The Primary Respiratory Mechanism

Cranial osteopathy and craniosacral therapy³ are variants of a health treatment regime that originated with Sutherland.⁴  Physicians (primarily osteopaths), physical therapists, occupational therapists, chiropractors, dentists, and others currently use forms of this method.  Although some practitioners make distinctions among various types of cranial therapy, we believe observations recorded here pertain to all.  Because Sutherland was an osteopath and because all variants known to us have their roots in his ideas, here we refer to them all as "cranial osteopathy."
    The biological model usually called upon to explain the various diagnostic and therapeutic ministrations performed by practitioners of cranial osteopathy has been given the name "craniosacral mechanism" or "primary respiratory mechanism" (PRM^4[p24]) and includes the following elements:^{3(pp10-11),4(pp51-53),5,6(pp165-166),7,8(p23)} 

• 1) inherent rhythmic motility of the brain and spinal cord,

• 2) rhythmic fluctuation of cerebrospinal fluid (CSF),

• 3) articular mobility of cranial bones,

• 4) mobility of intracranial and intraspinal dural membranes, and

• 5) mobility of the sacrum between the ilia.

    According to the model, intrinsic rhythmic movements of the brain (independent of respiratory and cardiovascular rhythms) cause rhythmic fluctuations of CSF and specific relational changes among dural membranes, cranial bones, and the sacrum.  Because several of these elements provide the biological/mechanistic underpinnings for the cranial rhythm - the focus of this review - we examine the primary respiratory mechanism in light of published, peer-reviewed research.

• Element 1: Inherent Rhythmic Motility of Brain and Spinal Cord
  Pressure pulses caused by respiratory and cardiac rhythms (transmitted to the cranium by the venous and arterial systems, respectively) produce minor movements of the brain.  It is also true that some glial cells possess small amounts of actin and myosin and are somewhat motile (as are many cells).  However, in spite of assertions to the contrary by proponents of the PRM,^{4(pp51-52),5,6(p165),7,8(pp23-24,34-35,42,86),9(pp4-5),10} the brain and spinal cord cannot be capable of intrinsically derived movement as organs (see also Becker¹¹) because neurons and glial cells lack the requisite microstructure (in particular, dense arrays of actin and myosin filaments).  Claims of "a subtle, slow, wormlike movement" ⁷ and "coiling and uncoiling of the cerebral hemispheres" ^6(p165) and "rhythmic expansion and contraction of the brain and spinal cord" ⁵ and "dilation and contraction of the [cerebral] ventricles" ^4(p52) are scientifically groundless.

• Element 2: Rhythmic Fluctuation of CSF
  Again, although respiratory and cardiovascular rhythms produce minor movements of CSF, it is not these fluctuations in which cranial practitioners take special interest.  Their focus is on a purportedly independent rhythm palpable throughout the body, produced by brain movement (element 1),^{4(p51-52),6(p165),7,8(p35)} by rhythmic variations in CSF production,^3(pp11‑12) by extracranial muscles,^11,12 as an amalgam of other physiological rhythms,^13,14,15 or by other factors.  Practitioners believe that these pulses within the cranium are translated, through movements of dural membranes (element 4) and bones of the cranium (element 3), to the surface of the head, where they can be palpated by properly trained individuals.  This palpable rhythm goes by several names, including "cranial rhythmic impulse" ¹⁶ and "craniosacral rhythm."  We are aware of  six published attempts to assess quantitatively the reliability with which practitioners could measure, through palpation, parameters relevant to cranial osteopathic diagnosis.  Five of these reports focused on frequency of the cranial rhythm and published interexaminer reliabilities of approximately zero; the sixth (Upledger, 1977)¹ is badly flawed.  Here we summarize all of these reports, reanalyze some of their data, and give Upledger (1977)¹ its own section.  

• Element 3: Articular Mobility of Cranial Bones
  Movement between the bases of the sphenoid and occipital bones "is an essential part of Sutherland's functioning model [the PRM]" ^3(p10) and many practitioners claim that movement is possible here throughout life.^{3(pp10,16),4(pp24,31),6(p165),7,9(p23),10,17}  However, research on large samples of fresh tissue (unembalmed) or living tissue (CT scans) has shown that these two bones undergo complete fusion at their bases between the ages of 12 and 19 (Table 1; see Becker¹¹ for another treatment of this issue).  Similarly, some practitioners claim that, even after ossification, "palpable deformation of the sphenobasilar junction can be appreciated even in the elderly" ¹⁰ and that adult humans show "intra-osseous bone flexibility throughout the cranial base." ²¹  This suggests the palpable deformability of the heavily mineralized matrix of solid bone.  These claims are so completely lacking in scientific support that they border on ridiculous.
  
  Likewise, although palpable "articular mobility [at vault sutures] . . . is . . . the basis for important diagnostic and therapeutic procedures" ^8(p32) (see also Sutherland^4[p23]), movement here also is impossible in most adults.  Examination of many hundreds of specimens has revealed that, by age 30 or so, most vault sutures also have begun to ossify.^23,24,25  It is difficult to imagine the basis for claims such as: "sutural obliteration does not appear to occur normally during the aging process" ^6(p165) or "motion . . . persists throughout life." ^8(p32)

• Elements 4 and 5: Mobility of Dural Membranes and the Sacrum
  Whereas elements 1through 3 of the primary respiratory mechanism easily are invalidated, the relationship of elements 4 and 5 to biological reality or to parameters of biomedical importance is more difficult to assess.  Although elements 4 and 5 also may be of dubious merit,^12,26,27 they require no further attention here because they cannot be playing the biomedical role ascribed to them by practitioners of cranial osteopathy if elements 1through3 are invalid.  Furthermore, others^28,29 have reported results inconsistent with one of the correlates of elements 4 and 5 - the contemporaneous mechanical linking of cranial and sacral movements.

      Although cranial osteopathy is widely practiced, there is little more scientific support now for its presumptive mechanism of action than when first presented more than six decades ago.^11,12,26,30  In fact, elements of the PRM are contingent on anatomical and physiological phenomena that, if proven real, would require reconsideration of some of what is now considered impossible in those disciplines (e.g., motility of neural organs and palpable deformability of solid bone).
      Ordinarily, manual diagnoses and treatments will be more meaningful and effective (respectively) if based on a working understanding of relevant anatomical relationships and physiological principles. Even then, only if relevant parameters are measured reliably can there be any assurance that diagnosis will lead to an appropriate treatment.  Setting aside the likelihood that the PRM itself is invalid, relatively little attention has been given to interexaminer reliability associated with cranial osteopathy and it is to this we now turn. 

The Cranial Rhythm

Except for Upledger (1977),¹ all published reports of interexaminer reliability associated with cranial osteopathy have focused on frequency of the cranial rhythm.  Therefore, we concentrate now on elements two and three of the primary respiratory mechanism: the rhythmic movement of cerebrospinal fluid and cranial bones.  For practitioners of cranial osteopathy, the cranial rhythm (CR) plays a central role in diagnosis and treatment.^{3(p243),5,6,7,8(pp86,107),13,17,31,32(p95)}  For example: "utilization and management of this fluctuation becomes of prime importance" ^8(p107) and "all of craniosacral therapy . . . is predicated on the therapist being able to monitor the patient's craniosacral rhythm.  Attempting to use these techniques without the sensitivity to monitor the craniosacral rhythm guarantees that the treatment will fail." ^32(p195)  It is presumably because the cranial rhythm is at the mechanistic center of cranial osteopathy and plays a central role in diagnosis and treatment that it has so frequently been chosen by practitioners for tests of interexaminer reliability.  Again, only if reliably measurable could such a parameter be related predictably to meaningful features of a patient's physiology and possibly useful as a biomedical tool.

• Estimates of Interexaminer Reliability for the Cranial Rhythm

The CR is believed to exhibit several features measurable through palpation, including frequency, amplitude, regularity, symmetry, and quality.  Of these, almost all quantitative estimates of interexaminer reliability have been for frequency.  We are aware of five blinded attempts to measure practitioners= ability to assess CR rate through palpation.^{28,29,33,34,35}  For each of these reports (Table 2), subject samples were small (ranging from 9 to 40).  Examiners were variously trained and their experience at palpating the CR ranged from periods of 11 months to at least 20 years.

Several coefficients have been used to assess interexaminer reliability for determination of CR frequency (Table 2, next-to-last column): intraclass correlation coefficients ICC(1,1) and ICC(2,1),³⁶ Pearson product-moment correlation coefficients, and coefficient alpha.³⁷  Instead of the published product-moment correlations (r), we report coefficients of determination (r² - the percentage of variance shared by two variables) because they are better measures of reliability³⁸ and will support more useful comparisons.  Of coefficients published so far, we believe that ICC(2,1) is most appropriate for comparative purposes³⁶ and have calculated this coefficient from raw data supplied by several authors who did not report it themselves (Table 2).  We used SYSTAT 9 ³⁹ to calculate product-moment correlations and mean squares required for computation of intraclass correlations.  The ICC(2,1) coefficients, themselves, were computed using the appropriate equation^36(p423) and a computer program written by us for that purpose (Microsoft FORTRAN, Version 5.0).

We can conceive of only two explanations for this finding.  Perceived cranial rhythms may be primarily an amalgam of each examiner's own cardiovascular, respiratory, or other rhythms.¹⁵  To date, there is little scientific support for this idea.^1,15,34,35  The only alternative we can conceive of is that the rhythm is the perceptual product of psychological phenomena manifested in the examiner.  It is not unusual for individuals to perceive an imaginary sensation only because they have been told to perceive it,⁴⁰ and "this is especially true when the stimulus is vague or ambiguous or when clear observation is difficult."^41(p37)  In particular, humans often perceive motion, that they themselves are the source of, as being externally produced (Aautomatisms");⁴² see especially Spitz⁴³ and numerous references therein), with "the anticipation of a given result being the stimulus which directly and involuntarily prompts the muscular movements that produce it." ⁴⁴  Perhaps a cranial practitioner's expectations result in minute, subconscious contractions in hand/arm muscles of the practitioner, leading to tactile sensations that appear to confirm PRM-related expectations of cranial movements and rhythms in a patient. 

Upledger (1977)

Upledger (1977)¹ evaluated interexaminer reliabilities and agreements for 19 palpatory diagnostic parameters used in his version of craniosacral therapy and compared his subjects' CR rates with the subjects' and examiners' cardiovascular and respiratory rates.  For each of his 25 subjects :

• a. the subject's and first examiner's cardiac and respiratory rates were measured,^{1(Table 7)} then

• b. the first examiner (one of four osteopaths) evaluated the subject's CR rate^{1(Table 7)} and the other 19 parameters (on a five-unit scale^{1[Appendix B]}), then

• c. the subject's cardiac and respiratory rates were measured again, as were those of a second examiner, then

• d. the second examiner re-evaluated the subject's CR rate and the other19 parameters.

     Upledger and one of three other osteopaths evaluated each subject.  This yielded 25 pairs of CR rates and 25 sets of paired measurements for each of the 19 other parameters.  These served as the basis for numerous estimates of agreement and reliability.^{1(Tables 2-5)}  This work has been published at least three times^{1,3(pp345-356),45} and is frequently cited.  This is the only published, detailed consideration of interexaminer reliability for diagnostic features (other than rate of the cranial rhythm) used in cranial osteopathy (or craniosacral therapy).

Although Upledger¹ did not publish interexaminer reliability for his measurements of CR rates, data in his Table 7 (column 4) have permitted us to do so.  The intraclass correlation coefficient ICC(2,1) is not strictly appropriate for data from all 25 of his subjects because all were not measured by the same examiners.³⁶  However, we provide it for comparative purposes and, using Upledger's data, we believe the estimates of CR interexaminer reliability likely to be biased least by small sample size are those computed over all subjects.  Over all 25 subjects and all 4 examiners, we calculated a coefficient of determination of .41 (probability that r = 0 < .001) and an ICC(2,1) of .59 (p < .001; Table 2, final column).  These are perhaps "too low to support [Upledger's] conclusion that the examination can be conducted with an acceptable degree of reliability," ³⁵ but they are statistically significantly greater than zero.  Together with his published global estimate of 86% interexaminer agreement (compared to a chance expectation of 52%) for the other 19 parameters (his Table 6), these estimates reflect substantially greater reproducibility of CR rate measurements than anyone else has achieved (Table 2). 

Perhaps Upledger¹ achieved relatively high CR rate reliability because his subjects were so young (ages 3-5 years).  In that case, if the cranial rhythm is real, there may have remained sufficient mobility among bones of his subjects' crania to permit reliable measurement of it.  Alternatively, Upledger and his colleagues may have been more skillful in perception of the CR than other examiners reported on here.  However, all 10 of Drengler and King's³³ examiners and all six of Norton's²⁸ examiners also were trained as osteopaths.  Likewise, most practitioners represented here had considerable experience in monitoring cranial rhythms (Table 2, column 5).  Finally, one of the highest reliabilitiesy published since Upledger's¹ report was achieved by two students in physical therapy,³⁴ each with less than a year's experience in palpating the CR (Table 2).

Such factors may help explain the wide disparity between the CR rate interexaminer reliability manifested by Upledger¹ and all reported since, but only if the PRM is a real biological phenomenon.  Given the paucity of data supporting this notion, we believe a more likely explanation for Upledger's apparent success is suggested by the fact that his study shows extraordinary evidence of carelessness and poor design. 

1. Although all other workers have reported tiny (or negative) CR rate interexaminer reliabilities, Upledger's were relatively high (e.g., r² = .41).  By contrast (Table 4), the first measures of his subjects' cardiac and respiratory rates were poorly correlated with respective second measures (r = .09 and .31).  One possible explanation for this finding is that actual measurement reliability was similar for all rates, but CR rate reliabilities appeared greater because CR rates were simply more stable over the interval between first and second measures.  Although some claim that the CR rate exhibits greater short-term temporal stability than cardiac and respiratory rates,^3(p6) we are aware of no published data that support this claim.  Another possibility is that all rates were equivalently stable over the interval between first and second measurements but CR rate interexaminer reliability was much higher.  However, it is difficult to imagine how a rhythm many cannot perceive and many others cannot detect without training (the CR) could be measured more reliably than cardiac and respiratory rhythms, which can be measured objectively and accurately by anyone who can count. 

2. Some cardiac and respiratory rate pairings presented in Upledger's Table 7 are difficult to explain.  For example, when first measured, subject 19 had a cardiac rate of 120/minute and a respiratory rate of 24/minute.  When measured again many minutes later (after evaluation of CR rate and the other 19 parameters), this subject showed a much lower cardiac rate (92/minute) and a respiratory rate that had almost doubled (40/minute).  Similarly, subject 5's cardiac rate jumped from 84 to 120/minute while respiratory rate actually dropped slightly from 30 to 28/minute.  These findings are, physiologically, very difficult to explain and suggest measurement error, transcription error, or some other form of carelessness.

3. Upledger^1(p891) indicated that each child's CR rate was recorded "as counted for one minute."  Elsewhere,⁴⁶ he reported that: " We did count cranial rhythmic impulses as well as heart and respiratory rate, but we only counted each for 15 seconds and multiplied by 4 to get the rates per minute."  Indeed, examination of rates reported in his Table 7 ¹ shows that many more are divisible by four than would be expected by chance, suggesting that many rates were extrapolated from 15-second counts.  However, many are not divisible by four, suggesting that rates were recorded over different intervals on different occasions.  Not only were perceived respiratory and CR frequencies too low for either to support accurate rate determinations using only 15-second counts, but this inconsistency could easily have affected results.

4. Upledger¹ showed little correspondence between his subjects' CR rates and cardiac and respiratory rates of his subjects and examiners and said this would "help establish the CRI as an independent physiologic rhythm."  However, as described earlier, these different rhythms were recorded over an unspecified time interval and none of the others was recorded simultaneously with CR rate.  Considering this and the fact that first and second measures of cardiac and respiratory rates for subjects do not even resemble one another in a physiologically sensible way (see items 1 and 2), it is hardly surprising that neither resembles CR rates measured twice over a similar interval.

5. We used raw data published in Upledger's Appendix B to verify measurement reliabilities (product-moment correlations; personal communication - JE Upledger, 1998) for Upledger's 19 diagnostic parameters (his Tables 2-5).  For Upledger paired with each of three other examiners (his Tables 2-4, columns 2) and for all four examiners and all 25 subjects combined (his Table 5, column 2), 26 of 76 (34%) of these reliabilities were misreported.  Six of the differences between his and our correlations may have resulted from rounding errors but others were off by a great deal more.  For example, for Dr. Upledger with Dr. Mitchell and parameters 13, 17, and 18, reported reliabilities were all zero (his Table 4, column 2).  However, we calculated correlations of .92, -.90, and -.32, respectively.

6. 6) Also in his Tables 2-5 (final columns), Upledger¹ published percentages of agreement for each of the 19 diagnostic parameters, allowing a difference of up to one unit on his 5-unit scale.  Again based on data from his Appendix B, 10 of the 76 values presented (13%) were miscalculated, being off by an average of about nine percent and about five percent too high overall.  (Although five of the eight averages presented in Upledger's Table 6 - each over all 19 parameters - are in error, none is off by more than two percentage points.)

    In his abstract, Upledger¹ said: "these data would seem to support the reliability . . . of the examination findings."  In his conclusion he said: "it is possible to achieve an acceptable degree of interexaminer reliability and percentage of agreement between examiners utilizing craniosacral examination methods and techniques . . . [this] lends considerable evidence to the existence of a real and perceptible craniosacral motion system."  However, evidence of poor experimental design and extraordinary lack of rigor call this central claim into question.  His measured reliabilities can be meaningful and his claims justified only if the experimental blind between paired examiners was maintained scrupulously during all assessments of the cranial rhythm; otherwise, findings of the first examiner in each pair may have influenced (consciously or unconsciously) findings of the second examiner.  With so much evidence for lack of scientific rigor, we cannot be confident that paired examiners were properly shielded from each other's findings.  We are not suggesting that Dr. Upledger intentionally mislead his readers.  However, before osteopaths (or others) can claim meaningful diagnostic and treatment control over the biomedical system labeled "cranial osteopathy" (or "craniosacral therapy"), it must be shown that biomedical parameters of this system can be measured reliably.  Practitioners should at least be able to count cycles of the CR.  After all, for both mechanistic explications and diagnostic and therapeutic judgements, practitioners rely on properties of this rhythm (e.g., amplitude, symmetry, and quality) that are even more derivative and subjective than its frequency.  To date, only Upledger¹ has submitted results at all suggestive of reliability and we here provide reason to doubt the value of even these findings. 

 Possible Points of Contention

Some cranial practitioners have asked how we justify such a critical review of cranial osteopathy, given the wide array of other poorly understood (including more conventional) treatment modalities.  Although much remains to be learned about mechanisms and effectiveness of many conventional treatments, the proportion of such treatments with these deficiencies is much overstated.⁴⁷   Most importantly, however, we are critical because the extraordinary biomedical implausibility of cranial osteopathy places it in a very different category than any standard, science-based method of health care known to us.

Some practitioners may disregard findings of poor interexaminer reliability for CR frequency summarized here because they are based on small samples of subjects and examiners.  However, reliabilities of essentially zero across all of these small studies (except Upledger¹) are ample support for our claims that: sample size is not at issue, the CR probably is not a biomedical feature of patients, and additional study using larger samples is unwarranted.  It now should be up to practitioners to find any important diagnostic or therapeutic parameter of cranial osteopathy that can be measured reliably.

Some practitioners have argued that low reliabilities of CR rate measurements taken one after another are not surprising, given that: " . . . it seldom happens that a therapist practicing CranioSacral Therapy can touch a patient for more than a minute or two without having some therapeutic effect on this very sensitive craniosacral system." ^{46,see also 21}  Not only is this ad hoc assertion without scientific support, but we consider it suspiciously convenient: if true, it would render reliability of all diagnostic claims related to cranial osteopathy completely untestable.  Also, this would leave unexplained the marked consistency of rate assessments across all subjects within individual examiners.  In general, whether interexaminer reliability of zero results from irregularity of the rhythm itself or an inability of consecutive examiners to measure a stable CR and reach similar diagnostic conclusions, the implication is the same: none of the features of the CR can be useful sources of diagnostic information.  Furthermore, others^28,29 examined comparability of CR rates measured simultaneously at two locations (head and sacrum, head and feet, respectively), and reported between‑location product‑moment correlations below zero.

Some practitioners have suggested that published reliabilities of essentially zero for measurement of CR rate are relatively unimportant, or even irrelevant, because rate for the cranial rhythm is less important clinically than its other qualities.  However, if phases of the phenomenon labeled "cranial rhythm" or "primary respiratory mechanism" cannot even be counted, then it is unlikely that its other, more derivative and complex features (e.g., amplitude, symmetry, and quality) can be evaluated reliably either.  We consider this to be the case, whatever the cause of the rhythm.  Also, given that this one presumed biomedical parameter of cranial osteopathy has been the nearly unanimous choice for reliability testing and has failed utterly, we are suspicious of practitioners who now claim that this parameter was a poor selection because of its minimal clinical value.

Perhaps in recognition of the weakness of the mechanistic framework of cranial osteopathy and reported interexaminer reliabilities of zero, some practitioners de‑emphasize the PRM in their defense of cranial osteopathy, instead focusing on perceived clinical efficacy. They assert that properly trained cranial practitioners must be diagnosing reliably because many decades of clinical experience has shown treatment to be effective.  We agree that, if cranial osteopathy were demonstrably efficacious, invalidity of mechanistic explanations and apparent lack of interexaminer reliability, although hard to explain, would be clinically moot.  Unfortunately, although many clinicians (and patients) have become convinced of the efficacy of cranial osteopathy, there are still no data, based on properly controlled research, supporting any claim that apparent symptom improvement following Acranial@ treatment has ever involved more than, at most, a form of placebo effect.  Recent summaries offer practitioners little cause for optimism.^48,49  Without careful scientific controls, weaknesses of perception and interpretation can fool both practitioners and patients into believing that a treatment is effective when it is not.^{41(pp195‑232),50,51,52,53,54, and work cited in these reviews}  We believe that these and other natural, human, psychosocial influences help to explain how cranial osteopathy has achieved the 21^st century without scientific support of any kind. 

Conclusion

There is little science in any aspect of cranial osteopathy:

• 1) there is no scientific support for major elements of the PRM;

• 2) the only publication purporting to show diagnostic reliability with sufficient detail to permit evaluation¹ is deeply flawed and stands alone against five other reports that show reliabilities of essentially zero; and

• 3) there is no scientific evidence of treatment efficacy. 

Until mechanistic claims associated with the PRM have been validated, until diagnostic reliability has been established, and until properly randomized and placebo‑controlled outcome studies have demonstrated symptom improvement following manipulation of relevant parameters, we believe cranial osteopathy should be excluded from required curricula of colleges of osteopathic medicine and from osteopathic licensing examinations. 

Acknowledgements

We thank Dr. King for sharing his CR rate data and we thank him and Dr. Hanten for sharing some details of their published work.  We thank Boyd Buser, Jane Carreiro, Mark Alain Dery, Mike Fillyaw, Viola Frymann, Judy Kimball, James Kneebone, George Pasquerello, Steve Shannon, Mary Smith, Ralph Thieme, Barbara Winterson, and many of our students for time they spent in valuable dialogue with us regarding some issues characterized here.  Many of those named here would not support some of our assertions and conclusions; if omissions or errors of fact or interpretation remain, we accept sole responsibility for these. 

*     *     *

While this manuscript was in press, another examination of interexaminer reliability for measurement of the cranial rhythm was published.⁵⁵  Again, four of eight published coefficients were negative and, again, data suggested these two osteopathic physicians may have been imagining the cranial rhythm.

References

• 1. Upledger JE.  The reproducibility of craniosacral examination findings: A statistical analysis. J Am Osteopath Assoc 1977;76:890-899.

• 2. Dodes JE.  The mysterious placebo [quotation from: Dubious dentistry: A dental continuing education course. Jarvis WT. Loma Linda, CA: Loma Linda University; 1990]. Skeptical Inquirer 1997;21(1):44-45.

• 3. Upledger JE,  Vredevoogd JD. Craniosacral therapy. Chicago: Eastland Press; 1983.

• 4. Sutherland WG. The cranial bowl. USA: Free Press Company; 1939.

• 5. Ettlinger H,  Gintis B. Cranial concepts. In: An osteopathic approach to diagnosis and treatment (2^nd Ed.). DiGiovanna EL, Schiowitz, S, eds. Philadelphia: Lippincott-Raven; 1997.

• 6. Greenman PE.  Principles of manual medicine (2^nd Ed.). Philadelphia: Williams & Wilkins; 1996.

• 7. Lay EM.  Cranial Field. In: Foundation for osteopathic medicine. Ward RC, ed. Baltimore: Williams & Wilkins;1997:901-913.

• 8. Magoun HI. Osteopathy in the cranial field (3^rd Ed.). Sutherland Cranial Teaching foundation; 1976.

• 9. Lippincott RC, Lippincott HA.  A manual of cranial technique. Ann Arbor, Michigan: Edwards Brothers, Inc; 1948.

• 10. Mitchell Jr FL. Clinical significance of cranial suture mobility, In: The Cranium and its sutures: Anatomy, physiology, clinical applications and annotated bibliography of research in the cranial field. Retzlaff EW, Mitchell Jr FL, eds. New York: Springer Verlag; 1987:13-26. 

• 11. Becker RF. Cranial therapy revisited. Osteopathic Annals 1977;5:316-334.

• 12. Ferguson A. Cranial osteopathy: A new perspective. Am Acad Osteopath J, 1991;[Winter]: 12-16.

• 13. McPartland JM, Mein EA. Entrainment and the cranial rhythm. Altern Ther 1997;3:40-45.

• 14. Nelson KE, Sergueef N, Lipinski CM, Chapman AR, Glonek T. Cranial rhythm related to the Traube-Hering-Mayer oscillation: Comparing laser-Doppler flowmetry and palpation. J Am Osteopath Assoc 2001; 101:163-173.

• 15. Norton JM. A tissue pressure model for palpatory perception of the cranial rhythm. J Am Osteopath Assoc 1991;91:975-994.

• 16. Woods JM, Woods RH. A physical finding related to psychiatric disorders. J Am Osteopath Assoc 1961;60:988-993.

• 17. Greenman PE, Mein EA,  Andary M.  Craniosacral manipulation. Man Med 1996;7:877-895.

• 18. Melsen B. Time and mode of closure of the spheno‑occipital synchondrosis determined on human autopsy material. Acta Anat 1972;83:112-118.

• 19. Madeline LA, Elster AD.  Suture closure in the human chondrocranium: CT assessment. Radiol 1995;196:747-756.

• 20. Okamoto K, Ito J, Tokiguchi S,  Furusawa T.  High-resolution CT findings in the development of the sph