We used epidural electrocorticographic (µECoG) and intracortical silicon probe recordings to examine the developmental profile and spatial-temporal characteristics of responses evoked by BC stimuli in the ACx of rats aged from P6 to P30 (total n = 22 rats), and compared them with the development of responses evoked by high-intensity acoustic shock waves (ASW), low-intensity AC auditory stimuli as well as with responses evoked by electrical stimulation of the ipsilateral cochlea (iCo) and the inferior colliculus (IC). The two latter types of stimulation were particularly instructive for ensuring recordings from the auditory cortical areas during the pre-hearing period when AC-evoked responses are not yet present.
BC stimuli evoke a biphasic response in the ACx of neonatal rats
Responses to BC stimuli were recorded from the ACx of head-restrained neonatal rats under urethane anesthesia using 60-channel µECoG arrays placed epidurally over the ACx (average centre: 4 mm caudally, 7 mm laterally from Bregma) (Fig. 1A,B,C). Example response to BC mechanical stimuli applied to the skull through the animal head fixation system (Supplementary Figure S1) in a P8 rat is shown on Fig. 1D. Biphasic BC-evoked responses were characterized by early positive P and delayed negative N components (Fig. 1F). The P component typically was of smaller amplitude and had a shorter duration than the N component. Modest spread over the cortical surface was observed for the N component (Video 1). The P and N components were largely co-localized (Fig. 1G,H), although the N peak was slightly shifted in the caudolateral direction from the P peak (see also Fig. 2C). Besides that, the P component occupied a smaller cortical area compared to the N component (see also Fig. 2E,2F). Concomitant ECoG and intracortical recordings revealed that the P peak corresponds to the maximal sensory-evoked potential (SEP) amplitude and the earliest sink in L4 of the ACx (Fig. 1I) as well as the maximal spiking frequency (Fig. 1J) in the thalamo-recipient L4 and L5/6 border, whereas the N peak corresponds to the maximal SEP amplitude in L2/3 (Fig. 1I). Therefore, we concluded that the P component represents activation of thalamo-cortical synapses, whereas the N component likely reflects transfer of excitation from L4 to L2/3 and further horizontal spread of excitation through the supragranular layers. It should be noted that AC sounds of low intensity (500 ms, 70 dB, 1–40 kHz) failed to evoke any detectable response (Fig. 1E) at this age.
Developmental changes in BC-evoked responses in the auditory cortex.
We further addressed developmental changes in the temporal and spatial organization of BC-evoked responses through the postnatal period. At P6-7, BC stimuli failed to evoke responses in the ACx (Fig. 2A). Starting from P8, BC stimuli reliably evoked biphasic P-N responses in the ACx in all animals, and the delay from stimulus decreased with age (Fig. 2A,B). Latency of both P (r = -0.71, p = 0.0002) and N (r = -0.74, p = 0.00008) peaks from the stimulus, as well as the delay between P and N peaks (r = -0.71, p = 0.0002) decreased with age to attain near-adult values by P15-16 (Fig. 2A,B, n = 22 rats). Also, the developmental trajectory of the P and N peaks from P8 to P16 projected to the P1 and N1 peaks of BC-evoked responses at P30 (Fig. 2A, bottom trace; see also Supplementary Figure S2). This suggests that the P and N components of BC-evoked responses (as well as those of ASW and AC–evoked responses, see below) in neonatal animals are precursors of the P1 and N1 components in adults. Besides the changes in temporal characteristics BC-evoked responses demonstrated an increase in the amplitude of the P (r = 0.55; p = 0.01) and N (r = 0.51, p = 0.0189) peaks (n = 21 P8-16 rats) with age. It is of note that the late P2 and N2 response components were not present in P8-P16 rats.
Developmental changes in the spatial organization of the P and N components were further explored through two approaches. First, we examined the relative positions of the P and N peaks over the cortical surface. P peaks were centred for all animals and the distribution of N peaks around this central P-point was analyzed (Fig. 2C). The distance between P and N peaks was on average 328 ± 40 µm and did not show any change with age (r = 0.105, p = 0.65; Fig. 2D, left). Following the conversion of cartesian peak coordinates to polar coordinates, the Rayleigh test revealed that N peaks were non-uniformly distributed (p = 0.6 10− 7) and located in the caudolateral direction relative to P peaks with a mean angle of -33 ± 6° (where 0° corresponds to the caudal direction; Fig. 2D, right). No age-dependent differences were observed for the P-N angles (r = 0.28, p = 0.21). Second, we also analysed the spatial distribution of P and N half-width areas, which are the cortical regions where the response amplitude was higher (lower in case of the N peak) than half of the peak amplitude. We found that the half-width area of the BC-evoked P component was largely nested within the larger N half-width area (Fig. 2C). Intersection of the P and N half-width areas occupied 74 ± 4 % of the P half-width area (n = 17 rats). The larger N component spread out of the intersection area in the caudolateral direction with the intersection of the P and N half-width areas occupying 48 ± 3 % of the N half-width area (n = 17 rats). Together, these results suggest that BC-evoked signals enter the ACx via thalamocortical connections and spread vertically through L4 – L2/3 connections within the thalamorecipient regions expressing both P and N components and horizontally via supragranular layers to adjacent P-lacking N regions. We also found that the P half-width area did not change in size with age (0.72 ± 0.07 mm2; r = 0.21, p = 0.41), and neither did the N half-width area (1.07 ± 0.08 mm2; r = 0.05, p = 0.85, n = 17 rats). Besides that, the portion that the intersection of the P and N half-width areas constitutes from the P half-width area decreased with age (r = -0.65, p = 0.004, n = 17 rats; Fig. 2E), however there was no such dependence for the N half-width area (r = -0.1, p = 0.69, n = 17 rats; Fig. 2F).
BC-evoked responses co-localize with the ASW and AC-evoked responses.
We further compared the spatial-temporal characteristics of the responses evoked by BC and auditory stimuli (Fig. 3, see also Supplementary Figure S3). While AC stimuli of low intensity (70 dB) failed to evoke responses at ages earlier than P13, AC stimuli of high intensity (~ 110 dB) presented in a form of ASW-evoked cortical responses with electrographic appearance and topography very similar to that of BC-evoked responses starting from P8 (Fig. 3B) (n = 6 P8-12 animals). The mean distance between the P peaks of BC and the ASW-evoked responses was 184 ± 81 µm (N peaks: 297 ± 91 µm; Fig. 3D,H, left) without any preferable direction in the position of BC and ASW P peaks (p = 0.12; N peaks: p = 0.69). The mean absolute area for the P peak of ASW responses was 0.44 ± 0.06 mm2 (N peak: 1.13 ± 0.1 mm2). Intersection of the BC and ASW-evoked P half-width areas over the BC P half-width area was 62 ± 8 % (N: 74 ± 10 %; Fig. 3D,H, middle). The P peak of ASW-evoked responses was slightly delayed from the P peak of BC-evoked responses by 6.7 ± 2.3 ms (N peaks: 5.5 ± 4.4 ms ; Fig. 3C,H, right). Together, these findings point to similarities in the developmental emergence and spatial-temporal organization of BC and ASW–evoked responses starting from P8.
Next, we performed stimulation by AC sounds at low intensity (70 dB SPL) in the frequency range 1–40 kHz (Fig. 3A). In keeping with the results from previous studies12, AC sound of low intensity reliably evoked cortical responses starting from P13 (n = 4 P13-16 animals). Another characteristic feature of cortical AC-evoked responses12,13 that we observed was tonotopic organization with more prominent tonotopic maps in P16 animals than in P13-14. We found that starting from P14 BC-evoked responses are co-localized with responses to a high frequency AC sound (20–30 kHz). The mean distance between the P peaks of BC and the closest AC-evoked responses was 193 ± 30 µm (N peaks: 266 ± 31 µm; Fig. 3E,H, left) with no preferable relative position of the AC and BC P peaks (p = 0.96, N peaks: p = 0.89). The mean absolute area for the P peak of AC responses was 0.75 ± 0.14 mm2 (N peak: 1.37 ± 0.22 mm2). Intersection of the BC and AC-evoked P half-width areas over the BC P half-width area had a mean value of 60 ± 5 %, (N: 82 ± 7 %; Fig. 3G,H, middle). The P peak of AC-evoked responses was delayed from the P peak of BC-evoked responses by 0.667 ± 0.72 ms (N peaks: 1.0 ± 0.68 ms; Fig. 3F,H, right). Together, the similar spatial and temporal characteristics of BC and AC-evoked responses suggest that BC stimuli evoke responses in the auditory cortical areas.
No responses were observed to the stimulation at 40 kHz, but stimulation at other frequencies (30, 20, 10, 1 kHz) reliably evoked cortical responses in P13-P16 animals. In all examined animals (n = 4 rats) BC-evoked responses were co-localized with AC-evoked responses. However, in P16 animals (n = 2) BC-evoked responses were closest to the responses to the highest AC sound frequency (30 kHz) and furthest from the responses to the lowest AC sound frequency (1 kHz) (Fig. 3I,J,K). The half-width areas of BC-evoked responses also largely overlapped with the half-width areas of the responses to 30kHz AC sound but did not overlap with the half-width areas of the responses to the 1-kHz AC sound (Fig. 3K). Cortical responses to AC sound demonstrated frequency-dependent localization as illustrated by the example responses from a P16 animal on Fig. 3. These responses were organized in two opposite frequency gradients (low-to-high) with one running in the caudo-rostral direction and the other in the latero-medial direction (Fig. 3K, black arrows). Both gradients were considerably reproduced among animals (caudo-rostral in n = 3 rats; latero-medial in n = 2 rats)
BC-evoked responses co-localize with the iCo and IC-evoked responses.
We further performed electrical stimulation of the iCo (n = 4 rats) and the ipsilateral IC (n = 5 rats) using bipolar electrodes (50–100 µs, 60–70 V, 0.1 Hz; Fig. 4, see also Supplementary Figure S4). The mean distance between the P peaks of BC and the iCo-evoked responses was 459 ± 194 µm (N peaks: 488 ± 260 µm; Fig. 4B,H, left) with no preferable relative position of the iCO and BC P peaks (p = 0.456; N: p = 0.3). The mean absolute area of the P peak of iCo-evoked responses was 0.58 ± 0.06 mm2 (N peak: 0.85 ± 0.15 mm2). The intersection of the BC and iCo-evoked P half-width areas over the BC P half-width area was 46 ± 18 % (N: 51 ± 18 %; Fig. 4D,H, middle). The P peak of iCo-evoked responses was delayed from the P peak of BC-evoked responses by 2.0 ± 2.48 ms (N peaks: 2.75 ± 4.37 ms; Fig. 4C,H, right). The similar spatial and temporal characteristics of BC and iCo-evoked responses suggest that BC stimuli also evoke responses in the auditory cortical areas during the pre-hearing period.
Stimulation of the IC reliably evoked cortical responses starting from P6 (the earliest examined age). In a P8 animal the BC and IC-evoked responses were co-localized (Fig. 4E,F,G). However, in P10-13 animals the localization of the IC-evoked responses depended on the depth of stimulation (Fig. 4I,J,K,L). The mean distance between the P peaks of BC and the closest IC-evoked responses was 1263 ± 306 µm (N peaks: 947 ± 231 µm; Fig. 4G,H, left) with IC-evoked responses located in the caudal direction relative to the P peaks with a mean angle of -8 ± 12° (p = 0.007; N peaks: -8 ± 19°, p = 0.03; Fig. 4G). The mean absolute area for the P peak of IC-evoked responses was 1.01 ± 0.28 mm2 (N peak: 1.16 ± 0.25 mm2). The intersection of the BC and IC-evoked P half-width areas over the BC P half-width area was 14 ± 10 % (N: 20 ± 7 %; Fig. 4G,H, middle). The P peak of IC-evoked responses was earlier than the P peak of BC-evoked responses by 7.8 ± 2.06 ms (N peaks: 4.25 ± 2.94 ms; Fig. 4F,H, right). Another interesting observation was that the stimulation of the IC reliably evoked short latency antidromic spikes in the presumptive L5b of the ACx (Supplementary Figure S5) which resulted in the emergence of the P’ ECoG component preceding the P component (Fig. 4F). Since these antidromic spikes in adult animals signify the existence of cortico-collicular projections29, our findings suggest that this type of connections is present prior to the onset of low-threshold hearing.
In addition to the tonotopic maps obtained for responses to AC-sound, stimulation of the IC also evoked cortical responses in a topographic manner depending on the depth of stimulation (n = 3 rats). The BC-evoked responses were localized closer to the responses to stimulation of the deepest part of the IC and further from the responses to stimulation of the superficial part (Fig. 4K). Since in adult rats the central nucleus of IC (CIC) was reported to be arranged in a laminar pattern that represents different frequency bands, with higher frequencies represented in deeper parts of the CIC1 (Fig. 4L) we conclude that our findings reflect tonotopic organization of the collicular-cortical pathway and suggest that BC-evoked responses are localized in high-frequency processing ACx regions.
Bilateral cochlear ablation eliminates BC-evoked responses.
We further investigated whether the BC-evoked responses involve cochlear activation (Fig. 5) in two animals (P8 and P16). We found that after bilateral cochlear ablation (Fig. 5A; see also Supplementary Figure S6) BC stimuli failed to evoke any detectable cortical responses in the cortical areas where the responses to BC-sound were observed in control recordings (Fig. 5B,C). Amplitude of the P peak dropped to 0.033% and 0.02% of the control values in a P8 and a P16 animal (N peak: 0.06% and 0.014%) respectively. These results suggest that BC-evoked responses involve cochlear activation and propagate through the auditory system up to the neocortex.