, Germaine Cornelissen2 and Franz Halberg2
(1)
Department of Chronomics & Gerontology, Tokyo Women’s Medical University Medical Center East, Arakawa-ku, Tokyo, Japan
(2)
Halberg Chronobiology Center, University of Minnesota, Minneapolis, MN, USA
Abstract
Franz Halberg (1919–2013) developed chronobiology and founded the field of chronomedicine including chronomics, chronoastrobiology, and chronobioethics. He coined the term circadian, after documenting that biologic rhythms tip the scale between health and disease and even between life and death. His work is summarized in his over 3600 scientific publications, in cooperation with colleagues from around the world. Progress of the chronomedicine exactly depends on his history of research. Thus, herein we introduce him briefly.
Beyond blood pressure and heart rate, the biomedically interested scientist will find an all-too-brief introduction to what students of sociology, biology, and medicine regard as a new science from the molecular to the sociologic and the epidemiologic time structure of life. We show that the large extent of cardiovascular variations can be exploited in the form of dynamic and other endpoints derived for each individual, for use in preventive as well as curative health care.
Keywords
ChronobiologyCircadian rhythmChronobiologically interpreted ambulatory blood pressure monitoringVascular variability disordersLinear-nonlinear cosinor spectrumIn approaching any problem, stress and strain in particular, it is tempting to use a “norm” as an initial single value or a daily or yearly average and to refer to it as a “baseline.” Over 60 years ago, Halberg wrote [25], citing Zbigniew Z. Godlowski (J. Endocrinol 1952; 8: 102): “… it is not surprising to encounter reference to ‘the great variation in circulating eosinophil counts, even in normal conditions, which makes it impossible to establish a base-line’ for the study of the physiopathology of these cells. It is indeed impossible to establish a base-line, at least a straight base-line, for eosinophil counts in normal conditions, because there is none in normal conditions.” Nor are there baselines for other variables, except in the wishful thinking of too many investigators, including the best, such as Hans Selye, as we will demonstrate. As of March 2012, 63,740 publications referring to “circadian” attest to the ubiquity of these variations; there are many extracircadian changes as well, all constituting control information to anything we do along the scale of time. We appeal to all authors and editors to debunk “baselines” and to use, whenever available, historical and concomitant controls in science (and art, most broadly in all “humanities”).
An earlier paper by Halberg was labeled by the editor of the Journal of Gerontology A: Biological Sciences and Medical Sciences “Future history.” This title applies to this historical sketch; it ends as an implied, here explicit recommendation for the future, to account for a complementary system of cosmic cycles in and around us. This ever-present transdisciplinary spectrum will have to be considered in focus upon anything singled out as a partial system, as the study of time structure in living matter (chronobiology) investigated in biological data aligned with space and other “weather” series (by chronomics). The invariably present complementary system, a sphere of the mind (noös), Vernadsky’s noösphere, undergoes aeolian cosmic cycles mirrored in human affairs, as in living matter broadly, in the chronousphere (Gk chronos, time + Attic Gk nous, mind + Gk sphairos, sphere, globe), the nonstationary glocal (global and local) diversity in space and time of the universe which happens to be our home.
We qualify Pierre Charron (1541–1603), who wrote that “The true science and true study of man is man” (Traité de la Sagesse: Preface du premier Livre), and Alexander Pope (1688–1744), who wrote “The proper study of Mankind is Man” (An Essay on Man, Epistle II), adding, of course, the need for the backtracking of human periodicities over an archaeon to the cosmos. This sketch, originally written for a meeting on the history of chronobiology, was later extended in response to a request by Dr. Botond Buda for Folia anthropologica.
1.1 From Chronobiology to a Unified Science
Living matter is variable in time and space, as is weather. The more constant we try to make our proximal environment, the better we recognize seemingly spontaneous variability that reflects the changes around us, near and far. I describe a journey that started with counts of circulating blood cells and action potentials of our brain and the then-surprising periodicities even in RNA and DNA formation and led to the recognition of the drastic importance of timing, among very many other stimuli, exposure to noise first (and X-irradiation and drugs afterward) in the 1950s. What I had then dubbed “circadian stage” made the difference between ~70 or 80 % death and ~70 or 80 % survival from the same drug, ouabain, or an adrenocortical inhibitor, respectively, or other stimulus. Eventually, I learned that extracircadian rhythms also needed to be mapped; infradians can determine whether we die suddenly because of cardiac malfunction, by our own hand or by that of others. The cycles of the cosmos are found transdisciplinarily and are a challenge as we try to account for them beyond the molecule, at the levels of atom and quantum, with respect to the ever-present associations with space weather.
A perspective in space gains from new complementary imaging tools, chronobiology and chronomics leading (this is humanity’s challenge) to chronobioethics. In having fun toward that goal, we can use a set of linear-nonlinear cosinor methods that, whenever possible, should be applied glocally in time and space: nonstationarities are found not only in geo- (and cosmo-)graphy but also in time, in each case requiring the analyses of the longest available time series as a whole (globally) and in sections varied systematically in length (locally). The resulting atlas, when completed, will, we trust, serve a unified science, providing new preventive and therapeutic marker endpoints for variability anomalies among which coexisting multiple circadian and multiple about-7-day cycles are already mapped and await application for strain relief during wellness. We resonate with the solar flares about 5 months and the solar wind speeds over 1-year-long periods, as well as with decadals (about 10 years), didecadals (about 20 years), paratridecadals (about 35 years), and transsemicentennials (about 600 years) of space weather. These contribute to more than blackouts in and around us, including war and terrorism. The study of space weather is thus relevant to human affairs, including an effect on our mood, long anticipated by Alexander Leonidovich Chizhevsky and indirectly documented by Joseph Vallot, validated herein in “the language of shared frequencies” and by remove-and-replace approaches whose application has just begun.
1.2 Variability: Foe in 1948–1949, Friend Thereafter
Late in 1948, when I arrived at the Peter Bent Brigham Hospital of the Harvard Medical School in Boston, word was that the adrenocortical hormone cortisone, isolated by Edward C. Kendall of the Mayo Clinic (manufactured by the Merck Co.’s Lewis H. Sarett, the likely source of information) given by Philip S. Hench to patients with severe rheumatoid arthritis [1–3], had reproduced the effect of Lourdes: figuratively speaking, patients who came in wheelchairs left walking. This was more impressive than the wheelbarrows in William Withering’s (1749–1794) garden, with which patients with congestive heart failure were said to have brought their ascites and presumably left them after receiving Digitalis purpurea [4]. Cortisone was then scarce and a substitute with similar activity was highly desirable. Hence, at Harvard I was assigned to test different substances for cortisone-like activity. Variability was great, so that occasionally the count was spontaneously zero in one or at most two consecutive samples. I was given a few mice for testing. One solution, with tests on one or a few mice, was to make sure to bring the count to zero for 24 hours, something I did not find occurring spontaneously; I used a large dose of 2.8 mg of cortisone, the truck on the right of Fig. 1.1 [5] as a reference standard for eosinopenic activity, then a cautious, but wasteful, approach warranted only in view of great variability in count and the limited numbers of mice available for testing and even though the reference standard was scarce. The dose response curve on the left and in the middle of Fig. 1.1 shows that 1 μg of cortisone was effective in inducing a relative eosinopenia in the ascending stage of the circadian rhythm in eosinophil count in the mice investigated (when a spontaneous decrease in count was most unlikely to occur).
Fig. 1.1
Gain in sensitivity from taking rhythms into account. Circulating eosinophil cell-depressing activity was detected with 1 μg of cortisone when tested in the ascending phase of the circadian rhythm in the count of these blood cells [25]. The increase in sensitivity of the assay is depicted by comparing the minuscule dose on the left, taking into account the rhythm, with the dose depicted as a truckload on the right, when the rhythm was “eliminated” by depressing the cell count to zero for 24 hours by 2.8 mg of cortisone used as the standard in my original bioassay in Boston [5]. Using chronobiology, a 2800-fold decrease in dosage proved successful (© Halberg Chronobiology Center)
1.3 Epinephrine Test of Adrenocortical Function?
Others in the department of medicine at the Peter Bent Brigham Hospital tested human adrenocortical function: Recant, Hume, Forsham, and Thorn [6–8] had postulated, in keeping with C.N.H. Long [9], that epinephrine (from the adrenal medulla) acts upon the hypothalamus, which in turn, via corticotropin-releasing factor, stimulates the pituitary [7, 8] to produce adrenocorticotropin (ACTH), which stimulates the adrenal cortex to secrete corticosteroid which in turn suppresses circulating eosinophil counts. Accordingly a test was recommended for the clinic based on epinephrine-induced eosinopenia, thought to depend upon the presence of the adrenal cortex and which would not occur as a ≥50 % fall in the absence of cortical adrenal tissue. It seemed somewhat surprising that epinephrine from the adrenal medulla would have to act via the hypothalamus and the pituitary to stimulate its next-door neighbor, the adrenal cortex. Hence, I tested the effect of epinephrine in adrenalectomized and gonadectomized mice with ectopic cortical tissue removed, as best I could, along both sides of the spinal column, in the large ligaments of females and the scrotal fat of males. The eosinophil count decreased after my epinephrine injections by ≥50 % when I happened to test. I found eosinopenia in response to epinephrine in mice without adrenals. I reported my inability to confirm the basis of the epinephrine test in mice, presumably deprived of all adrenal cortical tissue (for epinephrine to act upon), to my department head, the late George W. Thorn, Hersey Professor of Medicine at Harvard. Thorn admired my “sticking to my guns,” as he put it, but added that all his staff members and distinguished fellows could not be wrong. In the case of the epinephrine test, however, his staff at the time was hardly right [10]: in 1952, William R. Best et al. reported on the “clinical and statistical analyses of 702 4-hour eosinophil response tests to corticotropin, ephedrine, epinephrine and placebos in 284 normal and miscellaneous medical subjects” and concluded that “Greater than 50 % drop of eosinophils has been noted following epinephrine or ephedrine in patients with pituitary tumors and in patients adrenalectomized and receiving small doses of cortisone. … Tests with these substances are therefore of little value in the diagnosis of adrenal, hypothalamic, or pituitary disease, and do not accurately assess the functional capacity of these organs at the time of examination.”
1.4 Vascular Variability Disorders (VVDs)
Very many still treat variability as a foe, or ignore it. Typical is a former head of cardiology at the University of Minnesota who was given a week’s half-hourly around-the-clock data on one of his patients. He certainly “honestly” added that he would gladly check the results (on a variability disorder, with his single measurement!). He is not alone. A family practice department head who monitored himself for a week, in whom we found altered variability, went to his care provider in Rochester, Minnesota, and we never saw him again. These instances are the rule. A respected friend, a former cardiology department head at the Mayo Clinic, who monitored himself repeatedly for Franz Halberg’s sake only, as he put it, could not be convinced of the merit of monitoring, even when his own data showed that it was helpful [11].
It took more than a generation for the profession to accept the use of the blood pressure cuff at home (not yet to remove it from the care provider’s office), a very great progress, re-advocated in the 1970s [12, 13] and implemented in the third millennium AD [14].1 It may take another generation to develop a chronobiologically interpreted ambulatory blood pressure and heart rate monitoring (C-ABPM) system (Fig. 1.2a–f [15]), which may allow even the mapping of infradian, e.g., paratridecadal cycles [16] (Fig. 1.3) and perhaps some new information thereby concerning the health not only of individuals but also of societies (Fig. 1.4) [17], and may even detect the antecedents of earthquakes (Fig. 1.5) [18]. If we can have sensors in the tires of our cars and computer chips that continuously monitor pressure over the life of the tire, we should be able to measure, as-one-goes, all rhythms that can form the basis of diagnostic and therapeutic measures. When known or assessed, such rhythms resolve effects of aging and are particularly indicated in view of the epidemic of noncommunicable diseases, referred to as a “slow-motion disaster” [19; cf. 20].
Fig. 1.2
(a) Hypertension and normotension at same clock-hour on different days or even “prehypertension” or “normotension” in 24-hour average on same day of the week in different weeks (© Halberg Chronobiology Center). (b) Top: the MESOR is usually more accurate than the arithmetic mean, as it is less biased from sampling at non-equidistant intervals. Middle: the MESOR is usually more precise than the arithmetic mean, as it tends to have a smaller standard error (SE) since the variability accounted for by the usually prominent rhythmic pattern does not enter the error term. Bottom: illustration of a cosine function identifying four parameters of the oscillation: MESOR, period, amplitude, and acrophase (© Halberg Chronobiology Center). (c) Illustrative parametric (left) and nonparametric (right) approach bracket a sphygmochron (middle) from a MESOR-normotensive man with CHAT, a first tentative diagnosis requiring additional monitoring. After data covering preferably at least 7 days of blood pressure (BP) and heart rate (HR) are downloaded from the e-mail into a computer for analysis, the following results are provided (since the 1990s and currently cost free from corne001@umn.edu) for the patient as well as for the care provider: (1) A list of actual measurements and the times at which they were obtained. (2) A plot of data as a function of time, shown together with the time-specified prediction intervals (PIs) of acceptability for systolic (S) and diastolic (D) BP and HR characteristics. (3) A data summary and a report of any BP and/or HR excess in consecutive 3-hour intervals. This part of the report may be accompanied by a “rhythmometric summary,” which is just a more technical form from which the information is derived to prepare the: (4) “Sphygmochron.” A sample “sphygmochron” (center) illustrates how results are being reported. First, above, the participant’s name is kept confidential; a codename is used instead. Gender and age are listed, along with the date and time at which monitoring started, and for how long data were collected. The numerical report consists of two parts labeled “Characteristics” (parametric results) and “Indices of Deviation” (nonparametric results). In each case, results are shown for SBP (when the heart contracts) on the left, DBP (when the heart relaxes) in the middle, and HR on the right. Under “parametric results,” a mathematical model of a smooth curve is fitted to the data to assess their circadian variation, which is primarily characterized by four numbers shown in the left-hand section of the graph, one of which, the period, covers with its uncertainty the precise 24 hours, so that the other three numbers are given from the fit of a 24-hour cosine curve. One characteristic, called the “MESOR,” is the average value around which values fluctuate. It is very similar to the mean value, but yields more reliable results when the data are not collected at precisely regular intervals, and has a smaller error when the data are equally spaced. Another characteristic, called the “double amplitude,” is a measure of the predictable change occurring within a day, from the overall low values found usually during sleep to the high values during the daily active span. The third characteristic, called the “acrophase,” is a measure of the time when overall high values are likely to recur each day. For each of the three characteristics (“parameters”), the participant’s value is compared to a range of acceptable values, derived from data provided by clinically healthy people of the same gender and age group as the participant. For instance, in the example shown here in the sphygmochron, all parameters are within the range of acceptable values (rectangles), except for the double amplitudes of SBP and DBP. Under “nonparametric results,” the participant’s data are compared by computer with time-specified reference values, also derived from chronobiologic archives on clinically healthy subjects matched by gender and age. For this analysis, all data are stacked over an idealized 24-hour day. Whenever a given person’s profile exceeds the limits of acceptability of peers, the data are marked as being excessive or deficient. The “percentage time of elevation” reports the relative incidence of excessive values during a 24-hour day. It is common to have occasional high values, but in the example herein, there is reason for concern. The next item, the time of excess, becomes useful when drug treatment should be timed prior to the peak in excess. Excessive values may either be barely above the limit or in turn can be very much higher than the limit. It is therefore important to express the extent of deviation by the “area under the curve,” that is, the area between the values when they exceed the limit and this limit itself. Empirically, it has been shown that excess up to about 50 (mmHg × hour during 24 hours) may still be acceptable and accountable for by daily worries and/or physical activities. In the case summarized, the HBI is 60, in bold, and if confirmed in the next 7/24 profile, a reason for treatment. On the top right, an abstract illustration of excess and deficit is accompanied below by two cases that are similar in terms of percent time elevation. They are very different in terms of hyperbaric index. In patient #2, although the percent time elevation is 9 % smaller than that in patient #1, the hyperbaric index is several times larger. The “timing of excess” can be used as a guide to time the administration of nondrug or, if need be, of drug treatment once there is blood pressure excess above 50 (mmHg × hour during 24 hours) and/or an elevation in MESOR, taking into account the chronopharmacokinetics of the drug prescribed. When, e.g., a tentative diagnosis of MESOR-normotension with CHAT is made, with insight into information provided on the questionnaire given to the participant with the monitor, as a first step, additional analyses may be carried out. Additional monitoring is recommended to check on any abnormality detected during the first monitoring, and if confirmed, the need for intervention is reported to the person monitored so that it can be reported to the health-care provider. In one case summarized elsewhere, the follow-up 7-day monitoring showed that CHAT persisted for both SBP and DBP, while the MESORs were again acceptable. Thus, the diagnosis of CHAT with MESOR-normotension was confirmed. Consultation with a health-care provider was strongly and urgently recommended. In two cases of CHAT without an elevation of the blood pressure MESOR, when such recommendations were ignored, catastrophic disease and high cost occurred, a myocardial infarction in a man or eclampsia in a pregnant woman with pressures of 115/67 mmHg (SBP/DBP), leading to the delivery of a very premature boy hospitalized for 26 months at a cost of $1 million US.(© Halberg Chronobiology Center). (d) Abstract limits for acceptable I. MESOR and II. amplitude (left, first 2 rows and upper curves in bottom two rows) for III. acrophase, IV. frequency (period), V. pulse pressure, and VI. heart rate variability, associated with blood pressure and heart rate surveillance and their corresponding vascular variability anomaly (VVA) descriptions. I. MESOR-hypertension (MH) can be systolic (S) (S-MH), diastolic (D) (D-MH), or both (SD-MH), or mean arterial (MA-MH), demonstrated parametrically (by cosine fitting). II. Circadian hyper-amplitude-tension (CHAT), which can also be systolic (S-CHAT), diastolic (D-CHAT), both (SD-CHAT) or mean arterial (MA-CHAT), etc. (by cosine fitting). III. SBP, DBP, or SDBP ecphasia (odd timing of the circadian rhythm of BP but not of that in HR) (by cosine fitting). IV. Ecfrequentia (altered period of the circadian rhythm) (by cosine fitting). V. Excessive pulse pressure (EPP), when the difference in the MESORs of SBP and DBP for adults exceeds 60 mmHg, a threshold that remains to be replaced by reference values from clinically healthy peers (eventually) with disease-free long-life outcomes specified further by gender, age, ethnicity, and geography. VI. A deficient HR variability (DHRV), defined as a standard deviation of HR less than 7.5 beats/min, a threshold that remains to be replaced by reference values from clinically healthy peers (eventually with disease-free long-life outcomes) specified further by gender, age, ethnicity, and geography (© Halberg Chronobiology Center). (e) Illustrative results supporting the need for continued surveillance and for a chronomic analysis of blood pressure series. I: Nocturnal hypertension: data stacked from 11 days of around-the-clock monitoring. Office spot checks cannot detect nocturnal pathology. II A: Among risk factors, an excessive circadian BP amplitude (A) raises the risk of ischemic stroke most. II B: Among risk factors, an excessive circadian BP-A raises the risk of nephropathy most. II C: An excessive circadian BP-A is a risk factor for ischemic stroke independent from the 24-hour mean (MESOR). III A: Individualized assessment (by CUSUM) of a patient’s initial response and subsequent failure to respond to autogenic training (AT) (EO, F, 59y). III B: Individualized BP chronotherapy. Lower circadian BP-2A and MESOR after switching treatment time from 08:30 (left) to 04:30 (right). III C – Control chart assesses individualized anti-MESOR-hypertensive chronotherapy. Chronomics detects nocturnal escape from hypotensive treatment taken in the morning (I above) and conditions such as CHAT, associated with a risk of stroke and nephropathy greater than hypertension (IIA, B), even in MESOR-normotension (IIC), and monitors transient and/or lasting success of treatment (IIIA–C). Merits are as follows: − Detection of abnormality during the night when the dose of medication taken in the morning may no longer be effective in certain patients, facts not seen during office visits in the afternoon but revealed as consistent abnormality by around-the-clock monitoring; − Detection of abnormal circadian pattern of blood pressure (CHAT, “overswinging”) associated with a risk of cerebral ischemia and nephropathy larger than other risks (including “hypertension”) assessed concomitantly (IIA, B); − Finding that CHAT carries a very high risk even among MESOR-normotensives who do not need antihypertensive medication (IIC); – Availability of statistical procedures such as a self-starting cumulative sum (CUSUM) applicable to the individual patient to determine whether an intervention such as autogenic training is effective and if so for how long it remains effective (IIIA); − N-of-1 designs for the optimization of treatment timing: the same dose of the same medication can further lower the same subject’s blood pressure MESOR and circadian amplitude when the timing of daily administration is changed (IIIB, C), as ascertained by as-one-goes (sequential) testing and parameter tests, procedures applicable to the given individual (© Halberg Chronobiology Center). (f) Benidipine (taken once a day upon awakening, Rx2) was found in large Asian clinical trials to be associated with better outcomes than nifedipine (taken twice a day, in the morning and in the evening, Rx1). Reducing the incidence of CHAT may be the reason accounting for the difference (almost by a factor 2) in outcomes, whether strokes or all cardiovascular events are considered (© Halberg Chronobiology Center). (g) Concept of an international multilingual website serving all comers worldwide, as a project on The Biosphere and the Cosmos, BIOCOS, now does on a small scale (left half of graph) with the data available to care providers (right bottom) and (after aligning with epidemiological data on natality, morbidity, and mortality and on crime and terrorism as well as with philanthropy and physical environmental data) for research on medical and broader problems with special reference to effects of space weather. The Phoenix Group of volunteering electrical and electronic engineers from the Twin Cities chapter of the Institute of Electrical and Electronics Engineers (http://www.phoenix.tcieee.org) is planning on developing an inexpensive, cuffless automatic monitor of blood pressure and on implementing the concept of a website (www.sphygmochron.org) for collection and analysis of data collected with these instruments (© Halberg Chronobiology Center)
Fig. 1.3
(a) Display of original climate data (top) and map of environmental-biospheric paratridecadal Brückner-Egeson-Lockyer (BEL) cycles (bottom). The BEL cycle originally found in Brückner’s data (first five point-and-interval estimates of the period) is also detected by spectral analysis of military, economic, and ecological data and in physiological data from one man (RBS) and from two other men (FH and WB), as well as in many other population phenomena, the mechanisms of which become available to study on individuals for investigators who try to plan beyond their life-spans. An about-33-year cycle was reported as a spectral peak by Shanahan who studied sediments in Lake Bosumtwi in Africa. The spectral coherence peak covers periods of 30–50 years (full length of uncertainty bar shown in graph) but includes a double peak, the second peak corresponding to another spectral peak at 42 years. The tick mark at 40 years corresponds to the small trough on the published coherence graph between the two peaks. Silverman (Rev Geophys 30: 333–51, 1992) was first to report an about-33-year spectral peak in aurorae (S in graph) but did not provide uncertainty estimates. The latter are derived for aurorae observed in Central Europe as compiled by Krivsky and Pejml during 1001–1900, analyzed globally and during different spans corresponding to varying average numbers of reported aurorae as technology to detect them improved. The span from 1545 to 1724 was of special interest as it allowed a comparison with an independent set of data reported by Schroeder and Treder (© Halberg Chronobiology Center). (b) Top: time series of weekly mean heart rate data for a clinically healthy individual, RBS (a), and its spectrum (b). Middle: Original time series of sunspot number (Wolf’s number, W) and superposed spectral components with periods (a) 32.82 (top left), (b) 10.56 (top right), (c) 8.02 (bottom left), and (d) its three-component model (bottom right curve). The spectral components are listed with 95 % confidence intervals in parentheses after the period values. Bottom: influence of solar activity on the human cardiovascular system: the congruence of natural and physiological cycles with periods from several years to several decades. (1) Cycle of change in the polarity of solar magnetic field (Hale’s cycle), BSC; (2) relative sunspot numbers (Wolf’s numbers), WN; (3) geomagnetic index aa, as determined from data of antipodal observatories in Greenwich and Melbourne, Gaa; (4) systolic arterial pressure, SBP; (5) diastolic arterial pressure, DBP; and (6) heart rate, HR. The horizontal length of black columns reflects 95 % confidence intervals for respective periods. Thin near-vertical linking lines and shading indicate congruent periods (© Halberg Chronobiology Center)
Fig. 1.4
Just as removal and replacement of a gland led to endocrinology, so biological consequences of the loss of environmental spectral components are critical to chronomics, the study of chronomes (time structures) in and around us. The drifting, bi- or trifurcation, disappearance and reappearance of a component with a given frequency and, when present, the waxing and waning of its amplitude, visualized in a and c, is in keeping with external driving of an individual (c) and of a population (a and b). Note components with neighboring frequencies waxing at or near a transyear component in solar wind speed (top of a and of c), geomagnetism (second row of a), and in 39 years of terrorist activity (bottom of a), an association supported by an independent method in b (by the statistical significance of the fit of a 1.3-year far-transyear to interplanetary [SWS] and terrestrial [aa] magnetism and to terrorism). Analysis in a is glocal insofar as it is based on the entire series in the spectral window on the right (global) and in sections of the series in the gliding windows on the left (local). Analyses only on sections of the time series are local in b (complementing the global counterpart, as seen on the right of a) and local in c. (c) shows time courses of the frequency structures of the speed of the solar wind (SWS) (top) and of an elderly man’s (FH) systolic and diastolic blood pressure and heart rate (rows 2–4, respectively), examined by gliding spectral windows. Human systolic blood pressure selectively resonates with solar wind speed (SWS) (top 2 sections). No obvious resonance, only minor coincidence of apparent change, is apparent to some in diastolic blood pressure or heart rate (bottom 2 sections). Aeolian rhythms in gliding spectra of SWS and SBP change in frequency (smoothly [A] or abruptly [B, C, D], bifurcating [D, F], and rejoining [G]); they also change in amplitude (B) (up to disappearing [C,E] and reappearing) (© Halberg Chronobiology Center)
Fig. 1.5
Biospheric contributions to the understanding, if not prediction of earthquakes. Upper left: Locomotor activity of some of the mice telemetered around the clock was statistically significantly decreased starting 3 days prior to the magnitude 8.0 earthquake in Chengdu, China, on 12 May 2008 (data of Zhengrong Wang). Upper right: Human systolic blood pressure started increasing 2 days prior to the magnitude 9.0 earthquake in Sendai, Japan, on 11 March 2011, documented on the basis of weeklong records of around-the-clock ambulatorily obtained data from 13 Japanese (data of Yoshihiko Watanabe). Similar records from longitudinal and transverse controls differ in their time course, suggesting that the trend observed before the earthquake was related to it rather than being a feature of an anticipated weekly pattern. Lower left: the monthly incidence of major earthquakes since 1900 is characterized by the presence of cycles with periods of about 49.3, 12.2, 1.44, and 0.41 year(s), given with their uncertainties in parentheses. Lower right: The prominent about-50-year cycle is also documented in physiology, pathology, societal upheavals, and space weather. Nonlinearly estimated periods are displayed with their 95 % confidence intervals shown as the length of corresponding horizontal bars (© Halberg Chronobiology Center)
In this context, an immediate reward can be anticipated and has been obtained on a small but worldwide scale by recognizing, for instance, vascular variability anomalies (VVAs) by the cosinor method (Fig. 1.2a–g) ([21]; cf. [12, 22, 23]. When these VVAs (Fig. 1.2d) [15] persist in several automatic 7-day around-the-clock records and become vascular variability disorders (VVDs) [15], treatment is indicated and is sometimes as simple as changing the schedule of hypotensive medication (Fig. 1.2e bottom left) [24]. We try not to repeat the mistakes of the past, such as the failure to scrub before surgery. Measuring and interpreting chronobiologically blood pressure series may seem cumbersome, like scrubbing for antisepsis. Nonetheless, in a computer era, self-surveillance could soon be implemented by everyone, monitoring continuously and affordably. VVAs would gauge loads and teach us how to avoid VVDs as a feature of universal preventive health care. The merit of work with vs. without rhythms is apparent quantitatively in Fig. 1.1, showing the dose reduction for corticosteroid-induced eosinopenia by an order of magnitude (from 2.8 mg to one or a few μg [5] vs. [25]).
Qualitatively opposite results obtained in the case of differences in circadian phase started chronobiology in Minnesota (Fig. 1.6a, b) to where the physiologist Maurice B. Visscher (whom I, as an assistant to the dean, had met in Innsbruck before coming to the USA, while he was lecturing as a member of a Unitarian medical mission) offered me a chance to move in 1949. Visscher also gave me the task to study adrenocortical function indirectly by eosinophil counts in two groups of mice, with a high and low breast cancer rate, respectively. I compared their counts without realizing at first that they were under the influence of different synchronizers, light for a fully fed group (feeding in the daily dark span, as rodents do when food is freely available) and the meal time for a group on a calorie-restricted diet ([26], cf. [27]) (Fig. 1.6) that happened to be given in the morning and, since limited in calories (from carbohydrate and fat only), was promptly consumed. These initially startling opposite results led us, by the early 1950s, to find the synchronization of a circadian rhythm by timing a calorie-restricted diet when the restriction was by 50 %, yet did not involve any changes in the intake of protein and vitamins and thus to find its dominance over the synchronization with that by the lighting regimen, found by us earlier [28].2
Fig. 1.6
(a) Importance of rhythms in assessing intervention effects, illustrated in relation to stress or allergy. A. Eosinophil counts seem to be lowered by fasting (and/or stress), when a 50 % reduction in dietary carbohydrates and fats (with proteins, vitamins, and minerals similar to control group) was fed in the morning to C3H mice (dark column). (In this model, the naturally high incidence of breast cancer is lowered by a diet reduced in calories and by ovariectomy, not shown.) The result could have been interpreted as an adrenocortical activation and then assessed by eosinophil depression, with applications for treating breast cancer and for prolonging life. Steroids that depress eosinophil cell counts and perhaps mitoses could be a mechanism through which caloric restriction and ovariectomy act in greatly reducing cancer incidence. This tempting inference was never published. B. In view of the importance of this finding for the etiology of cancer, results were replicated on a larger group of animals; 1 week later, a follow-up study with more animals started at an earlier clock-hour yielded confusing results, showing no statistically significant difference between the two groups of mice. C. After another week, another study starting at an even earlier clock-hour yielded results opposite to those in the first experiment when considered alone. These findings in C in themselves could have been interpreted as an allergic response, certainly contrary to the “stress” response in A. D. Sampling at intervals of a few hours in the third study, the stages called 4 and 5, hinted at the reason for the confusion: by sampling at different clock-hours, two groups of mice were found to be characterized by a circadian rhythm with different phases. Opposite effects thus became predictable. E. Abstract illustration of two circadian rhythms in antiphase. Differences in opposite direction or no effect are then anticipated from sampling at different clock-hours (© Halberg Chronobiology Center). (b) Effect of food restriction on circulating eosinophil counts in mice. Follow-up study on Fig. 1.6a, with a phase difference greatly reduced by offering the restricted diet in the evening. A. Even after log10 transformation of the data expressed as percentage of mean, great interindividual variability is apparent in the raw data. B. Plots of timepoint mean for mice in each group reveal different circadian patterns. C–E. Parameter tests quantify differences, indicating that calorie restriction is associated with a lower MESOR (C–D), a larger circadian amplitude (D–E), and only a slight difference in acrophase (D). The difference in acrophase in this study, where calorie-restricted mice were fed in the evening, is much smaller than the almost-antiphase observed in prior studies (Fig. 1.6a), where calorie-restricted mice were fed in the evening (© Halberg Chronobiology Center)
This was the reason for an intergroup difference in circadian acrophase. Its major generalizable result was methodological; it illustrated the possibility of false intergroup differences in spot checks that resulted largely from an intergroup difference in phase of the very many variables that exhibit circadian and many other rhythms. The difference in circadian amplitude revealed the real intergroup difference that spot checks could not detect and thus led to the also generalizable inference that the characteristics of rhythms (cf. Fig. 1.6) are the endpoints that must replace spot checks. Similarly, differences in circadian timing can play havoc with aging research [29], as well as with cancer research [30] restricted to sampling daily at a fixed clock-hour rather than assessing rhythm characteristics. The evidence in Fig. 1.6 was my first hint that a medical science based on time- and rhythm-unqualified spot checks as a whole must eventually be replaced by one based on time series for continuous surveillance, a view validated by the ubiquity of circadian rhythms ([26, 31]; cf. [32]).
By 1953, we also learned that the circadian pattern of convulsions (recorded for prior decades in the same patient!) (Fig. 1.7 left) could be shifted in humans [33, 34] by a change in the timing of sleep and wakefulness (Fig. 1.7 right), a finding for which we subsequently encountered an animal model [35, 36] which revealed an increase in vulnerability during phase-shifting [36] (Fig. 1.8 top left), perhaps because we could also show that different variables shifted at different speeds and hence were transiently desynchronized among each other [36–40].
