Heart rate variability measures indicating sex differences in autonomic regulation during anxiety-like behavior in rats

Figures

• 

  Figure 1. No sex differences in behavioral measures of the Novelty Suppression of Feeding (NSF) task. (A) Cartoon of the NSF arena, where a food-restricted rat has access to food in the center of the arena (circle) which is brightly lit (yellow cone). The task consists of a series of progressive measures before food consumption, including latency to first approach, the number of approaches, time in center (tolerating the light), and latency to grab the food. (B) Food intake. (C) Latency to first approach the center. (D) Number of approaches. (E) Time in center. (F) Latency to grab food. No average differences, although variability of latency to approach was greater in females (&&& p < 0.001), and variability of time in center was greater in males (& p < 0.05).

• 

  Figure 2. Female–male differences in basal HR/HRV measures. (A) Examples of HR traces with lower HRV (top) and higher HRV (bottom, exaggerated cartoon version). (B–E) Relative to males, females had (B) greater HR, (C) lower SDNN, (D) no rMSSD differences, and (E) lower SDNN/rMSSD ratio. To perform two-way ANOVA, data were log-normalized for SDNN, rMSSD, and SDNN/rMSSD. (F–I) Males show larger baseline-to-NSF changes in several HRV measures, when calculating percent change in a given measure within-rat. *,**,*** p < 0.05, p < 0.01, p < 0.001.

• 

  Figure 3. Description of how to calculate different HRV measures, including how they could relate to PNS and/or SNS. Specifically, some pharmacological agents are considered to impact the SNS, through the beta-adrenergic receptors (which are blocked by propranolol), or the PNS, through the muscarinic cholinergic receptors (which are blocked by atropine) [this interpretation does require caution since laying down is related to low SNS (17, 18, 23, 24)].

• 

  Figure 4. Males eat more food after NSF. Because the NSF is a brief test (10 min), amount of food consumed is typically not corrected for body weight. In support of this, rats had 30 min to eat food in the home cage after the NSF and LDB testing, and (A) male rats ate significantly more food in the post-NSF intake period than females (MW = 172.5, p = 0.0065), as seen previously (5). (B) Also, if rat intake in the 10 min NSF task was titrated by body weight and nutritional need, one might predict that rats that ate more during NSF would eat less in the later home-cage intake period. However, NSF intake and later home cage intake were not correlated in females (F (1, 21)=0.118, R2 = 0.006, p = 0.7348) or males (F (1, 25)=1.328, R2 = 0.050, p = 0.2601). Also, females had a trend for greater intake post vs. during NSF (Wilcoxon matched-pairs signed-rank test, diff in medians = 0.600, p = 0.0633), while males had significantly more intake post vs. during NSF (Wilcoxon, diff in medians = 1.300, p < 0.0001). Thus, it is reasonable to conclude that food intake during NSF was small enough that it does not need to be corrected to body weight, supporting the interpretation that there was no sex…

• 

  Figure 5. Across all rats, females with lower SNS indicators ate more food. (A–C) Food intake did not correlate with HR (A) at baseline (females: F(1, 21) = 1.180, R2 = 0.053, p = 0.2897; males: F(1, 25) = 0.158, R2 = 0.006, p = 0.6946), (B) during NSF (females: F(1, 21) = 0.554, R2 = 0.026, p = 0.4651; males: F(1, 25) = 0.472, R2 = 0.019, p = 0.3984), or (C) change from baseline to NSF period (females: F(1, 21) = 2.944, R2 = 0.123, p = 0.1009; males: F(1, 25) = 0.021, R2 = 0.001, p = 0.8851). (D–F) Food intake did not correlate with SDNN (D) at baseline (males: F(1, 25) = 1.225, R2 = 0.047, p = 0.2789) or (E) during NSF (males: F(1, 25) = 0.975, R2 = 0.038, p = 0.3330), but (F) females with larger drop in SDNN from baseline to NSF ate more food (males: F(1, 25) = 0.824, R2 = 0.032, p = 0.3728). (G–I) Food intake did not correlate with rMSSD (G) at baseline (females, F(1, 21) = 1.160, R2 = 0.052, p = 0.2938; males, F(1, 25) = 0.146, R2 = 0.006, p = 0.7057), (H) during NSF (females, F(1, 21) = 0.153, R2 = 0.007, p = 0.6992; males, F(1, 25) = 0.228, R2 = 0.009, p = 0.6373), or (I) change from baseline to NSF period (females: F(1, 21) = 1.547, R2 = 0.069, p = 0.2273; males: F(1, 25) = 0.179,…

• 

  Figure 6. Females (but not males) with higher SNS indicators, including greater HR at baseline, had slower latency to first approach. (A,B) Females with higher HR (A) at baseline or (B) during NSF had significantly slower initial latency to approach the center (males: basal: F(1, 25) = 1.413, R2 = 0.053, p = 0.2458; NSF: F(1, 25) = 0.264, R2 = 0.010, p = 0.6118). (C) While basal SDNN did not correlate with slower latency (males: F(1, 25) = 0.059, R2 = 0.002, p = 0.8099), (D) females with lower SDNN during NSF had slower initial approach (males: F(1, 25) = 0.011, R2 = 0.001, p = 0.9175). (E,F) Females with lower rMSSD (E) at baseline or (F) during NSF had significantly slower initial latency to approach center [males: (E), F(1, 25) = 0.114, R2 = 0.114, p = 0.7386; (F), F(1, 25) = 0.115, R2 = 0.005, p = 0.7373]. No HRV measures related to male latency. Related correlations of log-transformed HR/HRV measures are shown in Supplementary Figure 11. *,**,*** p < 0.05, p < 0.01, p < 0.001.

• 

  Figure 7. Males (but not females) with lower SNS indicators had more approaches to and time in center. (A–D) While basal SDNN did not correlate with (A) number of approaches (males: F(1, 25) = 0.433, R2 = 0.017, p = 0.5167; females: F(1,19) = 1.319, R2 = 0.065, p = 0.2650) or (C) time in center (females: F(1, 19) = 0.017, R2 = 0.000, p = 0.8986), males with lower SDNN during NSF had (B) more approaches and (D) more time in center. (E–H) rMSSD did not correlate any measure, at baseline (E,G) or during NSF (F,H), or for (E,F) number of approaches or (G,H) time in center. (E), males: F(1, 25) = 0.000, R2 = 0.000, p = 0.9964; females: F(1, 19) = 2.800, R2 = 0.128, p = 0.1107; (F), males: F(1, 25) = 0.641, R2 = 0.025, p = 0.4309; females: F(1, 19) = 3.212, R2 = 0.145, p = 0.0891; (G), males: F(1, 25) = 1.031, R2 = 0.040, p = 0.3196; females: F(1, 19) = 0.071, R2 = 0.004, p = 0.7925; (H), NSF: males: F(1, 25) = 1.421, R2 = 0.054, p = 0.2444; females: F(1, 19) = 0.258, R2 = 0.013, p = 0.6171). (I–L) While basal SDNN/rMSSD did not correlate with (I) number of approaches (males: F(1, 25) = 0.454, R2 = 0.018, p = 0.5065; females: F(1, 19) = 0.001, R2 = 0.000, p = 0.9784) or (K) time in center…

• 

  Figure 8. In male consumers, greater SNS indicators predicted higher food intake. (A) Same data as Figure 3F but only males, with non-consumers at 0 g food intake, and consumers above that. (B–D) Male non-consumers had (B) greater SDNN/rMSSD, (C) trend for greater SDNN, and (D) fewer approaches, than males that consumed food during NSF. (E–J) data are for rats that consumed >0 g food during NSF. (E) Greater basal SDNN in females, but not males, was correlated with higher intake. (F) Greater SDNN during NSF predicted higher food intake in males but not females. (G) For the drop in SDNN (from basal to NSF), smaller drop in males related to higher intake, with the opposite trend in females (bigger SDNN drop for more intake). (H–J) SDNN/rMSSD showed a similar pattern as SDNN, with (H) higher basal ratio relating to greater female intake, (I) greater ratio during NSF associated with more male intake, and (J) opposite trends in females and males [similar to (G)]. *,** p < 0.05, p < 0.01. ## p < 0.01 female–male difference in slopes.

• 

  Figure 9. Limited changes in HR/HRV measures across time, centered on the time to grab food, an important step in the NSF. See Methods for how the by-time data were generated. (A–D) Both females and males show a dip in heart rate after the food grab. Note that, for a one-way ANOVA, data were first normalized by Z-score to a baseline period (150 to 70 s before grab food), then compared across baseline, an interval 40 to 30 s before food grab (gray rectangles in figure), and an interval 18 to 22 s after food grab (white rectangles, about the peak of the heart rate drop after food grab). (E–H) Females showed an increase in SDNN/rMSSD both before and after the food grab, while males did not. (I–L) No changes in SDNN before food grab. * p < 0.05.

• 

  Figure 10. Summary of sex differences in the relation between HRV patterns and specific NSF behaviors. For females, greater HR at baseline and during NSF predicted slower initial latency to approach, while lower SNS correlated with greater food intake. In males, lower SNS was associated with increased number of approaches and time in center, while greater SNS predicted higher food intake in rats that consumed food. “--” indicates no relation between HRV and behavioral measures.