar in saline was used to cover the craniotomies. Ten minutes prior to recording, guinea pigs were administered an i.m. injection of Pancuronium Bromide (0.1 mL).

The stimulating electrode (Platinum-iridium, concentric bipolar electrode with a 2–3 μm tip, impedance 200 K, World Precision Instruments) was connected to an A-M Systems Isolated Pulse Stimulator (Model 2100). Neurosound software controlled the timing of electrical stimuli. Electrical stimulation was delivered as shock trains (50 ms duration) with a pulse duration of 0.5 ms and at a rate of 200 Hz. The intensity of the electrical current ranged from 100 μA to 1 mA. To increase the chance of observing an effect in the MGN, the current used to stimulate the PFC was initially applied at maximum (1 mA). The current was then decreased until threshold if the recording remained stable. Threshold was determined as the lowest possible current intensity that could elicit a response that was different to the neuron’s firing pattern.

Single neurons in the MGN were identified using pure tone and broadband noise stimuli as a search stimulus. Noise stimuli were delivered to the left ear while the right ear was blocked with plasticine. A tungsten in a glass recording electrode was used to record single neuron activity in the MGN of the thalamus (Merrill and Ainsworth, 1972). Entry into the right (contralateral to noise stimulus) MGN was verified by the presence of a robust cluster response to broadband noise stimulus (50 ms duration, 1 ms rise and fall times, sample rate of 96 kHz) (Cook et al., 2021; Zimdahl et al., 2021). Electrophysiological recordings of single neuron activity were recorded in MGN in response to stimulation of the right (ipsilateral to MGN) PFC.

When a single neuron was isolated, characteristic frequency (CF) and acoustic threshold at CF were determined as described previously (Mulders et al., 2011; Barry et al., 2015, 2019). Spontaneous firing rates (SFRs) were measured for 10 sec. Peristimulus time histograms (75–100 stimulus presentations) were constructed at CF at 20 dB above threshold to determine neuronal response. Then histograms (500 ms samples; 50–100 sweeps) were collected to determine the effect of PFC stimulation as follows. Firstly, effects of PFC stimulation were investigated on spontaneous firing, followed by effects on sound evoked activity (at 20 dB above neuronal threshold). If an effect was observed, current was decreased in small steps (200–300 μA) until threshold was reached. For analysis, changes in firing rate were determined by comparing the total spike counts with and without electrical stimulation of PFC.

Once recordings were finished, animals received an overdose of Lethabarb (Sodium Pentobarbitone 325 mg/mL; 0.3 mL i.p injection). Guinea pigs were transcardially perfused by methods outlined by previous studies (Barry et al., 2015, 2017). Brain tissue was collected and stored overnight in 4% paraformaldehyde in 0.1 M phosphate buffer (PB), followed by cryoprotection in 30% sucrose in PB for another 24 h. Brains were cut on a freezing microtome-cryostat into 60 μm sections. Sections were mounted onto slides and allowed to dry overnight. Once dried, slides were stained with Toluidine Blue and coverslipped. A Nikon Eclipse 8oi was used to perform light microscopy. This procedure verified the position of stimulating and recording electrode tracks within the PFC and MGN.

Multiple t-tests, corrected using Holm-Sidak method, were used to determine if sham or acoustic trauma had an effect on peripheral auditory thresholds and if hearing thresholds were different in the experimental and control group. A Mann-Whitney U Test was used to determine differences in spontaneous firing rates, characteristic frequencies (CF) and neural (acoustic) thresholds at CF between animals with normal hearing and hearing loss. To characterise the effect of PFC stimulation on single neuron MGN activity PSTHs without and with stimulation were compared using a paired t-test, with each time-bin as a separate variable. Proportions of neurons displaying no effect, inhibition or excitation were compared using a Chi-square analysis.

The effects of acoustic trauma (AT) are shown in Figure 1. CAP thresholds in sham animals at the start of experiment were the same as in the AT group before the AT (Figure 1A). Immediately after AT was finished a significant increase in CAP thresholds was observed at all frequencies with the largest increase seen at and above the AT frequency (Figure 1A; multiple t-tests, corrected using Holm-Sidak method) in line with our previous publications (Mulders and Robertson, 2013; Mulders et al., 2016; Cook et al., 2021). After 2 weeks of recovery, at the time of the final experiment, CAP thresholds had partially recovered but a difference remained between sham and AT animals (Figure 1B). CAP thresholds showed a significance increase only at 20 and 22 kHz (p < 0.05), with threshold at 12 kHz approaching significance (p = 0.063, multiple t-tests, corrected using Holm-Sidak method). Right ears (not exposed to AT or sham surgery) showed similar thresholds at the final experiment (Figure 1C) as measured during initial surgery in the left ears of either group, confirming the absence of an effect of the AT on the contralateral side. In Figures 1B,C the original left ear thresholds of shams are plotted as well to illustrate the stable thresholds in the sham group and the normal thresholds in the right ear at the final surgery.

Histology was used to verify the position of stimulating electrodes in all animals (Figure 2 showing an example). Five animals from the AT group and seven animals from the sham group showed placement of the stimulation electrode in the PFC (Rapisarda and Bacchelli, 1977). Only these animals showed effects on MGN recordings and were therefore included in further analyses. Electrode placement was explored in some animals by changing the depth of the electrode. Stimulation at a depth of approximately 5 mm from the cortical surface, most likely reflecting placement in prelimbic cortex (Hennessy et al., 2018), yielded the largest and most consistent effects on MGN activity.

Data were obtained from a total of 136 neurons with 68 neurons from control animals and 68 from AT animals. Characteristic frequency (CF) varied between 0.24 and 24.5 kHz (11.3 ± 0.8 kHz; mean ± SEM) in sham animals and between 0.6 and 19 kHz (8.1 ± 0.7 kHz; mean ± SEM) in AT animals. Mean CF was significantly different between the groups (unpaired t-test p = 0.017). It should be noted that CF could not be determined in seven neurons of the AT animals and four neurons of the sham animals though these neurons did show a response to noise. Mean SFRs were 0.66 ± 0.13 in AT animals (median 0.1) and 0.63 ± 0.15 in sham animals (median 0.1) and a Mann-Whitney showed no statistical difference (p = 0.899) between the groups.

Response type was based on a PSTH at CF 20 dB above threshold and could be determined in 57 of the neurons in AT animals and 61 of the neurons in sham animals. Examples of the PSTHs of the different response types are shown in Figure 3. Response types were in agreement with data described previous studies (Bordi and LeDoux, 1994; Barry et al., 2015, 2017). In both the sham and AT group, the majority of MGN neurons had an onset response to pure tone (46 neurons, 75.4% and 42 neurons, 73.7%, respectively), with a response latency ranging from 10 to 65 ms from the onset of the sound stimulus (Figure 3A). About a fifth of the neurons (12 neurons, 19.7% in sham group and 13 neurons, 22.8% in AT group) exhibited an on and off response (Figure 3C). An off response (firing related to offset of the tone burst) was observed in 1.6% (one neuron) of sham animals and 3.5% (two neurons) of AT animals (Figure 3B). The remainder of neurons exhibited a “sustained response” to pure tone (two neurons, 3.3% in sham animals, none in AT group) (Figure 3D). No difference was found in the proportion of response types between the sham and AT group [X2 (3, N = 118) = 2.422 p = 0.4895].

Since inhibitory effects cannot be ascertained in neurons with a zero or very low SFR when using extracellular recordings, potential effects on SFRs were only investigated in neurons with a SFR > 0.4 spikes/sec. This meant that effects of PFC stimulation on SFRs were assessed in 26 neurons in AT animals and 28 neurons in sham animals. Mean CF was 10.35 ± 1.36 kHz (mean ± SEM) in sham animals and 7.0 ± 1.17 kHz in AT animals and this difference was not statistically different (unpaired t-test p = 0.0711). Note that in four of the sham animals and five of the AT animals CF could not be determined. In AT animals, excitatory effects were seen in three neurons (12%), inhibitory effects in eight neurons (31%) and no effect in 15 neurons (58%) (Note percentages rounded, so not adding up to 100%). In sham animals, excitatory effects were seen in four neurons (14%), inhibitory effects in 9 (32%) and the remaining 15 neurons showed no effect (54%). These proportions in type of effect were not different between the AT and sham animals [X2 (2, N = 54) = 0.1278 p = 0.9381].

The effects of PFC stimulation were also assessed on sound evoked activity (at CF, 20 dB above threshold) in 57 neurons from AT animals and 59 neurons from sham animals. Mean CF was 10.82 ± 0.83 kHz (mean ± SEM) in sham animals and 8.43 ± 0.69 kHz in AT animals and this difference was statistically different (unpaired t-test p = 0.0291). In one neuron of both sham and AT animals CF could not be determined. In AT animals, excitatory effects were seen in two neurons (4%), inhibitory effects in 21 neurons (37%) and no effect in 34 neurons (60%). In sham animals, excitatory effects were observed in one neuron only (2%), 17 neurons showed inhibitory effects (29%) and the remaining 41 neurons showed no effect (69%). Chi-square analysis showed again no significant difference between the prevalence of effect types between the AT and sham animals [X2 (2, N = 116) = 1.374 p = 0.5032].

Effects of PFC stimulation could be collected on both spontaneous and sound evoked activity from nine neurons in sham animals and 16 neurons in AT animals. From the nine neurons from sham animals, seven showed no effect on both spontaneous and sound evoked responses, one neuron showed excitatory effects on both and the remaining neuron inhibitory effects on both. From the 16 neurons from AT animals, five showed no effect on both responses and 4 neurons showed inhibitory effects on both. Four neurons showed an inhibitory effect on the sound evoked response without an effect on spontaneous firing, two neurons showed no effect on sound evoked responses but excitatory effects on spontaneous firing and the remaining neuron showed no effect on sound evoked responses but an inhibitory effect on spontaneous firing.

Although most neurons did not show an effect in response to PFC activation, it should be noted that they were located in the same tracks as neurons that displayed either inhibition or excitation. In addition, each animal included in the analysis showed some inhibitory or excitatory effects. It is therefore unlikely that the lack of effect was due to incorrect placement of the stimulating electrode. Threshold of stimulation, the lowest possible current intensity that could elicit a response, was collected in 29 neurons from AT animals and 18 neurons from sham animals. Mean threshold was 740 ± 46 μA in AT animals and 711 ± 59 μA and these values were not significantly different (unpaired t-test, p = 0.7221).

The type of effect from PFC stimulation, excitatory, inhibitory or no effect, did not correlate with CF or thresholds. This is shown in Figure 4. Data from SFR measurements are shown in panels 4A (sham) and B (AT) whereas effects on sound evoked firing rates are shown in panels 4C (sham) and D (AT). Please note that in this figure neurons that could not be tuned to a pure tone are shown arbitrarily with a CF of zero and an arbitrary threshold. Examples of inhibitory effects on sound evoked and spontaneous firing are shown in Figure 5. This figure also illustrates the effects of decreasing stimulation intensity. Examples of excitatory effects at different stimulation intensities are shown in Figure 6.

The magnitude of the effect from PFC stimulation was calculated as the percentage change from the PSTH of the unstimulated condition (spikes from 51 to 500 ms in PSTH to exclude contamination of stimulus artefact) and results are shown in Figure 7. The average inhibition in sham animals was 62 ± 5.1% (ranging from 41 to 83%) for SFRs and 48 ± 4.5% (ranging from 18 to 78%) for sound evoked responses. In AT animals the average inhibition for SFRs was 65 ± 9.2% (ranging from 35 to 100%) and for sound evoked responses 60 ± 5.1% (ranging from 24 to 100%) (Figure 7A). Excitatory effects on SFRs in sham animals ranged from 209 to 3740% (mean ± SEM, 1,348 ± 810). In the only neuron that showed an increase in sound evoked responses, the increase was 138%. In AT animals, excitatory effects on SFRs were seen in three neurons, ranging from 37 to 487% (193 ± 147%) and in two neurons in sound evoked response at 57 and 58% (Figure 7B). The magnitude of inhibition was significantly different between sham and AT animals with regards to the sound evoked response (unpaired t-test p = 0.045), but not for the SFRs. Due to the low numbers of neurons showing excitatory effects we did not run an analysis for statistical differences between groups.

To investigate the temporal pattern of effects of PFC stimulation, the average absolute magnitude of effect on SFRs and evoked FRs over time was calculated from the PSTHs (per bin) before and after PFC electrical stimulation for both sham and AT animals. Results are shown in Figure 8. The effects on SFRs are shown for sham (Figure 8A) and for AT animals (Figure 8B). In agreement with the findings above no apparent differences could be observed. The average effects on SFRs are dominated by the excitatory effects even though only few neurons showed an excitatory effect. This can be explained due to the general low SFRs, which means that even 100% inhibition is equivalent to only a small absolute change. The excitatory effects occurred within the first 50 ms after electrical stimulation had seized (grey column in graphs). The effects on sound evoked FRs are dominated by a large inhibitory effect in the first 50 ms after stimulation had stopped (Figure 8C, shams and 8D, AT animals). The significant increased inhibitory effect on sound evoked FRs in AT animals compared to shams is clearly apparent when comparing Figures 8C,D.