The main circadian pacemaker responsible for maintaining the sleep/wake cycle in mammals is located in two bilaterally symmetrical suprachiasmatic nuclei (SCN) 1. Each SCN comprises its own oscillator 23, but the two structures are thought to work in tandem to generate a single rhythm. However, when both sides of the body are monitored simultaneously (such as with wrist actigraphy), the rhythms for each side vary slightly 2 and the activity of the dominant hand has been shown to peak before the non-dominant hand 24. There are also sexually dimorphic activity patterns, such that females reach acrophase earlier in the day than males 5 and female chronotype is considered to be morning-type while males are deemed evening-types 67. Cerebral dominance may be readily assessed, but how patterns of differences in anatomical and functional asymmetries are correlated to cerebral dominance are not as straightforward 8.

While sex, hand and handedness all appear to influence circadian regulation of activity patterns, their relative contributions remain unknown. The primary goal of the present study was to advance our understanding of the mechanisms which regulate bilateral circadian rhythms by investigating the basis for sex differences in activity patterns. More specifically, this study addressed the role of gonadal hormones assayed indirectly by analysis of digit ratios in the regulation of circadian activity, assessed with wrist activity monitors.

It is well documented that circulating hormones can affect circadian rhythms; for example, estrogen has been shown to shorten circadian period and advance phase in hamsters 9. Gonadal hormones can also modulate circadian activity, as in mice 10 and Octodon degus 11. There is direct evidence of sex hormones influencing SCN function: estrogen and progesterone receptors are present in the human SCN but females have more estrogen receptors in the SCN than males 12. Androgen receptors have been localized in the core of the mouse SCN and have greater expression in males than females 10. Furthermore, gonadectomy alters male mouse circadian activity, which can be restored with testosterone treatment 10. Thus, the sexually dimorphic combinations of steroid receptors in the SCN could account for sexually dimorphic patterns of circadian activity.

Although circulating hormones can directly affect circadian activity, the role of prenatal hormone exposure on the development of sexually dimorphic activity patterns has not been fully investigated. One non-invasive approach to assessing prenatal exposure to hormones is the use of digit ratios, a measure of the relative lengths of the second digit (most often a finger) to the fourth (2D:4D) which is generally lower in males than females 1314. The sex difference is subtle with overlapping ranges for males and females 1315, and the difference is more pronounced on the right hand than the left 151617. Increasingly in the past decade, a number of researchers have explored the relationship of digit ratio, a sexually dimorphic anatomical trait, with other morphological, physiological, and behavioral traits 181920212223.

Evidence from multiple lines of research indicates that digit ratio is correlated with prenatal androgen exposure: for example, (i) the sexual dimorphism in digit ratio is seen by the age of two and is relatively stable through puberty, especially in the right hand 1424; (ii) 2D:4D is sexually dimorphic in children from a wide ethnic range 25; (iii) children with congenital adrenal hyperplasia (CAH), who are exposed to high androgen prenatally, have lower 2D:4D (are more masculine) than healthy controls 26; (iv) low 2D:4D is related to a polymorphism in the androgen receptor that increases the sensitivity to testosterone 27; (v) females from opposite sex twins have lower 2D:4D than females from same sex twins 28; (vi) children over age one produce very low levels of sex hormones (until puberty), yet the sex difference in digit ratio is apparent and constant at young ages, suggesting 2D:4D is established during prenatal hormone exposure in utero 1425. In fact, the use of digit ratio as a putative marker for prenatal androgen exposure has become routine 182930 since direct measurement of prenatal hormone levels is not feasible in humans 13.

Despite the fact that there are similar patterns of sex and hand differences in digit ratio and circadian wrist activity, the relationship between the two components has never been examined. This study compares the two measures in male and female college students. It is reasonable to hypothesize that prenatal androgen exposure may lead to sex differences in circadian rhythms because: (i) many sexual dimorphisms such as personality traits, sexual orientation, and disease predisposition are linked to prenatal androgen exposure 131822 and (ii) it is already known that the same hormones circulating in adulthood significantly alter circadian activity. If the contribution of prenatal androgen exposure to the development of digit ratio also affects SCN development, we predicted that patterns of circadian activity would be correlated with 2D:4D. In order to eliminate brain lateralization as a variable, only right-handed subjects were included in the study.

We predicted that patterns of circadian activity would be correlated with 2D:4D if the prenatal androgen exposure that contributes to the development of digit ratio also affects SCN development. Results showed that: 1) 2D:4D was lower for males than for females, with greater differences in the right hand; 2) while total amount of bilateral circadian activity was the same for males and females, patterns of activity were sexually dimorphic such that female activity peaked earlier in the day and females were more morning-type; and 3) there was a robust lack of correlation between digit ratio and any measure of circadian activity.

Previous findings for sex and hand differences in 2D:4D were replicated in this study. Mean values for 2D:4D (0.958 ± 0.004 for males and 0.972 ± 0.004 for females) are in the range of those reported for adults 1315. Ethnicity and sexual orientation, factors known to correlate with digit ratio 131621252937, were not accounted for in this study. However, as McFadden and Shubel suggest 17, the effect of ethnicity may be on the value of the digit ratio and may not affect the sex difference, which appears to be relatively stable. The lack of significant sex differences in left 2D:4D in this study parallels previous reports and supports a case for right-sided sexual dimorphisms 172730.

Results demonstrating sex differences in phase of circadian activity 5 and activity preferences 67 replicate findings reported elsewhere. Additionally, earlier acrophase for females was correlated with preferences for earlier activity (based on ME survey data) while later acrophase for males corresponded with preferences for activity later in the day, indicating that behavior was strongly correlated with stated preferences and presenting the possibility that ME could accurately substitute for Actiwatch data in certain situations. Of note, only the right acrophase generated a correlation with female ME, while ME was correlated with left and right acrophase in males. Perhaps the ME survey is more sensitive to day/night differences in activity, making it less able to capture variation in female acrophase which occurs within daylight hours. These patterns of relationships warrant further investigation.

Our sample did not compare left- and right-handers but we did find that for combined subjects, activity peaked first in the dominant hand of right-handed individuals and that the right hand was used more; this provides corroboration of work by Natale 2. Thus overall, the activity data reveal effects for both sex and hand.

The simultaneous action of posterior HOXA and HOXD genes in the development of the genital bud and limbs has been used to explain the link between androgen and the establishment of digit ratio 3839. Androgen is also present during the period of prenatal brain organization 2940 and if early androgen exposure contributes to the regulation of bilateral circadian activity later in life, digit ratio should be correlated with circadian activity rhythms. However, we found no evidence to support such a relationship; for example, low 2D:4D was not associated with increased mean activity as might be predicted from prior associations of low digit ratio with increased aggression and sensation seeking 181920. 2D:4D was not correlated with acrophase (Figure 2), degree of right-handedness, or ME preferences. In fact, no measure of digit ratio was associated with any measure of circadian activity. Thus, we conclude that the prenatal hormone exposure which results in sex and hand differences in digit ratio, is independent from regulatory mechanisms that yield sex and hand differences in circadian wrist activity during adulthood. It is possible, however, that any effect of prenatal androgen is masked or countered by later hormone exposure. Future research could explore the relationship between digit ratio and circadian activity in pre-pubescent individuals.

Sexual dimorphisms in circadian activity have been shown in the diurnal rodent Octodon degus and are dependent upon postnatal/pubertal exposure to gonadal hormones 11. Sex differences in phase and period of degu activity appeared following a delay from puberty, so Hummer and co-workers 41 concluded that circulating hormones are not responsible for the differences in activity, but that gonadal hormones modulate the circadian system, perhaps through androgen receptors in the SCN. Findings from human studies also suggest that sexually dimorphic activity patterns occur post-pubertally and with some delay 5.

Our results revealed that both sex and age had main effects on phase of activity (Figure 3A and 3B) with female and older individuals active earlier in the day. Thus, the relationship of the age-sex cohorts is in the order of old females<young females<old males<young males. The same main effects of sex and age were present for ME preferences (Figure 4).

One possible interpretation of the results is that the phase delay in males (relative to females) represents the lag in puberty for males vs. females, and that the sex differences in phase delay reflect the timing (delay) of puberty for males. It must be noted that the cohorts have significant sex differences in age (p = 0.0004) with females being younger than males. However, since young individuals are more male-like (phase delayed), the fact that females are younger than males increased the difficulty of demonstrating sex differences. Based on our findings, future circadian research using college students should consider age as an important variable since significant differences were found within this narrow age group.

Recently, the relationship between digit ratio and digit length has been investigated as a means of comparing effects of prenatal vs. pubertal androgen: while the ratio of the digits (especially the second to fourth) has become routine as a putative marker for prenatal exposure 1318293142, the length of the digits has only recently been used as a proxy for pubertal androgen exposure 30343536. The rationale supplied is that sex differences in finger length appear following puberty due to aromatase actions on androgen in the tissue. One study found modest associations of personality traits with digit length and fewer with 2D:4D 35. A second study used correlations of 2D:4D and digit length to predict handedness and suggested that release of similar levels of androgen prenatally and pubertally (e.g. low 2D:4D and long digits, corresponding to two rounds of high levels of androgen) could predict right-handedness 30. Most recently, Bescos et al. 34 explored the relationship of finger length with sport performance in women and Voracek 36 found "spurious" associations of superstitious beliefs and digit length in women.

Subsequent analyses of our data were conducted to explore a possible relationship of digit length with the regulation of bilateral circadian activity. For purposes of comparison with published results 30, we averaged lengths of all four digits (D), and also averaged 2D and 4D across hands. Analysis of 2D and 4D showed predicted sex differences: males had significantly longer digits. However, digit length was not correlated with circadian activity, ME, or degree of handedness. Thus, there is no evidence that the putative markers for prenatal or pubertal androgens are associated with circadian activity and neither marker provides an explanation for sexually dimorphic patterns of activity.

In summary, digit ratio was lower for males than for females with greater differences in the right hand than the left. Phase of activity was found to be sexually dimorphic and age-dependent. ME preferences correlated with activity for males and females and was also age-dependent. The lack of correlation of circadian activity rhythms with either digit ratio or digit length suggests that the sexually dimorphic regulation of circadian activity is independent from the mechanisms of hormone exposure that affect digit growth. Future research should continue to explore the basis for post-pubertal sexual dimorphisms in phase of activity, especially the relative contributions of prenatal, pubertal and circulating gonadal hormones.

The authors declare that they have no competing interests.

DBM proposed and designed the study. CR collected the data. CR and DBM analyzed the data, wrote drafts and revised the manuscript. All authors read and approved the final manuscript.