Figure 1: a) Median sleep score and b) median sleep duration for 6 days after one night of bad sleep across three age groups. Here, 0 denotes the day of the bad sleep. The shaded regions show the standard error of the mean. c) Median sleep duration on Day 2 and Day 3 for the three age groups. Error bars denote the 95% CI. *, **, and *** denote p-values of <0.05, <0.01, and <0.001, respectively. n.s. denotes no statistical significance.
Recovery from a single night of poor sleep takes 3-4 days
We compared the overall metrics of sleep as provided by the Ultrahuman app to determine how the recovery of sleep profiles occurs across different age groups. We found that sleep scores improved appreciably on day 1 in all three age groups, and this night seems to have had the largest increase as compared to all days monitored. Sleep scores for the 40-50 and 50-60 age groups continued to trend upwards after day 1; surprisingly, the youngest age group’s sleep score plateaued on day 2 and even decreased slightly on day 3 (Figure 1a). This resulted in both the 40-50 and 50-60 age groups having substantially higher sleep scores on days 2 and 3 compared to the 20-30 age group. Despite the lower scores on days 2 and 3, the 20-30 age group’s sleep improved to a similar level as that of both the older groups by day 4 (Figure 1a). Our results also display that by day 4, all three groups’ sleep scores had crossed their initial average of 70.
Older users increase their sleep volume to a greater extent than younger users
Similar to sleep score, we found that sleep duration also increased in all three age groups post day 0. However, beyond day 1, the 40-50 and 50-60 age groups continued to increase their sleep volume until day 3, while the 20-30 age group’s sleep volume only increased after day 2 - and to a lesser extent (Figure 1b). The increase in sleep volume, observed in both groups of older users occurred around the same time as their increase in sleep disturbances (~day 3; Figure 2a). Our results also show that sleep duration was significantly higher in the 50-60 age group, compared to the 20-30 age group on day 2. Moreover, sleep duration was significantly higher in both the 40-50 and 50-60 age groups, compared to the 20-30 age group on day 3 after poor sleep (Figure 1c).
Figure 2: a) Median sleep movements for 6 days after one night of bad sleep, where 0 is the night of the bad sleep. The shaded region denotes SEM. b) Median sleep movements on days 2 and 3 for the three age groups. Error bars represent the 95% confidence interval. * denotes p-value of <0.05. n.s. denotes no statistical significance. c) Percentage change in median HRV for days 1, 2, and 3 relative to day 0 across three age groups. The shaded region represents the standard error of the mean.
Within-session awake episodes increase across all ages post-acute sleep deprivation
The Ultrahuman application defines sleep movements as small episodes of wakefulness within a larger sleep session. Our results show that sleep movements increased in all three age groups post day 0 (night of poor sleep), with the 40-50 and 50-60 age groups experiencing more profound increases on day 1. In contrast, sleep movements in the 20-30 age group increased more gradually, reaching a peak around day 2 and then staying fairly stable until day 5. On the other hand, the 40-50 and 50-60 age groups displayed two peaks in sleep movements during the week following poor sleep - highlighting more variable periods of wakefulness and sleep disruption (Figure 2a). We also found differences in sleep movements across age groups on day 3; users in the 20-30 group recorded significantly fewer disturbances than those in the 40-50 and the 50-60 age groups (Figure 2b).
Sleep HRV shows age-specific recovery patterns after poor sleep
Sleep HRV is a marker often used as a proxy for assessing nocturnal recovery and daily readiness.8 Higher HRV is associated with better recovery and sensitivity to autonomic nervous regulation. Our investigation revealed distinct age-related trends in HRV changes following a poor sleep session (Figure 2c). On day 1 post acute sleep deprivation, all age groups reported an increase in sleep HRV, with the 50-60 age group showing the most substantial rise (~12%), followed by the 20-30 age group (~7%), and the 40-50 age group (~5%). On day 2 we found that the 40-50 age group showed the highest increase (~10%), followed by the 20-30 age group (~7%), while the 50-60 age group experienced a notable decrease to about 2% above baseline. On day 3, we observed further changes: the 20-30 group maintained similar levels of increase as day 2, while the 40-50 group showed a decrease in relative change. Moreover, there was no relative change in HRV for the 50-60 group on day 3. No significant differences between age groups were found here.
Conclusions, Limitations and Future Directions
There is a lot of social dialogue around recovery after a single night’s sleep. There are single person experiential reports or social perceptions on practices to “cope” with such sleep deprivation. Our study provides one of the first insights into age-related differences in sleep recovery patterns after a single night of poor sleep. Contrary to conventional belief, we observed that older adults (40-60 years) demonstrated more consistent improvement in sleep scores and volume, while younger adults (20-30 years) underwent a more volatile recovery. We discovered that it generally takes 3-4 days to recover across all age groups. It is possible that older users engage in more active coping strategies by increasing their sleep volume more significantly based on previous experience. They, however, seem to have had more fragmented sleep in the first couple of days. This finding highlights the importance of considering sleep fragmentation as a critical metric in sleep quality assessment. In contrast, among younger ring users, who may have more work and family commitments, the sleep debt persisted and recovery was relatively slower. It is equally likely that relying on their younger physiology, the 20-30 age group continued to maintain their workloads, knowing that the strain could be met.
HRV recovery patterns following poor sleep also differed across age groups; it is important to note the variation we found, as shown by the shaded regions on the graph. No statistically significant differences were found between the groups, likely due to this high variability. Our findings provide support to reevaluate the previous assumptions of age-related sleep recovery and have important implications for personalised sleep recommendations and interventions.
This study's primary limitation is the lack of information about the underlying causes of poor sleep, which could have influenced the observed recovery patterns. The progressive decrease of HRV with age is also a biological limitation that might impact the interpretation of our age-specific HRV findings.9 Additionally, our focus on acute sleep deprivation may not capture the complexities of chronic sleep issues that may be present in a subset of these users. The use of wearable technology, while providing valuable real-world data, is usually associated with limitations and is not directly comparable to controlled clinical sleep studies. Finally, our sample, consisting of Ultrahuman Ring AIR users, may not be fully representative of the global population.
We hope that this preliminary investigation displays the merit of future research in exploring the forces driving these age-related differences in sleep recovery. The drivers could be physiological and psychological, as well as socio-cultural. Additionally, integrating data on daily activities, diet, and stress markers could offer a more comprehensive understanding of variables influencing sleep recovery. Further investigation into the relationship between sleep fragmentation and overall sleep quality across age groups, as well as the role of nocturnal HRV in recovery and stress mitigation, could enhance our understanding of sleep health and inform the development of more sophisticated wearable sleep-tracking technologies.
Reach out to support@ultrahuman.com for commercial queries and science@ultrahuman.com for scientific queries.
