Validation of subjective manual palpation using objective physiological recordings of the cranial rhythmic impulse during osteopathic manipulative intervention

doi:10.21203/rs.3.rs-1581927/v1

Objective

Reticular brain activity described by intermediate (IM) band physiology exhibits parallels to controversial concepts of cranial osteopathy (CO), the primary respiratory mechanism or cranial rhythmic impulse (CRI). Interestingly, evidence for PRM/CRI activity has been reported using manual palpation, yet with questionable validity. Validation is tested combining manual palpation, instrumented tracking, and algorithmic objectivation of frequencies and phases.

Methods

Using a standard CO intervention, Vault Hold (VH), two CO experts palpated and digitally marked CRI frequencies in 25 healthy adults. IM band autonomic nervous system activity was examined as momentary frequency of highest amplitude (MFHA) in examiner and participant forehead photoplethysmography (PPG). Palpation errors and frequency expectation bias during VH were examined using MFHA and CRI phases.

Findings

Palpated CRI frequencies matched those reported in the literature. Mean MFHA and CRI frequencies were highly correlated and showed constant low deviating 1:1 coupling in 77% of participants (responders), and high deviating 1:2 coupling in 23% of participants (non-responders). Both groups showed harmonic waves in (very) low and IM bands in >98% of palpated intervals. Phase analyses suggested synchronization between MFHA and CRI in responders but not in non-responders confirming that examiners palpated CRI activity in participants.

Conclusions

Decreases in forehead PPG of responders from IM activity (~0.14 Hz) to LF activity (0.072 Hz) suggest successful palpation may be explained by IM and LF activity, whereas CRI in non-responders was not associated with PPG activity.

osteopathic manipulative therapies

cranial osteopathy

primary respiratory mechanism

palpation

cranial rhythmic impulse

autonomic nervous system

reticular brain activity

low frequency band

high frequency band

0.15 Hz rhythm band

intermediate rhythm

phases

amplitudes

Osteopathic manipulative treatment (OMT) focuses on enhancing the body’s self-regulation capacity by influencing autonomic nervous system (ANS) activity. Increasing evidence of heart rate variability (HRV) research in healthy non-symptomatic participants suggests OMT may have a positive effect by modulating the parasympathetic branch (1–6). However, OMT in the cranial field, cranial osteopathy (CO), further rests on the controversial concept of the cranial rhythmic impulse (CRI) as rhythmic movements on the head referred to as primary respiratory rhythm (PRM) in the body. Said to oscillate with inspiratory (flexion) and expiratory (extension) phases, modulating the PRM/CRI is instrumental for the diagnosis and therapy of somatic dysfunctions. Both terms relate to rhythmic physiological phenomena associated with autonomic nervous system (ANS) activity, which previous studies on CRI/PRM dynamics have identified as Traube-Hering (Mayer) waves centered at approx. 0.1 Hz (7–11). To this day, however, CRI/PRM are still conundrums with open questions. Technical approaches investigating the physiology underlying the CRI have yielded promising yet inconclusive results due to their missing link with palpation (6). Furthermore, appearance of both CRI/PRM and concomitant TH waves in blood pressure, heart rate, respiration, and skin blood flow, has been considered the physiological correlate for the palpation of the CRI/PRM within all parts of the body. However, palpated CRI rates differed frequently by a factor 2 from rates recorded using technical instrumentation (10).

Hitherto, ANS activity has been widely accepted to precipitate in a low frequency (LF, 0.05–0.15 Hz) and a high frequency (HF, 0.15–0.4 Hz) band. This notion and that of TH waves accounting for the CRI/PRM have recently been challenged (13). Pelz suggested considering the CRI/PRM as manifestation of a ‘0.15 Hz rhythm band’, which has been shown to emerge in various peripheral systems in humans during hypnoid relaxation (14–17). These authors had hypothesized that the emergence of the ‘0.15 Hz rhythm band’ in humans would reflect findings in canine unspecific neurons in the reticular formation of the lower brainstem, therefore termed ‘retR’ (16, 18). These terms have been more recently supplanted by referencing it as intermediate (IM) frequency band for it is wedged between LF and HF bands at 0.12–0.18 Hz. Recent functional magnetic resonance imaging (fMRI) studies suggest that IM band fluctuations in humans may also originate in the brainstem (19). Furthermore, IM band neurobiology demonstrated in heart rate variability (HRV) and respiration was associated with blood oxygen level-dependent activity in an interoceptive processing network (20). This is also in line with considering interoception as key neurological concept in osteopathy (21).

The broad frequency range of the IM band (0.15 Hz ± 0.03 Hz), the formation of beat oscillations, and integer number phase-synchronization with respiration, HRV, and blood pressure (16) are in support of explaining the origin of the CRI/PRM. Moreover, lower and upper harmonics of IM primary frequencies have been described for IM band activity (17), which may account for regularly reported variations in palpated CRI rates as a persistently puzzling phenomenon of the CRI/PRM (10). The possibility of lower harmonics as a source of rhythmic activity has previously been discussed by osteopathic researchers (11).

Results of our study group further supported the hypothesis set forth by Pelz (13). We found that a standard CO treatment, Vault Hold (VH), triggered an immediate decline from activity at the border of LF and IM band activity at 0.12 Hz during control conditions (eyes open, eyes closed) to stable activity in the LF band at 0.07 Hz in forehead skin blood oscillations, which represents lower harmonics of IM band activity. However, a smaller subset of participants remained at IM band frequencies.

Open questions concern the validation of palpated events with ANS related rhythmic activity in skin blood flow waves. A previous study demonstrated a wide lack of interrater reliability, which put the validity of OMT further into question (22). Also, the observation that CRI rates palpated by unexperienced examiners were found to be consistently above those palpated by more experienced examiners needs further examination (23). Yet another question concerns the current OMT model used to describe palpated events as flexion and extension events. Flexion and extension account for one complete CRI cycle. However, one CRI cycle does not consistently amount to complete rates (1:1 ratio) attained by instrumented recordings (10) for higher palpated rates at a 2:1 ratio has also been reported (24). To resolve these questions concerning the validity of the CRI/PRM concept of CO, we designed a hardware-software approach, which combined a Bluetooth-equipped footswitch to mark palpated rhythmic activity with semi-linear high temporal resolution algorithms for the analyses of frequency, amplitudes, and phases. This approach allows testing the validity of ANS physiological activity and concomitant palpated CRI rhythmic activity during a specific CO intervention (Vault Hold, VH).

Employing these techniques, we tested the following hypotheses:

(1) time intervals palpated at the skull match CRI rates reported in the literature for experienced examiners;

(2) palpated CRI rates are equivalent to autonomic nervous system (ANS) rhythmic activity in photoplethysmographic forehead skin recordings in frequencies of IM-band physiology (0.12–0.18 Hz) and their lower harmonics (0.06–0.09 Hz);

(3) continuous wavelet transform (CWT) detects primary IM-frequencies and harmonics of lower amplitude responsible for rhythmic activity that confounds palpation results;

(4) phasic footswitch responses contain information on palpated activity which is reflected also by dynamics of momentary frequency of highest amplitude (MFHA);

(5) phases of momentary frequencies of highest amplitude correlate with manually palpated extension and flexion.

2.1. Participants

A convenience sample of twenty-five healthy adults (14 female, age 42.4 (± 13.1) years) was recruited by the investigators through word of mouth. Five participants completed two measurements at an interval of 1 week, resulting in a total of 31 measurements. One measurement had to be excluded as palpation records were faulty. All participants were non-smokers, none had any OMT in the preceding three months. Exclusion criteria were mental health symptomatology, a history of or acute neurological or cardiovascular disorders, current use of psychoactive medication as well as high-performance sport. Furthermore, participants were asked to abstain from caffeine for 4 hours and alcohol consumption for 48 hours prior to testing. The experiment was conducted according to the Code of Ethics of the World Medical Association (Declaration of Helsinki, 2008). All participants provided written, informed consent, and all protocols were approved by the Institutional Review Board of the state of lower-saxony/Germany (EK vote from 04/03/2017).

2.2. Examiners

Two DGOM-certified osteopaths participated in the study as examiners. At the time of the study, examiner A had been in private practice for 34 years and was also teaching cranial osteopathy. Examiner A estimated that 60% of patients under his care would undergo at least some cranial treatment, whereas cranial treatment would be the major treatment regimen for approximately 50% of patients. Examiner B had been in private practice for 25 years and was also teaching cranial osteopathy. Examiner B estimated that approximately 80% of patients under his care would receive at least some cranial treatment whereas this would be the major treatment for approximately 60% of patients.

2.3. Study Design

The study used a double-blind design with examiners as well as participants being naïve to data recording and offline data analysis. All participants were allocated to Vault Hold intervention (VHI) employed to palpate CRI activity. Experimental sessions were performed at comfortable room temperature. They consisted of five consecutive sections (300/330 seconds each; 30 seconds were added after the first measurements to account for artifacts due to section transitions, which were clipped from all signals prior to analysis), with two eyes-open (EO) and eyes-closed (EC) sections serving as within-subject control conditions. During VHI eyes were kept closed as well (the order was: EO, EC, EC + VHI, EC’, EO’). The two resting state phases were “hands-off” during which the physician’s hands stayed off the test person. The third phase was the only “hands-on” phase during which the VHI was administered using the appropriate hand positioning (see also Fig. 1).

As standardized osteopathic hands-on technique, VH incites defined cranial regions and cranial bones at very low pressure on the scalp. Examiners utilized commonly agreed techniques to execute VHI (25). Upon recognition of an arriving tide (in OMT commonly denoted as extension) the switch was pressed, upon subsiding of the tide (denoted as flexion) the switch was released. Therefore, the interval between two consecutive footswitch operations marks the extension part of the CRI. This procedure is in line with most CRI trials that recorded laser Doppler flowmetry (LDF) data of skin blood perfusion (9). Photoplethysmography recordings used a reflective PPG probe placed on the glabella forehead region of participant and examiner. The examiner was seated behind the head of the participant who remained in a supine position.

2.4. Instrumentation

An in-house modified commercially available footswitch was used to track momentary action using a Bluetooth module. This device served to mark palpated intervals of CRI activity. Examiners had sufficient time to get used to and practice the use of the footswitch.

PPG data loggers were in-house designed to record using only the reflected red light fraction of the signal at 660 nm at a sampling rate of 125 Hz (see Fig. 2). The synchronized recorded data stream was transmitted to a local computer using Bluetooth and the complete measurement was uploaded to a private cloud for further offline processing.

2.5. Data processing and statistical analyses

Serial analyses of all PPG data were performed using Numpy (26), SciPy (27), and Matplotlib (28). This included analyses of time-frequency distributions for the various frequency bands and conversion of these compressed time series to aggregated values used to compute statistical analyses. Continuous wavelet transformation (CWT) using Morlet wavelets were computed and then reduced to 2D-time series of momentary frequencies of highest amplitude (MFHA, also known as time-frequency ridge), used to identify objective physiological correlates of palpated CRI activity.

MFHA analysis may discard important information on frequencies of lower amplitude and may thereby account for a misleading display of frequencies. Importantly, frequencies of lower amplitude are of particular relevance in our context since they may represent lower or upper harmonics of IM or LF band activity. To gain a better understanding of frequencies of PPG data and palpated extension intervals we used multiple Gaussians to fit the wavelet amplitude spectrum (WAS), since the wavelet response is falling off with the distance to the underlying frequency:

$$g\left(f\right)=\sum {a}_{i}\cdot {e}^{\frac{-{\left(f-{b}_{i}\right)}^{2}}{2{c}_{i}^{2}}}$$

where f is the frequency, a is the height of the curve's peak, b is the position of the center of the peak, and c is the width of the curve. The number of dominant frequencies may be different at every point in time, so we repeated the fit with n = 1–6 Gaussians and selected the function which described the WAS best (see Fig. 5).

Processing routines used for serial analysis of the entire study sample followed data analyses detailed in Fig. 2. Datasets with obvious sections of artefacts were excluded from further data processing. Of the remaining 31 datasets, the number of corrected data of affected time series was < 1%. Statistical analyses were performed in Python as well as with the Statistical Package for the Social Sciences (SPSS), version 27 (SPSS Inc., Chicago, IL, USA). A p-value of < .05 was considered statistically significant. MFHA values as well as frequencies of palpated waves were computed for the Vault Hold intervention section and imported to SPSS. Descriptive statistics as well as differences between mean MFHA and mean palpation for each participant were computed. Furthermore, Pearson’s r correlations were computed between mean MFHA and mean CRI rates for responders and non-responders.

3.1. Frequencies of palpated waves and momentary frequency of highest amplitude

Our examiners palpated in a group of participants responding to VHI (N = 24) with a sudden decline from 0.14 Hz (primary IM band activity) to 0.072 Hz (LF activity) 377 extension intervals. In a smaller group of participants (N = 7) who did not show this response examiners palpated 124 extension intervals. In a first step of signal analysis the palpated ’extension’ time intervals, calculated extension frequencies, and CRI frequencies computed thereof were superimposed with lowpass filtered (5 Hz) PPG data (Fig. 3A and B). This showed that palpated extension intervals captured waves in PPG signals at IM band frequencies between 0.12–0.16 Hz (≡ 7.2–9.6 cpm; see e.g., recording ‘participant 25’ in Fig. 3). Dividing extension rates by 2 yielded CRI rates (i.e., a complete cycle of extension and flexion) in the LF band between 0.06–0.07 Hz (≡ 3.6–4.2 cpm). Such rates are in agreement with rates reported in the literature for experienced examiners (10, 23).

For further visual inspection, MFHA of PPG forehead skin blood flow was superimposed on CRI rates during VHI (see Fig. 4A-D for two representative subjects). The raw PPG data of participant 7, a responder, showed prominent large waves, also known as beat, resonance, or spindle waves at stable low oscillating LF – band activity at 0.07 Hz (Fig. 4A). Superimposing this graph with computed CRI rates exhibited a distinct fit (Fig. 4B). In raw PPG data of participant 25, a non-responder (Figs. 4C and D), there were no beat or spindle waves during VHI. Here, MFHA exhibited unstable IM band activity with two short lived episodes of LF activity and hardly any fit between MFHA and computed CRI activity. Table 1 lists computed mean CRI activity, corresponding MFHA rates, and differences between these (in [Hz]) for the two subsets of participants. In non-responders computed CRI frequencies and MFHA frequencies were at a 1:2 ratio whereas this ratio was 1:1 in responders. In the literature, a 1:2 ratio is reported for the relation between palpated waves and instrumentally recorded TH waves (Nelson et al., 2006). Correlation analysis of mean MFHA and mean computed CRI frequencies showed high correlations in responders (r = .72, p < .001) as well as in non-responders (r = .83, p < .05). In responders the absolute numbers of mean MFHA matched the mean of computed absolute CRI frequencies in most cases (M_MFHA = .071 Hz; M_CRI = .066 Hz). In non-responders, absolute numbers of mean MFHA and the mean of computed absolute CRI rates were highly correlated (M_MFHA = .147 Hz; M_CRI = .068 Hz) but differed by a factor 2. The ratio 1:2 between absolute mean MFHA numbers and computed absolute CRI frequencies in non-responders demands further clarification.

Table 1

Frequency of CRI activity and momentary frequency of highest amplitude (MFHA). CRI activity was computed from palpated extension time intervals.
Group	Participant	CRI activity (Mean ± Std [in Hz])	MFHA (Mean ± Std [in Hz])	Difference (MFHA – Palp. [in Hz])
Responders	1	0.082 (± .001)	0.108 (± .053)	.027
	2	0.087 (± .015)	0.086 (± .006)	− .001
	3	0.075 (± .009)	0.090 (± .029)	.014
	4	0.063 (± .010)	0.068 (± .027)	.006
	5	0.057 (± .005)	0.058 (± .005)	.001
	6	0.058 (± .015)	0.062 (± .022)	.004
	7	0.059 (± .008)	0.058 (± .004)	− .001
	8	0.067 (± .012)	0.063 (± .004)	− .004
	9	0.073 (± .008)	0.071 (± .007)	− .002
	10	0.060 (± .011)	0.064 (± .006)	.004
	11	0.087 (± .013)	0.101 (± .034)	.014
	12	0.062 (± .015)	0.087 (± .035)	.024
	13	0.075 (± .008)	0.077 (± .027)	.002
	14	0.059 (± .008)	0.057 (± .005)	− .002
	15	0.058 (± .008)	0.053 (± .004)	− .004
	16	0.063 (± .011)	0.083 (± .039)	.021
	17	0.060 (± .007)	0.056 (± .005)	− .004
	18	0.053 (± .008)	0.052 (± .003)	− .001
	19	0.060 (± .021)	0.057 (± .017)	− .003
	20	0.050 (± .009)	0.058 (± .023)	.008
	21	0.081 (± .048)	0.061 (± .010)	− .020
	22	0.072 (± .030)	0.089 (± .037)	.017
	23	0.050 (± .008)	0.070 (± .039)	.020
	All	0.066 (± .011)	0.071 (± .016)	.005
Non-responders	24	0.054 (± .021)	0.097 (± .038)	.043
	25	0.077 (± .011)	0.198 (± .061)	.121
	26	0.045 (± .008)	0.118 (± .051)	.074
	27	0.069 (± .007)	0.130 (± .064)	.061
	28	0.094 (± .010)	0.204 (± .079)	.111
	29	0.086 (± .010)	0.157 (± .079)	.072
	30	0.051 (± .013)	0.127 (± .035)	.077
	All	0.068 (± .019)	0.147 (± .041)	.079

Note: Ratios of mean CRI/MFHA for non-responders ranged from 0.38 to 0.55, which yielded an average of 0.46 (approx. 1:2 ratio).

Figures 5A and 5B illustrate this by presenting two examples of harmonic waves discarded by MFHA analysis. Figure 5A shows in a (LF) responder a dominant MFHA wave at 0.064 Hz and a smaller wave at 0.15 Hz, which represents most likely an upper harmonic of the former. However, this wave was not displayed by MFHA. The corresponding extension section was palpated at 0.154 Hz, CRI activity was computed at 0.075 Hz (≡ 4.5 cpm) and was slightly off the actual MFHA at 0.064 Hz. Figure 5B shows in an (IM) non-responder a dominant broad based MFHA wave at 0.157 Hz and a slightly smaller narrower based wave at 0.03 Hz (VLF range), most likely a lower harmonic of the former, which was also discarded by MFHA. Here, the corresponding extension section was palpated between the two peaks of MFHA and VLF activity at 0.065 Hz, CRI activity was at 0.035 Hz (≡ 2.09 cpm). Further, we macroscopically assessed the presence of harmonics in wavelet amplitude spectra. In both responders and non-responders, harmonic waves were present in excess of 98% in dominating VLF, LF, and IM activity. While frequencies in both examples were widely similar, the mismatch between Fig. 5A and 5B is striking. In the first example (Fig. 5A) palpation captured a smaller IM band wave, and the CRI corresponded with the larger LF wave. In the second example a large IM band wave was apparently missed by palpation (interval #9) while the CRI corresponded with VLF activity. Quantification of palpated frequency bands showed that 12.1% of all palpations in non-responders were within the very low frequency (VLF) range (below 0.05 Hz) whereas in responders only 3.4% of all palpations were VLF frequencies.

A more precise estimate of deviations of palpations from MFHA was obtained by computing the root mean square of differences between MFHA and CRI (Fig. 6A-C). There were low deviations in responders and high deviations in non-responders. The palpation deviation was also highly correlated with LF interval durations (r = − .97, p < .001). This suggests that a stable strong MFHA frequency in the LF range was related to palpations with low deviation from MFHA, i.e., a relatively exact palpation. However, MFHA fluctuating between the LF, IM and HF range was related to a larger deviation to palpations (Fig. 6B). Responders showed relatively small deviations of palpation and MFHA across VH intervention whereas non-responders showed an increasing palpation error as experimental time progressed (Fig. 6C).

3.2. Analysis of palpation phase

Further analysis of phases of palpated frequencies and MFHA at the time of push (start) and release (stop) of the foot switch were performed to rule out the possibility of ‘frequency or rhythm’ bias, i.e., that examiners could have followed a certain expected rhythm without responding to real palpated sensations (Fig. 7A-F).

Distinct peaks for start and stop of phases of palpated frequencies and almost identical peaks for the phases of MFHA suggest that palpation occurred not arbitrarily or randomly. Both histograms of responders show peaks at approx. -140 degrees and 40 degrees for both start and stop. Histograms of phases of palpated frequencies index that the footswitch was pressed and released at approx. the lowest point and shortly after the highest point of the wave, resp. The equal probability of each phase for start and stop indicates that extension could either be a correlate of the rising (-180°-0°) or falling (0°-180°) edge of the PPG signal.

To determine whether the high correlation of CRI and MFHA phases of participants was arbitrary and would occur at similar probability between CRI and MFHA phases of examiners, scatterplots of CRI and MFHA phases for participants and examiners were created (see Fig. 8). We observed in VH non-responders that the scatterplots of CRI and MFHA phases for both participants and examiners showed random distributions (Fig. 8D) with low correlations of CRI and MFHA phases (participants: r = .24; range − .1 to .6; examiners: r = .44; range .1 to .6). On the other hand, scatterplots of CRI and MFHA phases for VH responders showed systematic distributions with clusters around − 150° and 30° in participants (r = .80; range .42 to 1.0). In some cases, the negative phase was related to the examiner’s start signal and the positive phase to the stop signal, while the opposite was observed in other participants (see also Fig. 7). Interestingly, more than half of examiners’ scatterplots showed less consistent patterns at rather random distributions of CRI and MFHA phases (N = 14, Fig. 8A), more widespread distribution of phases (N = 7, Fig. 8B) or similarly clustered peaks of phases as in participants (N = 2, Fig. 8C). For responders, the CRI and MFHA phases were significantly less correlated in examiners (r = .40; range − .2 to .93)) compared to participants (t(22) = 5.7, p < .001; Fig. 8E).

In this communication we validated cranial rhythmic impulse (CRI) activity using digitally marked palpation of extension intervals as well as frequency and phase computations for forehead skin perfusion oscillations. Two examiners marked subjective sensations during manual palpation using a hardware-software system. Palpations could be related to high temporal resolution objective PPG data of forehead skin perfusion. Several reasons necessitated the probing of physiological correlates of palpation. Firstly, palpation of rhythmic activity in humans has a longstanding history in medicine in general. Secondly, palpation is of crucial relevance in the practice of osteopathic manipulative treatment (OMT) (29). Thirdly, though CRI/PRM has been the subject of numerous studies, the statement of McPartland and Mein (8) still holds today: “The CRI phenomenon is poorly understood, and its functional origin remains unknown, despite its significance in CST (cranial-sacral treatment)”.

In a previous unpublished study, we have identified autonomic nervous system (ANS) activity in the low frequency (LF) and in the intermediate (IM) range in forehead skin perfusion in response to a Vault Hold intervention. We showed that the majority (77%) of participants responded to VHI with distinct highly stable low oscillating activity at approx. 0.07 Hz, while the remaining participants showed unstable primary IM-band activity at 0.142 Hz. In the current communication, this data set was analyzed with respect to palpations marked by our two examiners during VHI. Palpations consistently marked similar extension sections in PPG waves (Fig. 3). Mean CRI rates computed by multiplying extension intervals by a factor two were 0.066 Hz for responders and 0.068 Hz for non-responders. This is in line with our first hypothesis that currently observed CRI rates were comparable with those reported for experienced osteopathic practitioners (0.08 Hz; 16). Using wavelet TFD analysis, we found a distinct fit of palpations on MFHA only in responders. This was accompanied by prominent spindle or resonance waves in original PPG and highly stable LF activity in MFHA. Spindle waves were missing in non-responders and there was but a poor fit of palpation with MFHA (see Fig. 4A-D).

Using wavelet amplitude spectra, we found ANS related VLF, LF, or IM activity and harmonics in 98% of responders and non-responders, yet with differing distributions of amplitudes (see Figs. 5A and B). This means that a dominant wave at 0.07 Hz was accompanied by an upper harmonic with lower amplitude around 0.15 Hz (see e.g., Fig. 5A). Of note, we observed a high correlation between the mean of absolute MFHA frequencies and the mean of computed absolute CRI frequencies [in Hz] in responders and in non-responders, which confirmed our second hypothesis. However, there was a stable 1:1 ratio in responders and a stable 1:2 ratio in non-responders (mean MFHA_responders = 0.071 Hz; mean MFHA_{non−responders} = 0.147 Hz). This difference of objectively recorded and subjectively assessed rhythmic activity in forehead skin perfusion of non-responders has already been reported in the literature (for a review, see Fig. 1 in 4) as the ratio between low CRI rates and rates of Traube-Hering (TH) waves (range: 0.1–0.15 Hz). This historic TH model, however, falls short as an explanation of the physiological backgrounds of the 1:2 ratio. Not only are TH or TH-Mayer waves associated in humans to a narrow frequency range between 0.1–0.15 Hz. Furthermore, it is unlikely that these oscillations are of the same etiology as those which have been described by the authors they have been named after. Originally, Mayer waves have been observed at 0.05 Hz in deteriorating preparations of anesthetized spontaneously breathing rabbits (30). Furthermore, Traube’s and later authors’ pivotal works used mostly dogs for their experiments, which may exhibit rhythms of sympathetic origin as well as those related to reticular activity. However, while the former is subject of scaling, meaning that the same wave appears at higher frequencies in smaller species (31), oscillations related to the retR have been shown at comparable frequencies in humans and in dogs (17). Also, if oscillations recorded by us were of sympathetic origin as those reported for blood pressure, they were to exhibit tonic, that is non-rhythmic qualities (32). This, however, is not the case since our findings are distinctly rhythmically modulated.

Therefore, we have suggested the phenomenon of 1:2 ratios between palpated CRI and instrumentally recorded MFHA to be accounted for by IM band activity. This model incorporates lower (0.075 Hz) and upper (0.3 Hz) harmonics of primary IM frequencies (0.12–0.18 Hz). It has been demonstrated to apply to different species thereby bypassing scaling effects (17). Moreover, this approach has accumulated mounting experimental evidence over the past years (19) and has already been discussed to account for the CRI/PRM (13, 20). Our current findings appear to corroborate this approach.

Responders showing in the former study highly stable LF activity consistently exhibited low deviations (palpation errors) between CRI and MFHA rates whereas deviations in non-responders increased over time. This might be due to n:m integer number coordination reported also for extended periods of hypnoid relaxation (16). These conditions apply in case of our responders as indexed by Fig. 8B + C as palpation errors were negatively associated with duration of LF intervals during VH, i.e., the longer LF activity prevails the more accurate became palpation in terms of deviation from MFHA (see Fig. 6A-C).

Our findings suggest that the MFHA in forehead skin rhythms appears as physiological correlate of CRI palpations. The exact overlap of MFHA in participants and CRI rates further indexes that the two examiners seized palpable rhythmic sensations occurring at the head of participants. However, to rule out bias of palpations by focusing on anticipated rhythmic events we computed phases for CRI and MFHA in participants. In responders, histograms of phases of palpated (extension) sections of computed CRI during footswitch operation showed for start and stop two distinct peaks at approx. -140 and 40 degrees, thereby confirming our fourth hypothesis. In non-responders, there were also distinct peaks at -150 degrees for stop and at 30 degrees for the start (distance: 180°), but otherwise the phase prevalence was scattered randomly. This was confirmed by a scatter plot of the palpation error plotted against phases of palpated frequencies. This showed distinct accumulation of responders’ errors at approx. -150 degrees and 40 degrees, and again a rather random distribution of errors in non-responders (Fig. 7A-C). A similar pattern was obtained for phases of MFHA during start and stop of the footswitch. In responders, this showed again two distinct peaks at approx. -140 and 40 degrees for both start and stop, whereas in non-responders two distinct peaks at -150 degrees for stop and at 30 degrees for start (sum: 180°) showed randomly scattered phase prevalence. Palpation error (MFHA – frequency for each palpated interval) plotted against phase of MFHA showed distinct accumulation in responders at approx. -150 degrees and 40 degrees, and again a rather random distribution in non-responders (Fig. 7D-F). While these findings support the validity of palpated CRI rates and MFHA rates under conditions met by responders, they do not yet infer to the origin of palpated correlates.

To investigate synchronization of CRI and MFHA in examiners we further investigated phases of PPG data of examiners. This showed in a group of 14 responders a distinct coordination of start and stop of phases of CRI and MFHA, but merely a weak coordination between palpated and physiological phases in examiners (Fig. 8A). A smaller group of responders (N = 7, Fig. 8B) showed widely identical coordination of CRI and MFHA with all stop phases at approx. -130° and all start phases at approx. +50°. The coordination in their examiners was widely identical, but with some reversed orders of start and stop still indexing a possible synchronization between participants and examiners. Yet fewer responders (N = 2, Fig. 8C) showed comparable distribution of CRI and MFHA phases in participants and examiners, but as opposed to Fig. 8B, the order of start and stop phases was reversed. Overall, correlations between mean MFHA and CRI phase was significantly higher in participants compared to examiners. In non-responders, however, CRI and MFHA showed only low correlations in participants as well as examiners suggesting poor synchronization between participants and examiners.

This does not curtail coordination but possibly reflects differences in palpated sensations accounted for by extension and flexion. These terms are widely described and practiced in OMT. While our current study sample is only of limited size it is of note that this analysis exhibited almost all response patterns possible, with a clear preponderance on correct palpation in responders. This suggests that osteopaths indeed palpated participants’ physiological rhythmic activity and not their own rhythms in case LF band activity dominated. This is relevant since confounding own rhythmic activity with those of the subject studied has been discussed already for inexperienced trainees of OMT (23). Thus, we demonstrated that validity of palpation appears to depend profoundly on responding to VHI with distinct LF rhythmic activity.

Origin of the 1:2 ratio between computed CRI and recorded MFHA

The observed differences in CRI to MFHA ratio between responders and non-responders pose the question on how to tell LF responders (showing reliable palpation) apart from IM (non) responders (showing poor palpation) on the physiological level. Interestingly, our two examiners reported for their palpation sensations no differences between responders and non-responders during VHI. However, despite the highly correlated MFHA and CRI rates in responders, the physiology of non-responders remains puzzling and cannot be explained easily.

A possible approach might be found in the widely matching frequency range of IM band activity, which is compelling as an explanation for the origin of the CRI and the PRM, respectively. This notion is supported by prominent features such as that it is broad banded (0.15 Hz ± 0.03 Hz) and the forms lower and upper harmonics. Yet, another prominent feature, phase-synchronization at 1:1 or 1:2 integer number ratios with respiration, HRV, and blood pressure (17), might become the source of confounding as the frequencies in the interacting subjects synchronize and rhythmic differences between subjects 'palpably' dissolve during phase synchronizations. This notion is supported by our findings on highly synchronized phases of CRI and MFHA (see Fig. 8). This, however, mandates further comprehensive data recording and data analyses in at least one additional peripheral system in future studies.

Flexion and extension: fact or fancy

Flexion and extension refer to palpation sensations in trained observers (7, 10, 24). These terms relate to physiological correlates apparently strong enough to be palpated as minute scalp motion by sufficiently trained therapists (33). Osteopaths commonly describe extension as movement towards the examiner and flexion as a movement away from the examiner. During our investigation the osteopaths marked extension phases of the CRI which we used to compute the complete CRI. To support the construct of flexion and extension we expected to observe similar phases during start and stop throughout all measurements. However, phases of operating the start and stop button were equally likely to be at -40 and 140 degrees and start and stop did not rotate during the measurement. Based on our data we cannot relate phases of forehead skin MFHA to manually palpated extension and flexion, whereas extension and flexion are well distinguishable by examiners. Therefore, a rising or falling edge of the PPG wave is obviously not permanently related to their palpated sensations.

Limitations

There are general limitations to the overall approach of research of physiological rhythms. Among these, the most important is the large variation of CRI rates amounting approx. 20% (11, 34). This is known to an even higher degree in biological systems at rest in organisms of comparably limited complexity, such as medulla fishes, which might amount 35% (35). Due to the small sample size and only two examiners generalizability of our results may be limited and should be confirmed in a larger sample. Also, recording and analyses of only PPG signals as the only physiological system is prone to leave questions unanswered as is the case for the yet to be identified source of missed palpation in non-responders.

Outlook

Cranial osteopathy has often been criticized for its limited interrater reliability. Instrumentation of the PPG signal combined with palpation records may offer a glimpse into limitations of interrater reliability by showing considerable and corresponding variability of MFHA and palpation frequencies combined with frequency modulation (switch from 1:1 to 1:2 ratio) during VH intervention. However, it is feasible that the interaction of participant and examiner itself may be enough to induce significant physiological changes. This individuality of interactions of two physiological systems may be an additional factor decreasing interrater reliability and should be considered in further investigations by examining more physiological subsystems as well as the interaction between participant and examiner. Furthermore, we defined two groups of participants responding to OMT standard stimulus in physiologically different modes. Further studies will show whether responses to VHI are related to behavioral parameters and to treatment outcome in general. This will be of paramount importance when investigating CRI and MFHA in patients.

Funding

This research was funded by grant number: HGS-03S-18072016 of the HEAD-Genuit-Foundation, Herzogenrath/Germany.

Conflict of Interest

The authors declare that they have no conflict of interest.

Data availability statement

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Arienti C, Farinola F, Ratti S, Daccò S, Fasulo L. Variations of HRV and skin conductance reveal the influence of CV4 and Rib Raising techniques on autonomic balance: A randomized controlled clinical trial. Journal of Bodywork and Movement Therapies. 2020;24(4):395–401.
Carnevali L, Lombardi L, Fornari M, Sgoifo A. Exploring the effects of osteopathic manipulative treatment on autonomic function through the lens of heart rate variability. Front Neurosci. 2020;1075.
Cerritelli F, Cardone D, Pirino A, Merla A, Scoppa F. Does osteopathic manipulative treatment induce autonomic changes in healthy participants? A thermal imaging study. Frontiers in Neuroscience. 2020;887.
Henley CE, Ivins D, Mills M, Wen FK, Benjamin BA. Osteopathic manipulative treatment and its relationship to autonomic nervous system activity as demonstrated by heart rate variability: a repeated measures study. Osteopathic Medicine and Primary Care. 2008;2(1):1–8.
Rechberger V, Biberschick M, Porthun J. Effectiveness of an osteopathic treatment on the autonomic nervous system: a systematic review of the literature. Eur J Med Res. 2019;24(1):1–14.
Ruffini N, D’alessandro G, Mariani N, Pollastrelli A, Cardinali L, Cerritelli F. Variations of high frequency parameter of heart rate variability following osteopathic manipulative treatment in healthy subjects compared to control group and sham therapy: randomized controlled trial. Front Neurosci. 2015;9:272.
Upledger JE, Vredevoogd JD. Craniosacral Therapy. Seattle: Eastland Press; 1983. Seatlle: Eastland Press; 1983.
McPartland JM, Mein EA. Entrainment and the cranial rhythmic impulse. Alternative Therapies in Health and Medicine. 1997;3(1).
Nelson, K.E., Sergueef, N., Glonek T. Cranial Manipulation Induces Sequential Changes in Blood Flow Velociy on Demand. The AAO Journal. 2004;
Nelson KE, Sergueef N, Glonek T. Recording the rate of the cranial rhythmic impulse. Journal of the American Osteopathic Association. 2006;106(6).
Nelson KE, Sergueef N, Lipinski CM, Chapman AR, Glonek T. Cranial rhythmic impulse related to the Traube-Hering-Mayer oscillation: Comparing laser-Doppler flowmetry and palpation. Journal of the American Osteopathic Association. 2001;101(1–2).
Rasmussen TR, Meulengracht KC. Direct measurement of the rhythmic motions of the human head identifies a third rhythm. Journal of Bodywork and Movement Therapies. 2021;26:24–9.
Pelz H. Inhärente Rhythmen - Komplexe psychophysische Synergismen durch Synchronisation. Osteopathische Medizin. 2015;16(3).
Perlitz V, Schmid-Schönbein H, Schulte A, Dolgner J, Petzold E, Kruse W. Effektivität des autogenen Trainings. Therapiewoche. 1995;26:1536–44.
Perlitz V, Cotuk B, Schiepek G, Sen A, Haberstock S, Schmid-Schönbein H, et al. Synergetik der hypnoiden Relaxation. PPmP Psychotherapie Psychosomatik Medizinische Psychologie. 2004;54(6).
Perlitz V, Cotuk B, Lambertz M, Grebe R, Schiepek G, Petzold ER, et al. Coordination dynamics of circulatory and respiratory rhythms during psychomotor drive reduction. Autonomic Neuroscience: Basic and Clinical. 2004;115(1–2).
Perlitz V, Lambertz M, Cotuk B, Grebe R, Vandenhouten R, Flatten G, et al. Cardiovascular rhythms in the 0.15-Hz band: Common origin of identical phenomena in man and dog in the reticular formation of the brain stem? Pflugers Archiv European Journal of Physiology. 2004;448(6).
Lambertz M, Langhorst P. Simultaneous changes of rhythmic organization in brainstem neurons, respiration, cardiovascular system and EEG between 0.05 Hz and 0.5 Hz. Journal of the Autonomic Nervous System. 1998;68(1–2).
Pfurtscheller G, Schwerdtfeger AR, Rassler B, Andrade A, Schwarz G, Klimesch W. Verification of a Central Pacemaker in Brain Stem by Phase-Coupling Analysis Between HR Interval- and BOLD-Oscillations in the 0.10–0.15 Hz Frequency Band. Frontiers in Neuroscience. 2020;14.
Keller M, Pelz H, Perlitz V, Zweerings J, Röcher E, Baqapuri HI, et al. Neural correlates of fluctuations in the intermediate band for heart rate and respiration are related to interoceptive perception. Psychophysiology. 2020;57(9).
D’Alessandro G, Cerritelli F, Cortelli P. Sensitization and interoception as key neurological concepts in osteopathy and other manual medicines. Frontiers in Neuroscience. 2016;10(MAR).
Moran RW, Gibbons P. Intraexaminer and interexaminer reliability for palpation of the cranial rhythmic impulse at the head and sacrum. Journal of Manipulative and Physiological Therapeutics. 2001;24(3):183–90.
Sergueef N, Greer MA, Nelson KE, Glonek T. Kranialer rhythmischer Impuls: Ist die palpierte Frequenz abhängig von der Berufserfahrung der Untersucher? Osteopathische Medizin. 2011;12(4).
Woods JM, Woods RH. A physical finding related to psychiatric disorders. J Am Osteopath Assoc. 1961;60:988–93.
Liem T. Schädeldachhaltung. In: Kraniosakrale Osteopathie: Ein praktisches Lehrbuch, 3 überarbeitete Auflage. Stuttgart: Hippokrates; 2001. p. 523ff.
Harris CR, Millman KJ, van der Walt SJ, Gommers R, Virtanen P, Cournapeau D, et al. Array programming with NumPy. Nature. 2020;585(7825).
Virtanen P, Gommers R, Oliphant TE, Haberland M, Reddy T, Cournapeau D, et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nature Methods. 2020;17(3).
Hunter JD. Matplotlib: A 2D graphics environment. Computing in Science and Engineering. 2007;9(3).
Kuchera WA. Glossary of osteopathic terminology. Foundation for Osteopathic Medicine, William and Wilkins édit, Baltimore MD. 1997;1126–40.
Malpas SC. Neural influences on cardiovascular variability: Possibilities and pitfalls. American Journal of Physiology - Heart and Circulatory Physiology. 2002;282(1 51 – 1).
Anrep G, Pascual W, Rössler R. Respiratory variations of the heart rate-I—The reflex mechanism of the respiratory arrhythmia. Proceedings of the Royal Society of London Series B-Biological Sciences. 1936;119(813):191–217.
Stauss HM, Stegmann JU, Persson PB, Häbler HJ. Frequency response characteristics of sympathetic transmission to skin vascular smooth muscles in rats. American Journal of Physiology-Regulatory, Integrative and Comparative Physiology. 1999;277(2):R591–600.
Rasmussen TR, Meulengracht KC. Direct measurement of the rhythmic motions of the human head identifies a third rhythm. Journal of Bodywork and Movement Therapies. 2021;26.
Lockwood MD, Degenhardt BF. Cycle-to-cycle variability attributed to the primary respiratory mechanism. J Am Osteopath Assoc. 1998;98(1):35–6.
von Holst E. Die relative Koordination als PNinomen und als Methode zentralnerv6ser Functionsanalyse. Gesammelte Abhandlungen I Miinchen, Piper. 1939;33–132.

No competing interests reported.

Validation of subjective manual palpation using objective physiological recordings of the cranial rhythmic impulse during osteopathic manipulative intervention

Status:

Journal Publication

Version 1

Abstract

Figures

1. Introduction

2. Methods

2.1. Participants

2.2. Examiners

2.3. Study Design

2.4. Instrumentation

2.5. Data processing and statistical analyses

3. Results

3.1. Frequencies of palpated waves and momentary frequency of highest amplitude

3.2. Analysis of palpation phase

4. Discussion

Declarations

References

Additional Declarations

Status:

Journal Publication

Version 1