Ventilatory phases of the Bullfrog.

(A) Extracellular recordings from cranial nerves (CN5, CN7, CN10 and SN1) displaying four phases of ventilation: B1 buccal-depression, B2 buccal-elevation, L1 lung-priming, and L2 lung-powerstroke. (B) Anatomic brainstem regions implicated in generating each ventilatory phase.

Functional ablation of motor neuron pools to focus on rhythmic interneurons using high-frequency nerve stimulation.

(A) Extracellular stimulation of SN1-R (red) with high-frequency selectively inhibits ventilatory output in stimulated nerve. (B) Bilateral high-frequency stimulation of CN10 and SN1 (highlighted in red) does not affect buccal and lung activity on CN5. (C) Quantification of activity (raw amplitude) pre- and post-high frequency stimulation illustrated in Panel B. Bars represent mean+SD (Two-way ANOVA, *** p<0.0001 and * p<0.05, n=34 for CN5 and CN10; n=21 for SN1).

Integrated nerve activity of minimum Buccal Area region capable of producing bursts.

(A) In this example, a brainstem segments spanning CN10 and SN1 roots produces reliable buccal-like bursts, especially when the network is excited with bath-applied AMPA. (B) After transection between CN10 and SN1, just caudal to the previously identified Buccal Area (i.e., Level 1 cut; n=6), buccal-like bursts can still be induced with bath-applied AMPA from the rostral slice (containing CN10) but not from the caudal slice (containing SN1). (C) After a slightly more rostral cut aimed at transecting the previously identified Buccal Area and yielding a thinner rostral and thicker caudal slice (i.e., Level 2 cut; n=6), both slices retain limited capacity to produce bursting with AMPA. Note transient activity with AMPA application, suggesting an optimal range of network excitation required for inducing bursting.

Group data of bursting activity following the transections illustrated in

Fig. 3. (A) The isolated buccal segment preparation that includes CN10 and SN1 maintains the capacity to produce bursts both at baseline (8 out 12 preparations) and with 125 nM AMPA (all preparations) (n=12). (B, C) Bursting in rostral CN10 slices (left panels) and caudal SN1 slices (right panels) resulting from the Level 1 cuts (n=6) and Level 2 cuts (n=6), respectively. While neither rostral (CN10) nor caudal (SN1) slices were able to generate spontaneous bursts without AMPA, AMPA caused a significant recruitment of bursting in rostral slices showing persistent bursting capacity in all preparations (compared to zero: p=0.031). In contrast, bursting in caudal slices was practically absent; AMPA failed to induce any bursts in thinner caudal slices, and only induced bursts in 3 of 6 thicker caudal slices which was not enough to reach significance across the group (compared to zero: p>0.25). Bars represent mean±SD. Asterisk symbols denote difference from zero (Wilcoxon one sample test); p values denote dose-response differences (Friedman test plus Dunn’s multiple comparisons).

Survey of rhythmic units in the buccal region.

(A) The buccal region was divided into 28 boxes based on a grid defined using anatomical landmarks to facilitate a systematic survey of units in the buccal region. (B) Single units were defined according to their patterning of activity. Subsequent figures illustrate the number of units identified in each anatomically defined box.

Summary of buccal segment preparations.

Number of buccal slices (and their corresponding levels – Level 1 or Level 2) that displayed bursting in response to different concentrations of AMPA (Wilcoxon One sample Test).

In the isolated intact brainstem, network mild excitation increases the density and distribution of premotor buccal units in the buccal region.

CN10 and SN1 (red) motor neurons were functionally ablated before unit surveys. The raster displays the survey results. The numbers are indicative of the frequency of units found in each location from all preparations. Top panel displays unit frequencies under baseline conditions for 145 units pooled from 15 preparations. Bottom panel is from 207 units pooled from 6 preparation in the presence of 60 nM AMPA to cause mild network excitation (Chi-squared test, Bonferroni correction for multiple comparisons; there were significant differences in the proportion of all categories of extracellular units between the two groups; p<0.01). Ext: extracellular unit recording.

In the buccal segment encapsulating the buccal region, mild network excitation increases the density and distribution of premotor buccal units.

CN10 and SN1 (red) motor neurons were functionally ablated before unit surveys. The rasters display the survey results from 247 units pooled from three preparations. The numbers are indicative of the frequency of units found in each location. Top panel displays unit frequencies under baseline conditions from 110 units; bottom panel is from 137 units in the presence of 60 nM AMPA to cause mild network excitation. (Chi-squared test, Bonferroni correction for multiple comparisons; there was a significant difference in the proportion of buccal units between the two groups; p<0.01).

Local network excitation transforms the operational identity of premotor neurons.

(A) Representative extracellular recording from the caudal brainstem showing activity of two sorted units following local AMPA injection (shaded red). Top trace: raw extracellular signal with spikes color-coded. Bottom trace: integrated CN5 activity with lung bursts (blue circles) and buccal bursts (green circles). Before local AMPA injection, Unit 1 (brown) is active during both buccal associated CN5 bursts, the operational definition of a buccal unit. In contrast, Unit 2 (blue) shows lung burst-associated spikes but no buccal burst-associated spikes once effects of a lung episode have subsided, the operational definition of a lung unit. However, following AMPA application, this lung unit is transformed into a buccal-lung unit, illustrating local excitation causes network reconfiguration. (B) Principal component analysis (PC1 vs PC2) of spike waveforms demonstrating clear separation between the two most prominent units (see Results for silhouette scores and Mahalanobis distances). Marker shape indicates burst association: open circles denote spikes occurring within ±0.8 s of a lung burst peak; filled squares denote buccal-associated spikes. (C) Spike-triggered waveform averages for Units 1 and 2 comparing activity during lung bursts (left) versus post-AMPA buccal bursts (right). Individual waveforms shown in brown or blue with mean waveform overlaid in black. The consistent waveforms across conditions confirm unit identity (spike correlation coefficients: r>0.95). (D) Phase histograms for Unit 2 (the lung unit) showing spike timing relative to CN5 burst phase (phase 0 = burst peak; ±π = inter-burst midpoints). Top: during lung bursts, Unit 2 fires preferentially near the burst peak. Bottom: during post-AMPA buccal bursts, Unit 2 exhibits a similar phase-locked firing pattern, demonstrating recruitment of this lung unit during buccal activity. Dashed red line indicates the Z-score threshold (Z > 1.65) for significant recruitment above uniform baseline; dashed black line indicates median spike count. (E, F) Meta-analysis across five lung units (rows) showing phase-recruitment heatmaps and spikes per burst represented as violin plots during lung bursts, respectively (n=3–5 lung bursts analyzed per unit). Blue color intensity in E indicates Z-score for recruitment above chance (uniform distribution). Dots in F indicate individual burst spike counts. (G, H) Recruitment of same five lung units during post-AMPA buccal bursts (n= consecutive buccal bursts with lung unit activity / total consecutive buccal bursts available to analyze). The spiking and recruitment pattern during post-AMPA buccal activity in E-G is consistent with local excitation causing network reconfiguration.

Demonstration that sufficiency and necessity of the Buccal Area is dependent upon the excitation state of the network.

(A) Effect of AMPA (∼5 nl, 5 µM; sufficiency test) and GABA (∼15 nl, 5 mM; necessity test) microinjections into the Buccal Area of an isolated hemisected brainstem preparation. AMPA causes a brief increase in buccal burst frequency and GABA causes the buccal bursts to momentarily cease; injections of AMPA immediately followed by GABA demonstrate that GABA can completely counteract the excitatory effects of local AMPA injections. (B) When the above protocol was repeated with bath application of 60 nM AMPA, the effect of AMPA persist but GABA injections fail to cease buccal bursting suggest the discrete buccal area is replaced by a distributed rhythm generating network. (C) Fractional change in instantaneous frequency before and during global network excitation with 60 nM AMPA (One sample t-Test; *, p<0.01; nsd, no significant difference). Bars represent mean+SD; n=7.