This study exploited multiple viral strategies to target spinoparabrachial projection neurons, defining a distinct population in the lateral region of LV. First, the propensity of PBN retro-AAV2 serotype injection to selectively transduce lateral LV SPBNs, but not the well described larger population of lamina I SPBNs was characterised. This unexpected result was directly compared with the labelling achieved by AAV9-RFP12. This identified limited overlap between retro-AAV2 and AAV9 labelled SPBNs, with those cells that did exhibit co-expression concentrated in the lateral LV region (LVlat). The intersection of retrograde PBN rAAV2-Cre and spinal AAV-Brainbow expression allowed morphological characterisation of LVlat SPBNs in the lumbar enlargement, with most exhibiting multipolar morphology, dendritic territories biased to the lateral and ventral spinal cord in the transverse plane, and extending over substantial rostrocaudal distances. Targeted recordings from ChR2-positive LVlat SPBNs in spinal cord slices confirmed optogenetic control, with reliable spiking achieved during photostimulation. In vivo spinal photostimulation under anaesthesia showed that the ChR2-expressing SPBNs and axon collaterals increased the expression pERK, a neuronal activity marker, in a range of local spinal circuits. Finally, recordings from randomly selected DH neurons in acute spinal slices showed that LVlat SPBNs made functional connections via local axon collaterals in 17% of recordings.
Retro-AAV2 serotype produces brainstem labelling consistent with retrograde transduction
To assess how reliably the AAV2-retro serotype transduced retrograde specific expression, the number of neurons labelled by the AAV2-retro versus AAV9 was compared at the injection site (lateral PBN - lPBN, Fig. 1A-B). Consistent with a retrograde specific action for AAV2-retro but not AAV9, this comparison confirmed that few lPBN neurons were labelled with AAV2-ChR2 or AAV2-GFP constructs (Fig. 1C-D: n = 5 animals, 7.2 ± 4 neurons; n = 2 animals, 52.5 ± 18 neurons, respectively), whereas many PBN neurons were labelled with the AAV9 construct (n = 7 animals, 419.6 ± 82 neurons, P = 0.000, N = 14). Also consistent with enhanced retrograde tropism, AAV2-retro produced far less fluorescence in terminals within lPBN, measured by mean pixel intensity (mpi), when compared to that in the superior cerebellar peduncle (SCP) (lPBN: 55.2 ± 20.4 vs SCP: 82.7 ± 26 mpi, difference: -27.5 ± 9 mpi, respectively). The opposite was true for the AAV9, with most injection sites showing a similar or less terminal fluorescence in SCP compared to lPBN (n = 7; lPBN: 75.5 ± 28 vs SCP: 68.5 ± 26 mpi, difference: 6.9 ± 11 mpi). The difference between lPBN and SCP fluorescence reinforced this distinction in viral labelling patterns (Average lPBN vs. SCP difference: AAV9 = 6.9 ± 11mpi vs rAAV2 = -27.5 ± 9mpi; z = 6.958, P = 0.00, N = 14). Importantly, the Rostroventolateral Medulla was not labelled in retro-AAV2, a well characterised source of descending input to the spinal cord; whereas AAV2-retro mediated cellular labelling was common in the contralateral Locus Coeruleus, ipsilateral vestibular nuclei, and paraolivary nuclei, all sources of PBN input and consistent with a retrograde action for this serotype.
Distinct spinal distribution of rAAV2-labelled SPBNs
Within the spinal cord, the rAAV2-ChR2/-GFP constructs labelled surprisingly few SPBNs in lumbosacral sections (n = 117, in 5 animals, Fig. 2A). Given this low overall number but similar overlapping distributions and densities of rAAV2-ChR2/-GFP labelled SPBNs across animals, the sample was pooled for analysis. Despite unilateral injections, rAAV2-labelled SPBN numbers were comparable between the ipsilateral (n = 65) and contralateral (n = 52) spinal cords, with the highest density observed in lamina V of the contralateral spinal cord (n = 30, 0.52 ± 0.50 neurons/50um), and the majority of these located in the lateral aspect of the DH, concentrated at the lateral grey/white matter border. The remaining contralateral SPBNs were infrequently distributed in other locations (LVI/VII: n = 16, 0.16 ± 0.21 neurons/50um; LSN: n = 5, 0.071 ± 0.08 neurons/50um, LVI, LX: n = 1, 0.01± 0.02 neurons/50um). rAAV2-labelled SPBN distributions in the ipsilateral cord were very similar to the contralateral side with the main density in lamina V (n = 32, 0.52 ± 0.62 neurons/50um), followed by LVI/LVII (n = 12, 0.13 ± 0.16 neurons/50um), and infrequent neurons within LSN, and LIII/LIV, (LSN: n = 8, 0.09 ± 0.06 neurons/50um; LIII-LIV: n = 11, 0.11 ± 0.13 neurons/50um). Finally, a single lamina I SPBN was identified in the ipsilateral spinal cord of a rAAV2-GFP animal and was the only instance across all sections analysed (n = 1, 0.01 ± 0.02 neuron/50um, from 91 sections).
The above rAAV2 SPBN labelling was directly compared with AAV9-RFP labelling as both viruses were included in PBN injections. Significantly more SPBNs were transduced by AAV9-RFP, with 1306 RFP profiles identified from the lumbosacral enlargement across 7 animals (Fig. 2B), greater than ten times those labelled by rAAV2-ChR2/-GFP. The distribution of AAV-RFP SPBNs was similar to our previous reports12, but starkly contrasted the rAAV2-ChR2/-GFP distribution. Most AAV9-RFP labelled cells were located in the contralateral spinal cord (n = 872, 66.8%) and concentrated in lamina I (n = 643, 6.8 ± 3.13 neurons/50um), followed by populations in the Lateral Spinal Nucleus (n = 130, 1.7 ± 1.05 neurons/50um), lamina V (n = 60, 0.87 ± 0.89 neurons/50um), and some across LIII, LIV, LVI, LVII and LX (n = 39; LIII: 0.17 ± 0.2 neurons/50um, LIV: 0.15 ± 0.2 neurons/50um, LVII: 0.08 ± 0.1 neurons/50um, LVI: 0.05 ± 0.1 neurons/50um: and LX: 0.04 ± 0.05 neurons/50um). A total of 289 SPBNs were identified in the ipsilateral spinal cord, with the main densities in lamina I (n = 84, 1.2 ± 1.0 neurons/50um), LSN (n = 102, 1.2 ± 0.5 neurons/50um), lamina V (n = 72, 0.99 ± 1.0 neurons/50um), and lamina IV (n = 23, 0.2 ± 0.3 neurons/50um), with negligible labelling elsewhere. Overlap between rAAV2 and AAV9 labelling was only assessed in LVLat as this was the only region to exhibit robust rAAV2-labelling (Fig. 2C-G). On the ipsilateral side, 38% (n = 22/61) of rAAV2-GFP SPBNs coexpressed AAV9-RFP, and conversely 24.7% (n = 22/89) of AAV9-RFP profiles exhibited coexpression of rAAV2-GFP. Overlapping expression was similar in the contralateral spinal cord, where 35.3% (n = 12/44) of rAAV2-GFP labelled SPBNs coexpressed AAV9-RFP, and 20% (n = 12/60) of AAV9-RFP labelled SPBNs coexpressing rAAV2-GFP. Thus, PBN injection of AAV9-RFP and rAAV2-GFP/ChR2 virus selectively identify distinct overlapping populations we collectively term LVLat SPBNs.
LVlat SPBN incidence and morphology
Across all labelling approaches, the LVlat SPBN population represented a discrete cell column positioned in the deep dorsal horn, concentrated in the lateral reticulated areas where parallel fasciculi and a narrow band of dorsal grey matter exist. The proportion of LVlat SPBNs in this region was assessed by comparing NeuN expression with RFP-labelling from unilateral PBN injections of AAV9-Cb7-Cl-mCherry (AAV-RFP, n = 3, Fig. 3). This analysis showed a relatively low overall density of NeuN labelled profiles, with AAV9-labelled SPBNs constituting approximately 12% of the LVlat cell column (Fig. 3: Animal 1: n = 48/338, 13.6 ± 0.02%, Animal 2: n = 35/273, 11.6 ± 0.04%, Animal 3: n = 74/688, 11.0 ± 0.02%). Using the relative proportion of AAV9/rAAV2 overlap from the above analysis, AAV9-RFP counts were extrapolated to estimate total LVLat SPBN proportions of 23.37% and 21.25% in the LVLat region of each animal.
Dense dendritic RFP labelling of LVlat SPBNs in the above tissue precluded detailed analysis of individual neuron morphology. This was addressed in 4 additional animals that received an rAAV2-Cre injection in the right PBN and an intraspinal injection of AAV9-Flex-Brainbow to the left spinal dorsal horn to produce intersectional multi-colour Brainbow labelling of LVlat SPBNs for reconstruction of somatodendritic dendritic morphology (Fig. 4). Cre-dependent labelling of LVlat SPBNs in this tissue, assessed using rAAV2-CreGFP, broadly reflected the expression achieved with other rAAV2 constructs. Specifically, CreGFP expressing neurons were concentrated in the LV lateral cell column along with small numbers in LI, LVI/LVII and LX (Supp Fig. 4). A total of 94 brainbow labelled LVlat SPBNs were identified contralateral to PBN injection. Interestingly, in 2 animals, a small population of SPBNs were located ipsilateral to PBN injection, also positioned in LVlat (n = 8). Overall, the number of LVlat SPBNs labelled per animal varied (Animal 1: n = 69; Animal 2: n = 1; Animal 3: n = 16; and Animal 4: n = 8, respectively), however, the overall distribution patterns were strongly overlapping. The maximum soma cross sectional area of LVlat SPBNs was similar across three animals but larger in the remaining one (Maximum cross-sectional area: Animals 1, 2, 4 vs. 3 (A1: 178.4 ± 39 mm2, A2: 189 ± 0 mm2, A4: 187.3 ± 89 mm2 vs. A3:242.2 ± 69 mm2; ANOVA: p = 0.000). The number of primary dendrites was consistent across all animals with a population average of 5.3 ± 1.2 dendrites (Primary dendrites = A1: 5.2 ± 1.2 vs A2: 5 ± 0 vs A3: 5.5 ± 1.1 vs A4 5.62 ± 1.6; ANOVA, p = 0.717). Close inspection of the primary dendrite relationship with LVlat SPBN somas suggested most of these neurons had multipolar morphology (primary dendrites > 4: 90/94; 95.7%), some exhibiting as many as 7 or 8 dendrites (n = 16 and n = 3, respectively). The remaining LVlat SPBNs exhibited pyramidal morphology with 3 primary dendrites (n = 4/94: 4.3%).
A detailed morphological analysis of was undertaken for 10 LVLat SPBNs, reconstructed in the sagittal plane over 4 consecutive 50µm sections (Figs. 4 and 5). Despite exhibiting a substantial total dendritic length (Range: 2585–5859 µm, Average: 3684 ± 875 µm), these cells had relatively simple dendritic profiles with primary dendrites exhibiting few branching points (Fig. 5B: Branching points: n = 54 dendrites: range: 1–8 total branches, Average: 2.8 ± 1.6 branches/dendrite; Primary dendrite length: range: 40-2503µm, Average: 573 ± 445µm). When comparing the territories occupied by LVLat SPBN dendrites, they were principally oriented in the rostrocaudal compared to dorsoventral axis (RC: 735 ± 198µm vs. DV: 342 ± 55µm; RC/DV ratio = 2.2 ± 0.8), averaging twice the territory in the rostrocaudal versus dorsoventral plane (Fig. 5C). Within the dorsoventral axis, there few dorsally oriented dendrites that extended outside the LVLat cell column, yielding a bias towards ventrally projecting dendrites (D:116 ± 49 µm vs. V: 228 ± 64 µm; D/V ratio = 0.57 ± 0.3). Reconstructions in the mediolateral plane, across multiple sections (Fig. 5D) highlighted a bias for LVLat SPBN dendrites towards the lateral white matter, with the majority of dendritic profiles remaining lateral to the cell soma (Fig. 5E: M: 505 ± 220 µm vs. L: 3219 ± 902 µm; M/L ratio = 0.17 ± 0.09). Further, comparison of the proportion of dendritic reconstructions across sections showed that only 13% occupied territory medial to the cell somas, 25% remained within the LVLat cell column, 37% was in the lateral white matter adjacent to the LVLat cell column, and 25% in the most lateral white matter region.
LVlat SPBN photostimulation recruits other spinal populations
To facilitate a search for postsynaptic collateral connections, LVLat SPBNs were optogenetically stimulated and the expression of the activity marker, phosphorylated extracellular receptor kinases (pERK) was assessed in the spinal cord. First, efficacious rAAV2-mediated transduction with ChR2 and photostimulation responses in LVLat SPBNs were confirmed using targeted patch clamp recordings in spinal cord slices (Fig. 6A-C). On-cell recordings avoided altering cell excitability as has previously been demonstrated for the whole-cell configuration (Gradwell 2022). All LVLat SPBN tested exhibited robust AP discharge in response to photostimulation (n = 3, Fig. 6D), confirming the model was adequate to assess connectivity. Next, mice with unilateral rAAV2-ChR2 virus injections (n = 9) received partial unilateral laminectomies and spinal photostimulation was applied under anaesthesia to elicit phosphorylated ERK expression in any activated neurons. Three paradigms were assessed to control for pERK expression unrelated to SPBN photostimulation: 1) rAAV2-mediated ChR2 expression with photostimulation as the active trial (ChR2:PS+; n = 4); 2) rAAV2-mediated ChR2 expression without photostimulation (ChR2:PS- ;n = 3) as a control for the spinal exposure and laminectomy procedures; and 3) rAAV2-mediated GFP expression with photostimulation as a control for spinal photostimulation (GFP:PS+;n = 2). At the conclusion of each paradigm animals were perfused and tissue processed to identify pERK-positive profiles (Fig. 6E-G). A total of 77 transverse spinal sections were analysed, 36 for the ChR2:PS + photostimulation trials, and 41 for the ChR2:PS- and GFP:PS + controls, with 7–10 sections assessed per animal (summarised in Table 1). A total of 2120 pERK positive profiles were identified across both ipsilateral and contralateral dorsal horns (relative to photostimulation). There were significantly more pERK profiles across the superficial (Fig. 6H) and deep DH (Fig. 6I) on both the ipsilateral and contralateral spinal cords of the ChR2:PS + trial animals, compared to the either control (SDH Ipsi: 7.5 ± 4.1 vs 20.2 ± 11.1 neurons; n = 50; z = 5.856, p = 0.000**, ChR2:PS- vs ChR2:PS+), (3.67 ± 2.8 vs 20.11 ± 11.1 neurons; n = 63, z = 4.452, p = 0.000**, GFP:PS- vs ChR2:PS+), and a comparison of all 3 conditions also highlighted the significantly increased pERK profiles in the active trial (ChR2:PS- vs GFP:PS + vs ChR2:PS+: 20.1 ± 11.1 vs 7.5 ± 4 vs 4.1 ± 3.1 neurons; ANOVA: 32.684, n = 77, P = 0.000). These results indicate that SPBN activation does engage spinal circuits through axon-collateral signalling.
Regarding the distribution of the photostimulation-activated cells, the greatest number of pERK positive profiles were detected in the SDH ipsilateral to photostimulation (Table 1; Fig. 6J), and there were significantly more pERK positive profiles per section in the ChR2:PS + versus controls (SDH-Ipsi: 20.1 ± 11.1, p = 0.000; Fig. 6A & 6D). Surprisingly, the contralateral SDH had the next most pERK profiles (SDH-Contra: 11.2 ± 9.6, p = 0.000), followed by the ipsilateral DDH (DDH-Ipsi: 9.5 ± 6.4; p = 0.000; Fig. 5A and 5F) and the contralateral DDH (DDH-Contra: 6.0 ± 6.1; p = 0.000; Fig. 6E). This indicates that photostimulation of LVlat SPBNs recruited neurons across the spinal cord, well above non-specific pERK activity induced by the surgical preparation and spinal cord light exposure alone. Given PBN injection of rAAV2 also transduced LVLat SPBN terminals in LX around the central canal, we compared pERK labelling between photostimulation and control trials in this area (Fig. 6K-L). This analysis showed a significant increase in pERK fluorescence in the ChR2:PS + trial to controls (2.1 ± 1.5 vs 1.2 ± 0.45 vs 1.1 ± 0.14 arbitrary units, ANOVA: 7.711, P = 0.001, N = 76; Tukeys HSD: ChR2:PS+** > GFP:PS+, ChR2:PS-). Together, these data confirm that as LVlat SPBNs are activated, their local axons collaterals recruit other DH populations across the superficial and deep dorsal horns bilaterally.
LVlat SPBNs provide synaptic input into superficial dorsal horn circuits
In light of the substantial activation of superficial DH neurons by LVlat SPBN photostimulation, we next sought to test for specific connectivity and the circuits underlying this observation. Channelrhodopsin-2 assisted circuit mapping (CRACM) of LVlat SPBN axon collateral inputs was undertaken in transverse spinal cord slices with brief (1ms) whole field photostimulation applied while patch clamp recordings were made from unidentified DH neurons (Fig. 7A). A criterion was used for identification of input arising from SPBN axon collateral activation, where putative optically evoked postsynaptic currents (oPSCs) needed to occur at relatively short latency following photostimulation in multiple trials and exceed the mean background instantaneous sPSC frequency ± 4SD (see Methods). In addition, a criterion latency of 8ms was used to differentiate monosynaptic and polysynaptic optically evoked postsynaptic currents (monosynaptic oPSC latency < 8ms; polysynaptic > 8ms). This threshold was derived from mean value (+ 4SD) of the recruitment delay between photostimulation onset and AP initiation in ChR2-expressing SPBNs (2.88 ± 0.67 ms, Fig. 6D), with 2 ms added for conduction and synaptic delays taken from previous paired recording studies in the dorsal horn 41. A total of 18 neurons (n = 18/106, 17.0%) met this criterion for input derived from local SPBN axon collaterals (Fig. 7B). The latency, jitter and amplitude of the photostimulation evoked inputs (Fig. 7D) allowed for differentiation into putative monosynaptic excitatory connections (n = 5/18, 27.7%: latency: 4.65 ± 2.6 ms: jitter (SD of latency): 0.8 ± 0.5 ms; amplitude: 31.85 ± 18 pA), polysynaptic excitatory connections (n = 11/18, 61.1%; latency: 23.59 ± 13.7 ms; jitter: 2.63 ± 1.0 ms; amplitude: 18.19 ± 14.0 pA), and polysynaptic inhibitory connections (n = 2/18, 11.1%; latency: 31.39 ± 19.4 ms; amplitude: 15.59 ± 1.2 pA). One recording exhibited a monosynaptic excitatory (inward) current and polysynaptic inhibitory (outward) current (Fig. 7C). When recorded in current clamp, all photostimulation evoked inputs remained subthreshold (ie, did not evoke AP discharge). A subset of recordings was also assessed following bath applied AMPA/Kainate receptor antagonist, 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; n = 5), which abolished all photostimulation-evoked responses confirming their dependence on glutamatergic signalling. The dorsoventral and mediolateral position of recorded neurons receiving LVlat SPBN-derived input was mapped on a transverse spinal cord template, highlighting the location of responsive neurons concentrated in the superficial DH and distributed across the mediolateral dimensions (Fig. 7E). Together, these findings show for the first time that axon collaterals from a population of LVlat SPBNs activate a range of circuits likely to influence spinal nociceptive processing.
Electrophysiological characteristics of cells receiving LVlat SPBN-derived input
Based on the well-established literature associating electrophysiological features with neuron identity in the DH (eg, excitatory vs. inhibitory) 42–44 the electrophysiological characteristics of neurons that received LVlat SPBN collateral input were assessed and compared (Fig. 8). This sample was separated into 3 populations: neurons with monosynaptic SPBN input (mono-oPSC); polysynaptic SPBN input (poly-oPSC); and neurons that did not exhibit connections (UN). First, baseline characteristics such as cell input resistance and membrane capacitance were similar across populations (UN vs mono-oPSC vs poly-oPSC; n = 83, 5, 12; Rin = 406.55 ± 195.9 MΩ vs 330.60 ± 104.5 MΩ vs 407.25 ± 241.7 MΩ, P = 0.71: Cm = 11.90 ± 6.4 pF vs 12.78 ± 3.9 pF vs 12.97 ± 12.4 pF, P = 0.86), as was the rheobase current required to evoked AP discharge and the AP threshold (rheobase: 57.68 ± 44 pA vs 64 ± 76.7 pA vs 49.23 ± 32.3 pA, P = 0.77; AP threshold: -29.34 ± 8.7 mV vs -34.77 ± 7.2 mV vs – 31.89 ± 5.3 mV, P = 0.26, UN vs mono-oPSC vs poly-oPSC; n = 70, 5, 12). In contrast, the incidence of neurons exhibiting specific AP discharge patterns did differ between the mono-oPSC, and poly-oPSC, and UN populations (Fig. 8A-B, Χ2(12, N = 92) = 21.124, p = 0.049). Specifically, tonic firing (which is associated with inhibitory cell phenotype) was most common in neurons that received a monosynaptic SPBN input (TF: n = 4, 80%), while the remaining monosynaptically connected neuron exhibited a delayed firing response. Cells that received a polysynaptic input (n = 12) exhibited a wider range of discharge patterns including initial bursting (IB: n = 4, 33.3%), tonic firing (TF: n = 3, 25.0%), Gap firing (GF: n = 2, 16.7%), and single examples of delayed firing, phasic firing, and single spiking. Finally, neurons that did not receive LVlat SPBN-derived input (n = 75) exhibited several discharge patterns typical of a random sample, including prevalent delayed firing (DF: n = 29, 38.7%), tonic firing (TF: n = 19, 25.3%), initial bursting (IB: n = 16, 21.3%), single spiking (SS: n = 6, 8.0%), reluctant firing (RF: n = 3, 4%), and individual examples of phasic firing and gap firing.
Consistent with a high prevalence of tonic firing in neurons with monosynaptic oEPSCs, AP discharge duration, discharge frequency, and spike frequency adaptation were significantly greater in this sample (Fig. 8C, mono-oPSC vs poly-oPSC vs UN: duration: 886.87 ± 144.9 ms vs 533.45 ± 298.3 ms vs 454.16 ± 391.6 ms, P = 0.000, Tukey HSD; mono-oPSC*<poly-oPSC < UN, P = 0.037, n = 5, 12, 60; AP frequency: 20.40 ± 9.3 Hz vs 8.67 ± 6.0 Hz vs 8.53 ± 7.5 Hz, P = 0.004, Tukeys HSD: mono-oPSC**>UN > poly-oPSC, P = 0.003, 0.010; Adaptation: 82.58 ± 20.9% vs 57.76 ± 26.5% vs 48.50 ± 29.9%, P = 0.037, Tukeys HSD: mono-oPSC*> UN, P = 0.036). In contrast, spike latency, mean interspike interval and spike height attenuation were similar across the sample (mono-oPSC vs poly-oPSC vs UN, n = 5, 12, 60 respectively; Latency to AP: 77.28 ± 119.7 ms vs 39.88 ± 39.6 ms vs 81.62 ± 162.8.0 ms, P = 0.68; interevent interval: 50.89 ± 43.7 ms vs 56.03 ± 38.3 ms vs 45.48 ± 47.1 ms, P = 0.76; attenuation: 77.63 ± 28% vs 84.69 ± 110% vs 59.18 ± 44%, P = 0.34). The incidence of different subthreshold voltage-activated currents was also in agreement with a high incidence of tonic firing in neurons receiving mono-oPCSs. Specifically, the incidence of IA potassium currents that supress AP discharge was lowest in cells receiving mono-oPSCs (mono-oPSC vs poly-oPSC vs UN: 33.3% vs 66.7% vs 63.4%), whereas the incidence of hyperpolarisation-activated IH currents that support repetitive AP discharge was common across the sample (~ 66.6%).
Finally, the properties of spontaneous excitatory synaptic input were compared across the sample (Fig. 8D-F). sEPSC frequency was lower in neurons that received mono-oPSC input versus those lacking SPBN axon collateral mediated connections (mono- oPSC vs poly-oPSC vs UN; 4.24 ± 2.3 Hz vs 4.92 ± 3.9 Hz vs 9.79 ± 10.7 Hz, Welchs ANOVA, P = 0.009). There were no differences, however, in sEPSC amplitude (mono-oPSC vs poly-oPSC vs UN; -19.13 ± 6.9 pA vs -18.1 ± 6.0 pA vs -16.76 ± 6.4 pA, P = 0.62), time-course or charge (rise time: 0.81 ± 0.1 ms vs 1.03 ± 0.1 ms vs 1.02 ± 0.ms, P = 0.18; Tau: decay time constant: 4.26 ± 1.3 ms vs 5.59 ± 1.1 ms vs 4.86 ± 1.1 ms, P = 0.07; charge: 100.17 ± 36.0 pA.ms vs 131.59 ± 71.1 pA.ms vs 108.85 ± 47.4 pA.ms, P = 0.34. Together, this profile of tonic AP discharge, low expression of IA currents, and low sEPSC frequency is consistent with monosynaptic LVlat SPBN input from axon collaterals preferentially targeting inhibitory DH populations 42,45–47.