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

Figures

• 

  Figure 1. Experimental paradigm. Each session was comprised of 5 sections, lasting 300/330 seconds each. Section 1 and 5 (eyes open) as well as section 2 and 4 (eyes closed) were performed “hands-off”. Section 3 was “hands-on” during which examiners applied the cranial vault hold (CVH), recording palpation intervals by operating the Bluetooth footswitch. During the entire session, forehead skin blood oscillations (photoplethysmography) were recorded for participant and examiner.

• 

  Figure 2. Data processing and analysis steps. (A) Photoplethysmography data was recorded using a sampling frequency of 125 Hz. Next, artifacts were removed by detection of large jumps in signal amplitude (limited to 3·σ) and removal of the first 15 s of each section due to proneness to movement artifacts. To improve further data processing, the signal was subsampled to 5 Hz and detrended. The preprocessed signal was then converted to a time–frequency distribution using the continuous wavelet transformation and a Morlet wavelet (with parameter σ = 5). The frequency ranged from 0.05 to 0.4 Hz and was sampled in 200 steps. Finally for each time step the largest frequency was extracted. (B) Extension intervals of palpated CRI activity in human scalp perfusion were marked with a Bluetooth equipped footswitch. The time series generated had a 1 kHz resolution to allow for high resolution analysis. In order to compute the lengths of complete CRI cycles, palpated durations of extension (in [s]) intervals were multiplied by a factor 2. These CRI durations (in [s]) were then converted to physical frequencies [Hz]. The frequencies computed this way replaced the original extension intervals on the…

• 

  Figure 3. Exemplary section from cranial vault hold of participant A14. (A) upper panel: Grey: lowpass-filtered (5 Hz) PPG signal; blue: slow PPG oscillations (lowpass-filtered at 0.5 Hz); vertical lines: palpated extension sections of cranial rhythmic impulse (CRI) with start (green; pressing of footswitch) and stop (red; releasing of footswitch). (B) lower panel: Grey: lowpass-filtered (5 Hz) PPG signal; blue: slow PPG oscillations (lowpass-filtered at 0.5 Hz); vertical lines: palpated extension sections of cranial rhythmic impulse with start (green; pressing of footswitch) and stop (red; releasing of footswitch); solid horizontal black lines: computed extension frequency; dashed horizontal black lines: computed CRI. The CRI activity was consistently palpated between 0.05 and 0.1 Hz.

• 

  Figure 4. Raw forehead photoplethysmography (PPG) data of (A) participant A5 and (C) participant A14: consecutive recordings of five 300 s blocks. White bars indicate extension of 15 s in the beginning and end of each section accounting for transition effects. X-axis: time in [s], y-axis: arbitrary units [A.U.]. Note distinct beat (or resonance or spindle) waves in response to CVH in (A). (B) participant A5: momentary frequency of highest amplitude (MFHA, blue line) obtained from Morlet wavelet time frequency distributions of forehead skin PPG recordings. During CVH the plot shows stable low frequency band activity at approx. 0.06 Hz. Palpated CRI intervals converted to frequency equivalents [Hz] (orange graphs) exhibit high agreement with PPG data. This is supported by a high correlation (r = .72) with MFHA. (D) participant A14: MFHA (blue line) of forehead skin PPG recording. During CVH, the plot shows IM-band activity oscillating at approx. 0.15 Hz with transient high frequency (HF) bursts and transient low-frequency episodes at 0.06 Hz, followed by changes between IM-band and HF-band activity. Palpated extension intervals [Hz] (orange graphs) exhibit poor, if any agreement with MFHA…

• 

  Figure 5. Testing amplitude-frequency distribution. Blue line: momentary frequency of highest amplitude (MFHA) computed from wavelet amplitude spectrum (WAS); green dashed line: palpated extension section; green solid line: computed CRI activity; red line: sum of multiple Gaussian fit; grey dashed lines: multiple Gaussian fit. X-axis: frequency [in Hz], y-axis: arbitrary units. (A) The first recorded palpation interval of participant A2 (1st measurement; LF-responder). MFHA shows a dominant wave at 0.065 Hz. The palpated extension section was at 0.154 Hz while the computed CRI activity was at 0.076 Hz (≡ 4.5 cpm), which corresponds well with a prominent wave centered at 0.065 Hz. (B) Ninth recorded palpation interval of participant A2 (2nd measurement; IM-responder). MFHA with dominant wave at 0.157 Hz. Palpated extension section at 0.064 Hz between the two dominant peaks of MFHA and VLF activity. CRI activity at 0.035 Hz (≡ 2.09 cpm). (C) MFHA of examiner A’s PPG during the first recorded palpation interval of participant A2 (1st measurement; LF-responder). Identical MFHA activity in participant and examiner palpated as an extension section at 0.154 Hz. CRI activity at 0.076 Hz (≡ 4.5…

• 

  Figure 6. (A): Histogram of deviations of palpated frequencies from momentary frequency of highest amplitude (MFHA), showing a low deviation in cranial vault hold (CVH) lower harmonic waves in LF-responders (blue) and a large deviation in IM-responders (primary IM band, orange). X-axis: deviation [in Hz], y-axis: number of examinations. (B) Scatterplot of deviation of palpation from MFHA and low frequency (LF) interval durations during CVH. Stable extended duration of LF MFHA was related to low deviation of palpation from MFHA. (C) Scatterplot showing accuracy of palpation in LF-responders (blue) and IM-responders (orange) throughout CVH intervention. Plot shows a constant low error for palpation in LF-responders (horizontal blue line) and increasing error for palpation in IM-responders (ascending orange line).

• 

  Figure 7. Plots of phases of palpated sections (A–C) and of phases of momentary frequency of highest amplitude (D–F) for LF-responders and IM-responders. Palpation phases were based on rates of the cranial rhythmic impulse (CRI) which were computed by multiplying durations of palpated extension sections by a factor two. (A) Histogram of phases of palpated sections (brown) of LF-responders for the CRI during press (start, blue) and release (stop, orange) of footswitch operation showing two separate distinct peaks at approx. − 140 and 40 degrees for both start and stop. (B) Histogram of phases of palpated sections (brown) of IM-responders for the CRI during press (start, blue) and release (stop, orange) of footswitch showing distinct peaks with greater overlap at − 150 degrees for stop and at 30 degrees for start (sum: 180°) yet scattered phase prevalence. (C) Palpation error (MFHA–CRI frequency for each palpated interval) plotted against phase of footswitch marked palpated CRI frequency. Constant low errors indicated by agglomeration in LF-responders at approx. − 150 degrees and 40 degrees. Higher randomly distributed errors at for IM-responders. (D) Histogram of phases of MFHA during start…

• 

  Figure 8. Comparison of start and stop of footswitch phases of palpated cranial rhythmic impulse (CRI, y-axis) with start and stop of phases of momentary frequency of highest amplitude (MFHA, x-axis) in participants (left columns) and examiners (right columns) for LF-responders (A–C) and IM-responders (D). (A) In most LF-responders (N = 14) start and stop of phases of CRI and MFHA responded mostly linearly in participants, and randomly in examiners. (B) In a smaller group of LF-responders (N = 7), start and stop of phases of CRI and MFHA in participants were linearly as in Fig. 8A. Start and stop of phases of CRI and MFHA in examiners were also linearly, yet with some outliers. Here, stop phases were at approx. − 130°, and start phases approx. + 50°. This may reflect synchronization between participants and examiners. (C) Fewer LF-responders (N = 2) showed comparable distribution of start and stop of phases of CRI and MFHA in participants and examiners, but at reversed order of start and stop phases as those shown in Fig. 8A and 8B. This may correspond to differences in palpated sensations of extension and flexion. (D) In IM-responder (N = 7), start and stop of phases of CRI and MFHA were…