Activation of the periaqueductal gray controls respiratory output through a distributed brain network

Figures

• 

  FIGURE 1. Viral expression profile for optogenetic stimulation of the lateral PAG: (A) Expression map for optogenetic virus and optic fibre locations. Green clouds indicate the extent of GFP expression in three coronal sections arranged from caudal (left) to rostral (right) along the core of the PAG. Clouds from each experiment are overlaid (n = 13). Purple dots indicate optic fibre tip locations. (B) Examples of immunofluorescence staining in the corresponding sections above in A for GFP (green), cFOS (red), and nNOS (white). Dashed lines demarcate the boundaries for PAG subdivisions based on nNOS fluorescence and histological landmarks (dm, dorsomedial; dl, dorsolateral; l lateral; vl, ventrolateral). Insets below show high power magnification of immunostained tissue in separate channels. Right panel highlights boundaries between lateral and dorsolateral PAG based on nNOS staining.

• 

  FIGURE 2. Optogenetic activation of PAG potently stimulates inspiratory and expiratory muscle activity and increases respiratory rate. (A) Example traces of raw diaphragm (DIA), genioglossus (GG), and abdominal (ABD) EMG activity together with breaths per minute (BPM) before and after 2 bursts of high frequency photostimulation (blue boxes). Photostimulation potently increases EMG activity at all three sites while also dramatically increasing BPM. Black box indicates the zoomed area shown in (B). (B) Expanded integrated traces of EMG activity from A before and after the onset of laser stimulation (blue box) showing the rapid enhancement of EMG activities along with increasing respiratory rate. (C) Peak amplitude of integrated DIAEMG and GGEMG activity during laser stimulation normalized to activity before stimulation (n = 15). Dots correspond to individual animals. Boxes define the 2nd and 3rd quartiles surrounding the median (solid line), whiskers show 1st and 4th quartiles. Dashed lines at 1 indicate the amplitude of baseline activity, stars mark significant differences from 1 as measured by a sign test or t-test at the p = 0.05 level. (D) Similar to C, dots indicate the mean amplitude…

• 

  FIGURE 3. Brief stimulation of the PAG elicits phase-locked EMG activity after a short delay in the same sequence as natural breathing. (A) Example integrated EMG traces for three different instances of brief (250 ms) pulse stimulation (blue boxes) delivered at different phases of the natural ongoing breathing rhythm. Note the increased DIA and GG amplitude and duration in the left trace and the rapidly-evoked activity shortly after laser onset in the middle and rightmost traces. (B) Stimulus-triggered overlay of integrated EMG activity. Five breathing cycles during which stimulus pulses arrived at different phases of the ongoing respiratory cycle are overlaid and time-locked to the onset of laser pulses (lag = 0s). Apparent reset occurs at short latencies and is also observed for the subsequent cycle. (C) Stimulus-triggered average across all cycles for the experiment shown in (A), (B). Integrated traces were z-scored to highlight the prominent stimulus-evoked activity. Shaded areas around the mean traces indicate ± S.E.M. (D) Mean stimulus-triggered activity for all brief PAG stimulation experiments (n = 26). Stimulus-triggered averages from each experiment were binned into phases of the…

• 

  FIGURE 4. PAG stimulation produces 2 types of respiratory reset. (A) Diagram of phase measures obtained from integrated DIAEMG traces used to quantify respiratory reset. Top trace exemplifies the expected phase of the cycle subsequent to photostimulation if stimuli did not elicit any changes to respiratory timing. Bottom trace shows the induced phase between stimulus and an advanced bout of EMG activity elicited by the laser pulse (in blue). Differences between the expected and induced phase indicate changes in respiratory timing. (B) Stimuli arriving to different phases of the respiratory cycle (example respiratory cycle aligned and overlaid at the top of the panel) altered the onset of the subsequent cycle in a phase-dependent manner. Induced phase for DIA activity across all experiments (n = 26, black), shaded area represents ± S.E.M. Expected phase (green) indicates the mean and 95% confidence limits derived from a spontaneously-derived surrogate dataset. Stimulation can be seen to significantly alter the induced from the expected phase of inspiration at multiple phase points of the ongoing rhythm. (C) Induced phase curve for an example experiment following phase-dependent reset. Dots…

• 

  FIGURE 5. cFOS expression following repeated PAG stimulation in the respiratory network. Overlaid immunofluorescence for GFP (green), cFOS (red), and NeuN (blue) in rats infected with the hSYN-ChR2-eYFP virus (left) and a naive animal (right) in the preBötzinger Complex (preBötC, A), nucleus of the solitary tract (NTS, B), the lateral parafacial (pFL, C), the ventral parafacial (pFV, D), the intermediate reticular formation (iRT, E) the locus coeruleus (LC, F) and the parabrachial nuclei of the Kölliker-Fuse (KF) and lateral parabrachial (LPB, G). Insets below show high power magnification of immunostained tissue in separate channels. 7n, Facial motor nucleus; cNA, nucleus ambiguus, pars compacta; PiCO, post-inspiratory complex; 4V, fourth ventricle; scp, superior cerebellar peduncle.

• 

  FIGURE 6. Stimulation of PAG terminals in the preBӧtC and pFL does not produce the same effects as direct PAG stimulation. (A) Example traces of raw diaphragm (DIA), genioglossus (GG), and abdominal (ABD) EMG activity together with breaths per minute (BPM) before and after 2 high frequency bursts of photostimulation of PAG terminals in preBӧtC (blue boxes). Photostimulation increases EMG activity in DIA and GG (but not ABD) sites and also increases BPM. Black box indicates the zoomed area shown in (B). (B) Expanded integrated traces of EMG activity from A before and after the onset of laser stimulation (blue box) showing the enhancement of EMG activities in DIA and GG traces (although not ABD) along with and increased respiratory rate. (C) Similarly to A, example traces showing raw EMG activity and respiratory rate effects elicited by stimulating PAG terminals in pFL. (D) Magnified traces from the example in (C). The effects of stimulation in pFL were more modest than that for the preBӧtC. (E) Peak amplitude of integrated DIA and GG EMG activity following high frequency burst photostimulation of the PAG (black, n = 15), PAG terminals in preBӧtC (purple, n = 8), and PAG terminals in pFL…

• 

  FIGURE 7. Brief stimulation of PAG terminals in preBӧtC does not elicit changes in respiratory EMG activities nor changes in respiratory timing. (A) Example integrated EMG traces for three different instances of brief (250 ms) pulse stimulation (blue boxes) delivered at different phases of the natural ongoing breathing rhythm. No apparent alteration in the rhythmic expression of EMG activities are observed. (B) Stimulus-triggered overlay of integrated EMG activity. Five breathing cycles during which stimulus pulses arrived at different phases of the ongoing respiratory cycle are overlaid and time-locked to the onset of laser pulses (lag = 0s). No stimulus locking of subsequent rhythmic EMG is apparent. (C) Stimulus-triggered average for all experiments. Integrated EMG activities (see legend) were normalized in both time/phase and amplitude and aligned to the onset of stimulus (0 lag, dashed blue line) and averaged to highlight the lack of locked responses post-stimulation across all experiments (n = 20). Shaded areas indicate ± S.E.M. (D) Stimulus pulses arriving at different phase of the respiratory cycle (aligned and overlaid above D) failed to evoke changes in the induced phase (purple…

• 

  FIGURE 8. Brief stimulation of PAG terminals in pFL does not elicit changes in respiratory EMG activities nor changes in respiratory timing. (A) Example integrated EMG traces for three different instances of brief (250 ms) pulse stimulation (blue boxes) delivered at different phases of the natural ongoing breathing rhythm. No apparent alteration in the rhythmic expression of EMG activities are observed. (B) Stimulus-triggered overlay of integrated EMG activity. Five breathing cycles during which stimulus pulses arrived at different phases of the ongoing respiratory cycle are overlaid and time-locked to the onset of laser pulses (lag = 0s). No stimulus locking of subsequent rhythmic EMG is apparent. (C) Stimulus-triggered average for all experiments. Integrated EMG activities (see legend) were normalized in both time/phase and amplitude and aligned to the onset of stimulus (0 lag, dashed blue line) and averaged to highlight the lack of locked responses post-stimulation across all experiments (n = 4). Shaded areas indicate ± S.E.M. (D) Stimulus pulses arriving at different phase of the respiratory cycle (aligned and overlaid above D) failed to evoke changes in the induced phase (green)…