Changes in heart rate variability during long-duration spaceflights
Changes in 24-h average HRV endpoints between space and Earth and between the two sessions in space are summarized in Table 1. Reflecting the dramatic changes in cardiovascular function taking place in space, parasympathetic activity was statistically significantly decreased during ISS01 and ISS02 compared to pre-flight in the 20 astronauts, as seen from 24-h averages of r-MSSD, pNN50, SDNNIDX5, SDNNIDX30 and HF-component. By contrast, HR (or NN-interval) and sympathetic activity estimated by the LF/HF ratio did not differ. The fractal scaling of HRV (slope β) was statistically significantly less steep in space than on Earth, as shown earlier [14, 15, 19], but did not change from ISS01 to ISS02.
While normal pre-launch, further increases during ISS01 are observed in SDNN (from 143.3 ± 42.5 to 153.4 ± 35.8 msec, P = 0.0681), SDANN5 (from 124.5 ± 37.6 to 133.8 ± 32.6 msec, P = 0.0775) and SDANN30 (from 117.5 ± 37.1 to 131.2 ± 33.3 msec, P = 0.0113), suggesting anti-aging effects, as previously reported [57, 58, 59].
Brain plasticity at night estimated by changed HRV dynamics in space
Changes in the circadian amplitude of NN intervals and the extent of sleep-related HR-dipping and HRV-rising during spaceflight served to assess brain plasticity at night (Table 1, bottom). Compared to pre-launch, the circadian amplitude of NN intervals increased from 116.9 to 134.7 msec (P = 0.0418) during ISS01 and to 132.6 msec (P = 0.1281) during ISS02. The nocturnal HR-dip became deeper during both ISS01 (P = 0.0309) and ISS02 (P = 0.0447), and the nocturnal HRV-rise accentuated during ISS02 in LF-component (P = 0.0454), LF-band (P = 0.1618) and MF1-band (P = 0.0087), which reflects DMN activity. The nocturnal rise in MF1-band and HF-band also increased statistically significantly between ISS01 and ISS02. In concert with the response in brain plasticity, suppression of the parasympathetic activity seen during ISS01 was no longer observed during ISS02 for r-MSSD, pNN50 and HF-component, even though nocturnal HR-dip and suppression of SDNNIDX5 and SDNNIDX30 lasted over both sessions in space (Table 1).
HRV measures before launch predict change in 12-h amplitude of NN intervals
Thresholds distinguishing Group I from Group II determined by ROC curve before launch were a 12-h amplitude of TF-component of 1529.4 or 2087.9 msec2 (sensitivity = specificity of 0.8182; AUC = 0.9798 ± 0.02604) and a 12-h amplitude of ULF-component of 1321.5 or 1776.2 msec2 (sensitivity = specificity of 0.9091; AUC = 0.9798 ± 0.02596).
Biological rhythm and sleep characteristics of Groups I and II
In Group I, the 12-h amplitude of TF-component increased about 2.4 times from pre-flight to ISS01 (from 801.6 to 1894.5 msec2, P = 0.0001), the 8-h amplitude increased about 1.4 times (P = 0.1130) and the 6-h amplitude about 1.7 times (P = 0.0317), Table 2. Similar changes were not statistically significant during ISS02, except for the 12-h amplitude (P = 0.0663).
By contrast, in Group II, only a decrease of borderline statistical significance in the 12-h amplitude of TF-component is observed from pre-flight to ISS01 (from 4339.3 to 1718.8 msec2, P = 0.0634), perhaps because all amplitudes pre-flight are higher in Group II than in Group I. Indeed, amplitudes (in msec2) in Group II vs. Group I are 5019.2 vs. 1842.6 (P = 0.0525; 24-h), 4339.3 vs. 801.6 (P = 0.0019; 12-h), 2621.2 vs. 1150.6 (P = 0.0589; 8-h), 2222.6 vs. 866.8 (P = 0.0133; 6-h), 533.4 vs. 322.5 (P = 0.0922; 3-h), and 502.8 vs. 196.1 (P = 0.0235; 90-min). During ISS01, differences between the two groups are no longer observed, perhaps due primarily to an amplification of the 12-h amplitude of TF-component in Group I.
Sleep duration (mean ± SD, in min) did not differ between Groups I and II (Pre-flight: 395 ± 86 vs. 410 ± 89, P = 0.7059; ISS01: 395 ± 38 vs. 412 ± 86, P = 0.5432; ISS02: 378 ± 50 vs. 426 ± 59, (P = 0.0685); Post-flight: 392 ± 84 vs. 398 ± 94, P = 0.8835). On average, the time of awakening (mean ± SD, in hr:mn) was also similar between the two groups (Pre-flight: 6:51 ± 1:10 vs. 6:25 ± 0:53, P = 0.3765; ISS01: 6:25 ± 0:48 vs. 6:20 ± 1:08, P = 0.8672; ISS02: 6:18 ± 1:06 vs. 6:18 ± 0:39, P = 0.9962; Post-flight: 7:08 ± 1:05 vs. 7:05 ± 0:57, P = 0.9165).
HRV endpoints of Groups I and II astronauts
A comparison of 24-h averaged HRV endpoints between Groups I and II is shown in Table 3. Pre-flight, SDNN (122.8 vs. 168.4 msec, P = 0.0126), SDANN5 (107.4 vs. 145.3 msec, P = 0.0204) and SDANN30 (101.6 vs. 137.0 msec, P = 0.0292) involving anti-aging effects, were lower in Group I than in Group II, even though all endpoints were within normal limits, as reported previously [38, 57, 58, 59]. Similarly, pNN50 (P = 0.0293) and TF- (P = 0.0082), ULF- (P = 0.0067) and VLF-components (P = 0.0199) were smaller in Group I than in Group II.
In Group I, statistically significant increases during ISS01 were found for SDNN (114.7%; from 122.8 to 140.9 msec, P = 0.0275), SDANN5 (114.2%; from 107.4 to 122.7 msec, P = 0.0493) and SDANN30 (118.9%; from 101.6 to 120.8 msec, P = 0.0201). This result suggests that the amplified 12-h rhythm of HRV may have promoted anti-aging effects in concert with the response in brain plasticity. Consequently, statistically significant differences between Groups I and II, seen during ISS01, are no longer observed during ISS02 since higher measures were maintained without changes in Group II (Table 3).
The fractal scaling of HRV (slope β) was statistically significantly less steep in space, with no changes observed between ISS01 and ISS02. Neither parasympathetic activity reflected by r-MSSD, pNN50 and HF-component, nor sympathetic activity estimated by the LF/HF ratio showed any definite differences between Groups I and II across the two records obtained in space.
12-h rhythm counteracted circadian desynchrony during long-duration spaceflight
Individual changes during spaceflight in the 24-h acrophase of HRV endpoints served to evaluate circadian desynchrony [60, 61, 62]. In Group I, the circadian acrophase of the TF-component was statistically significantly delayed by about 5 h (P = 0.0303) during ISS01 compared to pre-launch, and that of the ULF-component by about 6 h (P = 0.0055), Fig. 1 (upper left). The acrophases partly recovered during ISS02 but were still delayed by about 3 h (P = 0.0450) or 4 h (P = 0.0304), respectively (Table 4). In this group, circadian acrophases of the MF1-band did not differ statistically significantly between the two sessions recorded in space (Fig. 1, lower left).
By contrast, in Group II, during ISS01 compared to pre-launch, the circadian acrophase statistically significantly advanced by about 7.5 h in MF1-band (P = 0.0012) and by about 3 h in HF-component (P = 0.0095), r-MSSD (P = 0.0274) and pNN50 (P = 0.0318). Further advances were observed during ISS02, now averaging about 8 h (P = 0.0012) in MF1-band and 3.5 h in HF-component (P < 0.01), r-MSSD (P = 0.0221) and pNN50 (P = 0.0257) compared to pre-flight, Table 4. Then, the circadian acrophase of the LF-band and LF-component also advanced significantly by about 1.5 h (P = 0.0052) and 6.5 h (P = 0.0141), respectively (Table 4). The circadian phase advance in MF1-band of Group II is depicted in Fig. 1 (lower right). In this group, the ULF-component did not show any statistically significant phase shift between the two sessions recorded in space (Fig. 1, upper right).
12-h rhythm facilitates psychological resilience during daytime in space missions
Since neuroimaging studies showed that activity in low-frequency fluctuations corresponding to the DMN or SN was negatively associated with psychological resilience, the state of psychological resilience during daytime was assessed herein by changes in those HRV endpoints reflecting DMN or SN activity.
First, a comparison of Groups I and II during pre-launch shows that LF-, MF2- and HF01-bands and LFFs of Group I were statistically significantly lower than those of Group II, Table 5 (top). During ISS01 and ISS02, these HRV endpoints as well as HF01-, HF02-, HF03-, HF04-bands are also lower in Group I than in Group II. These differences are statistically significant in the case of MF2-band and HF01- and HF02-bands during ISS01 and of LF-, MF2-, HF01-, HF02-, HF03-, HF04-bands and LFFs during ISS02. The MF2-band in Group I remains lower than in Group II post-flight. These results suggest that Group I astronauts may have had higher psychological resilience than Group II astronauts.
Next, a comparison of ISS01 and ISS02 versus pre-flight in Group II shows that psychological resilience improved in space. During ISS01, LF- and HF03-bands and LFFs decreased to 71.8% (P = 0.0194), 79.6% (P = 0.0311) and 78.9% (P = 0.0378) of pre-flight values, respectively, and during ISS02, LF- and MF1-bands and LFFs decreased to 79.6% (P = 0.0322), 78.3% (P = 0.0325) and 79.1% (P = 0.0271) of pre-flight values, respectively (Table 5, top). In Group I also, MF1-band decreased to 90.1% (P = 0.0435) and 82.6% (P = 0.0365) of pre-flight values during ISS01 and ISS02, respectively, while during ISS01, MF2-, HF01- and HF03-bands decreased to 85.3% (P = 0.0737), 84.8% (P = 0.0834) and 67.8% (P = 0.0699) of pre-flight values, respectively.
Finally, psychological resilience, including life-satisfaction, was assessed in 3-h intervals from 06:00 to 21:00. LFFs improved mostly in the morning, Table 5 (bottom). During ISS02, LFFs in Group II decreased at 07:30 (i.e., between 06:00 and 09:00), 10:30 and 13:30 to 70.6% (P = 0.0433), 56.8% (P = 0.0387) and 74.8% (P = 0.0251) of pre-flight values, respectively, while during ISS01, LFFs decreased only at 07:30 (67.2%, P = 0.0198) and 13:30 (79.2%, P = 0.0497). In Group I, however, LFFs only decreased statistically significantly at 16:30 during ISS02. These observations suggest that Group II astronauts may have felt a sense of fulfillment during ISS01 and ISS02 and gained a great sense of satisfaction on days 159.2 (122 to 326) after launch (during ISS02), particularly in the morning.