Slow Breathing Can Be Operantly Conditioned in the Rat and May Reduce Sensitivity to Experimental Stressors

Figures

• 

  Figure 1. Use of light as negative reinforcement for respiratory rate feedback. Schematic depicts feedback procedure and representative respiratory traces from yoked control (top) and slowed respiratory rate (SRR) conditioned (bottom) animals over identical 5 s time intervals. Respiratory rate (RR) was continuously recorded via whole body plethysmography. For conditioning, real-time RR was compared to a predetermined target RR by a controller coupled to a transducer to trigger the light stimulus, such that the strobe light was turned off whenever current RR ≤ target RR (set to ≤ 80 breaths/min). Animals thereby experienced real-time changes in strobe illumination corresponding to their RR. This represents a closed-loop feedback control system for conditioning of SRR. In the example traces shown, the average RRs were 70.5 (conditioned) and 91.0 (yoked control) breaths/min.

• 

  Figure 2. Timeline for experimental procedures. All animals underwent 20 sessions of SRR conditioning or paired yoked control procedures (n = 11 per group). A subset of rats (n = 8 per group) also underwent 3-5 SRR retention tests with no reinforcement and only background red light (“dark sessions”), intermittent with conditioning sessions. In this same subgroup, open field testing occurred following Session 20 of conditioning. Nociception assays then took place over 4 days, 5–6 days after conditioning (except for one cohort of n = 2 pairs that was delayed approximately a month and underwent several SRR reinstatement sessions beforehand); a subset of rats were injected with formalin on Day 4 to evaluate the development of mechanical allodynia in a von Frey Test conducted 75 min later (n = 4 per group). Acute restraint sessions and blood collection for plasma corticosterone assays were performed during the first 3 days of nociception assays. Not shown here are acclimation and nose poke sessions (described in Materials and methods). A separate subgroup of FRR conditioned rats and their yoked controls (n = 4 per group) underwent identical testing procedures to those shown here but received…

• 

  Figure 3. SRR conditioned rats learned to breathe more slowly and spent more time breathing below the conditioned RR threshold of 80 breaths/min. Group differences in respiration emerged over the course of 20 sessions. (A) The percentage of samples where RR met criterion (≤80 breaths/min). SRR conditioned rats displayed an increasing incidence of breaths below the target RR (1-factor RM ANOVA), and a greater % of samples under threshold than yoked controls (2-factor RM ANOVA: significant main effect of Group) over training. Multiple comparisons analysis revealed a significantly higher % of samples below threshold in SRR conditioned vs. yoked rats during Sessions 4, 6, 9, 10, 12, 14–17, and 20. (B) SRR conditioned rats decreased their average resting RR by 10.6 breaths/min (1-factor RM ANOVA) and their RR was lower than that of yoked controls (2-factor RM ANOVA: significant main effect of Group) over training. Follow-up testing revealed a significant difference between SRR conditioned and yoked animals in resting RR during Sessions 10, 14–16, and 20; conditioned rat resting RR values were >6 breaths/min lower than yoked control values in all cases. Note in (A) that % of samples under…

• 

  Figure 4. Respiration became slower and more regular over 20 sessions in SRR conditioned rats. (A,B) SRR conditioning led to a leftward shift in RR values and a skewed distribution centered just below the target RR. Shown are frequency histograms depicting the number of occurrences of different instantaneous RRs (5 breaths/min bins) over one 2 h pre-conditioning period or three conglomerated 2 h conditioning sessions in a representative pair of SRR conditioned (blue) and yoked control (red) animals. (A) Distributions during pre-conditioning. (B) Frequency histograms show the gradual leftward shift in RRs. Decreased rates (on the x-axis) are easily visualized in the conditioned rat, as is the increasing tendency to breathe just below the feedback threshold of 80 breaths/min (vertical black dotted lines). (C) Mean standard deviation of resting RR was used to assess reduction in RR variability. This index was reduced at Sessions 14, 15, and 20 in SRR conditioned vs. yoked rats, indicating more regular respiration. *p < 0.05 SRR conditioned vs. yoked, post-hoc Holm-Sidak comparisons.

• 

  Figure 5. Reduced RR was retained between SRR conditioning sessions. (A) Individual frequency histograms for one pair of SRR conditioned (blue) and yoked control (red) rats during two representative retention sessions. These 2 h sessions were interspersed with normal conditioning and 4 days apart from one another (following Sessions 7 and 10, respectively). The RR threshold for conditioning (80 breaths/min) is shown with vertical black dotted lines. (B) Pooled results for all animals and retention sessions. SRR conditioned rats maintained trained reductions in RR during SRR retention sessions (*p < 0.05, paired t-test). Values averaged across SRR retention sessions were 87.9 ± 2.5 breaths/min for SRR conditioned rats (pre-conditioning: 96.3 ± 2.7 breaths/min) and 91.6 ± 2.0 breaths/min for yoked controls (pre-conditioning: 95.1 ± 3.2 breaths/min).

• 

  Figure 6. SRR conditioning reduced anxiety-like behavior in an open field and the respiratory response to acute restraint. Following 20 sessions of conditioning, rats were transferred to an open field and anxiety-like behavior was scored during 10-min sessions. (A) SRR conditioned rats showed a reduced latency to enter the center of the field vs. yoked controls (*p < 0.05, Mann-Whitney Rank Sum test). (B) SRR conditioned rats also showed a trend toward reduced distance traveled, in the arena as a whole (total) as well as in both defined areas (center and outside). (C) Rats later underwent 2 consecutive days of acute restraint. (C1) During these 10-min sessions, respiratory recordings were collected via electric field sensors positioned outside the restraint chamber. (C2) SRR conditioned rats had a lower RR response to restraint than yoked rats on the second day (*p < 0.05, Student's t-test), and decreased their RRs from Day 1 to Day 2 (*p < 0.05, paired t-test).

• 

  Figure 7. SRR conditioning prevented expression of mechanical allodynia following formalin injection. Shown are results for tests of thermal nociception in an open chamber (Hargreaves) or under restraint (Tail Flick), and mechanical nociception in an open chamber (von Frey) with or without prior formalin injection into the rat's right dorsal hindpaw. (A) There were no group differences in thermal sensitivity at baseline, despite SRR conditioned rats showing a trend toward increased responsiveness to the Hargreaves Test on Day 1 [10.4 ± 0.8 s vs. Yoked: 13.2 ± 1.1 s; t(14) = −2.02, p = 0.064; power (α = 0.05) = 0.47] and Tail Flick Test on Day 2 [4.8 ± 0.7 s vs. Yoked: 6.4 ± 0.5 s; t(14) = −1.81, p = 0.093; power (α = 0.05) = 0.39]. Yoked rats took more time to respond on Day 2 than Day 1 of the Tail Flick Test (*p < 0.05, paired t-test). (B) Results from Days 1 and 2 on the mechanical von Frey Test did not clearly differentiate the two groups of rats. (C) Mechanical nociceptive responses to formalin injection were reduced in SRR conditioned vs. yoked rats on Day 4 of testing (*p < 0.05, Student's t-test).