Cavitation Studies of Lumbar Zygapophysial Joints using Vibration Measurements 
 
 
by 
 
Preetam Nirmalkumar Bora 
 
 
 
 
A thesis submitted to the Graduate Faculty of 
Auburn University 
in partial fulfillment of the 
requirements for the Degree of 
Master of Science 
 
Auburn, Alabama 
May 9, 2011 
 
 
 
 
Keywords:  zygapophysial, accelerometer, cavitation,  
spinal manipulation, artifacts, vibration 
 
 
Copyright 2011 by Preetam Nirmalkumar Bora 
 
 
Approved by 
 
P. K. Raju, Chair, Thomas Walter Professor of Mechanical Engineering 
Gregory D. Cramer, Professor of Anatomy and Dean of Research, National University of 
Health Sciences, Lombard, IL 
Dan B. Marghitu, Professor of Mechanical Engineering 
 
 
 
 
 
 
ii 
Abstract 
 
 
The purpose of this research project was to locate the level of cavitation in 
zygapophysial (Z) joints in the lumbar spine (L1/L2-L5/S1) and the sacroiliac joints 
(SIJs) during side-posture positioning and spinal manipulative therapy (SMT). 
Forty healthy subjects were sought to complete this study, in which seven 1cm3 
accelerometers were placed on the L1-L5 spinous processes (SPs) and the S1 and S2 
sacral tubercles. Two more accelerometers were placed lateral to the midline, one 3 cm to 
the left and the other 3cm to the right mid-distance between the L4 and L5 
accelerometers. Subjects were randomized into one of the two groups: Group 1, SMT 
Group (n=30) or Group 2, Side Posture Positioning Group (n=10). Data from the 
accelerometers was collected from both groups. 
Two types of reliability studies were performed: one, comparing the results of two 
observers (blinded to the results of one another) in identifying vertebral level of 
cavitations and the other, comparing presence of a cavitation found by clinician, subjects 
and accelerometers.  
Artifacts (noise signals other than cavitation) were identified, described and 
categorized as Flexible Spine (or loose superficial fascia), Hyper-Muscular Subjects, 
Resisting or Not Relaxed Subjects, Unstable Hand Contact (UHC) and Miscellaneous, 
depending on notes taken during data acquisition, assessment, and the unique pattern of 
waveform signals.  
iii 
Vertebral levels of cavitations were identified and results were compared 
between: Group 1 vs. Group 2, targeted (L3/L4, L4/L5, and L5/S1) vs. non-targeted Z 
joints, left-side (up-side during side-posture positioning and SMT) vs. right-side.  
 Fifty-six (56) cavitations were recorded from 40 subjects. A reliability study 
comparing the results of two observers (blinded to the results of one another) to identify 
cavitations from the accelerometers recordings showed ?almost perfect agreement? 
between the two observers (Kappa = 0.841, Std E = 0.151). The reliability study 
comparing presence of a cavitation found accelerometers recordings of cavitations were 
in ?almost perfect agreement? with the subjects and clinician. The clinician and subjects 
were also in ?almost perfect agreement?. 
 Subjects randomized into Group 1 cavitated more often (96.67% of Group 1 
subjects) than those who were just held in side-posture positioning (30% of Group 2 
subjects). Most cavitations were recorded on the left side (up-side 94.63%) and at the 
targeted Z joints (73.20%). Multiple cavitations were recorded from the same joint. 
Double cavitations were recorded in seven joints and, interestingly, four cavitations in 
one joint. A maximum of six cavitations were recorded from a subject randomized in 
Group 1, with four cavitations at the same joint (Left L3/L4) and two cavitations at 
another joint (Left L2/L3). 
 
 
 
 
 
 
 
 
 
iv 
Acknowledgements 
 
 
I would like to express my deepest gratitude towards Dr. P.K. Raju, Dr. Gregory 
Cramer, and Dr. Kim Ross for their support, motivation and guidance throughout the 
research work. I would like to thank Dr. Dan Marghitu for being a positive support 
throughout my research work and for serving as a committee member. I would also like 
to thank David Cramer, who helped me in editing grammar for my thesis.  
I would like to thank all the project members who helped in making this research 
project successful particularly Drs. Adam Habeck and Scott Selby. I would also like to 
thank Frank Balester in helping to write programs in LabVIEW. I very much appreciate 
Ray McKinnis, PhD for helping with statistical analysis. I would also like to thank Robert 
Hansen for helping me in editing figures. I would like to thank the funding support of 
NIH/NCCAM (Grant #3R01AT000123-06S2) for the research project 
I wish to express my gratitude to all my family members for being supportive and 
keeping faith in me at all times. I am highly indebted to my parents, for their 
unconditional love and support, and for always encouraging me to follow my dreams. 
 
 
 
 
 
 
 
 
 
 
v 
Table of Contents 
 
 
Abstract ............................................................................................................................... ii 
Acknowledgements ............................................................................................................ iv 
List of Figures .................................................................................................................... ix 
List of Tables ................................................................................................................... xiv 
1. INTRODUCTION ...........................................................................................................1 
1.1 Review of Literature ...................................................................................................1 
1.2 Main Objective..........................................................................................................14 
2. ANATOMY AND PHYSIOLOGY OF LUMBAR SPINE ..........................................16 
2.1 Spinal Column ..........................................................................................................16 
2.2 Vertebrae ...................................................................................................................19 
2.3 Zygapophysial (Z) Joints ..........................................................................................20 
3.  ACCELEROMETERS AND MICROPHONES ..........................................................23 
3.1 Accelerometers .........................................................................................................23 
3.2 Types of Accelerometers............................. .............................................................23 
3.2.1 Piezoelectric Accelerometers ...........................................................................24 
3.2.2 Piezoresistive Accelerometers .........................................................................26 
3.2.3 Capacitive Accelerometers ..............................................................................27 
3.3 Applications ..............................................................................................................28 
3.3.1 Medical Applications .......................................................................................28 
vi 
3.3.2 Transport ..........................................................................................................28 
3.3.3 Consumer Electronics ......................................................................................29 
3.4 Microphones .............................................................................................................30 
3.5 Types of Microphones ..............................................................................................31 
3.5.1 Condenser Microphones ..................................................................................31 
3.5.2 Dynamic Microphones .....................................................................................32 
3.5.3 Piezoelectric Microphones ...............................................................................33 
3.5.4 Fiber Optic Microphones .................................................................................34 
3.6 Applications of Microphones in Research and Industry ...........................................35 
3.6.1 Research and Product Design ..........................................................................35 
3.6.2 Compliance ......................................................................................................36 
3.6.3 Environmental Noise Analysis ........................................................................36 
4. INSTRUMENTATION AND METHODOLOGY ........................................................37 
4.1 Instrumentation .........................................................................................................37 
4.2 Refining Previously Designed Accelerometry Techniques Used to  
Evaluate Cavitations Following Side-Posture Positioning and SMT .......................42 
4.2.1 Previously Designed Accelerometry Methods.................................................42 
4.2.2 Re-Design of the Accelerometry Method ........................................................44 
4.3 Testing the System ....................................................................................................46 
4.3.1 Re-Designed Methods ......................................................................................46 
4.3.2 Previously (Originally) Designed Method .......................................................50 
4.4 Clinical Study............................................................................................................55 
4.5 Data Acquisition for Subjects Randomized into Group 1 ........................................63 
vii 
4.6 Data Acquisition for Subjects Randomized into Group 2 ........................................65 
4.7 Information Collected for Subject Data Base during the Accelerometry  
Procedures .................................................................................................................67 
5. DATA ANALYSIS ........................................................................................................68 
5.1 Training in Data Analysis .........................................................................................69 
5.2 Data Analysis by Two Observers .............................................................................73 
5.3 Amplitude of Cavitation Waves................................................................................76 
5.4 Artifacts.....................................................................................................................79 
5.5 Types of Artifacts .....................................................................................................81 
5.5.1 Flexible Spine (or Loose Superficial Fascia) Subjects ....................................81 
5.5.2 Hyper-Muscular Subjects (e.g., Body Builders) ..............................................84 
5.5.3 Resisting or Not Relaxed Subjects ...................................................................86 
5.5.4 Unstable Hand Contact (UHC) (i.e., Clinician?s Hand Contact) .....................88 
5.5.5 Miscellaneous ..................................................................................................89 
5.6 Data Sheets................................................................................................................91 
5.7 Reliability Studies ...................................................................................................108 
5.7.1 Inter-Rater Agreement between Computer (Accelerometers), Clinician,  
         and Subject .....................................................................................................108 
5.7.2 Reliability Study for Agreement Assessing Accelerometry Data..................111 
5.8 Summary of Cavitation Assessment .......................................................................113 
5.9 Statistical Analysis ..................................................................................................113 
6. RESULTS ....................................................................................................................114 
6.1 Reliability Studies ...................................................................................................115 
viii 
6.1.1 Agreement between Accelerometers and Subjects ........................................115 
6.1.2 Agreement between Clinician and Subjects ..................................................117 
6.1.3 Agreement between Accelerometers and Clinician .......................................118 
6.1.4 Agreement Assessing Accelerometry Data ...................................................120 
6.1.5 Summary of Reliability Studies .....................................................................123 
6.1.5.1 Agreement between Accelerometers, Clinician and Subjects ..............123 
6.1.5.2 Agreement between Two Observers in Assessing  
Accelerometry Data ..............................................................................123 
6.2 Number and Percentage of Cavitations (CAV) Detected By the  
      Accelerometers at Each Joint ..................................................................................124 
6.3 Cavitations in Group 1 and Group 2 .......................................................................127 
6.4 Cavitations at Targeted and Non-targeted Area ......................................................128 
6.5 Cavitations at Left Side and Right Side ..................................................................130 
6.6 Multiple Cavitations................................................................................................133 
6.7 Amplitude of Cavitations ........................................................................................135 
7. DISCUSSION ..............................................................................................................138 
8. CONCLUSIONS..........................................................................................................145 
REFERENCES ................................................................................................................147 
APPENDICES .................................................................................................................153 
A. Technical Specifications .............................................................................................153 
B. Auburn University Institutional Review Board (IRB) Approval ................................166 
 
 
 
ix 
List of Figures 
 
 
Figure 2.1: Three views of the vertebral column ...............................................................18 
Figure 2.2: Parts of lumbar vertebrae ................................................................................20 
Figure 2.3: Lumbar zygapophysial joint ............................................................................21 
Figure 2.4: Motion in Z joints during flexion and extension .............................................21 
Figure 3.1: Piezoelectric effect ..........................................................................................24 
Figure 3.2: Piezoelectric accelerometer .............................................................................25 
Figure 3.3: Piezoresistive accelerometer ...........................................................................27 
Figure 3.4: Capacitive accelerometer ................................................................................27 
Figure 3.5: Cross section of condenser microphone ..........................................................32 
Figure 3.6: Cross section of dynamic microphone ............................................................33 
Figure 3.7: Cross section of piezoelectric microphone ......................................................34 
Figure 4.1: Block diagram of the accelerometer data collection system used in  
 this study ..........................................................................................................40 
Figure 4.2: Portable data collection cart ............................................................................41 
Figure 4.3: Position of accelerometers as per the original (previous) design ....................43 
Figure4.4: Position of accelerometers in the re-designed method .....................................45 
Figure 4.5: Testing the re-designed system .......................................................................47 
Figure 4.6: Oscilloscope showing complete time scale of the collected data for re-
designed system ...............................................................................................48 
x 
Figure 4.7: Expanded time and amplitude scales of Figure 4.6 .........................................49 
Figure 4.8: Testing previously designed method ...............................................................51 
Figure 4.9: Oscilloscope showing complete time scale of the collected data for the 
previous (original) designed system ................................................................52 
Figure 4.10: Expanded time and amplitude scales of Figure 4.9 .......................................53 
Figure 4.11: Error-resisting labeling methodology ............................................................57 
Figure 4.12: Additional error-resistant methods ................................................................58 
Figure 4.13: Labeled accelerometers placed in marked boxes ..........................................59 
Figure 4.14: Amplifiers marked with respective accelerometer number attached  
                     to input (I/P) and output (O/P) ports .............................................................59 
Figure 4.15: All nine accelerometers taped in the assigned position on a  
                     subject?s back ................................................................................................62 
Figure 4.16: LabVIEW data acquisition window ..............................................................64 
Figure 4.17: Spinal manipulative therapy (SMT) performed while recording from 
accelerometers...............................................................................................65 
Figure 5.1: LabVIEW data analyzing window before accelerometry data  
                   was loaded .......................................................................................................68 
Figure 5.2: Location of Left L1/L2 Z joint (joint that cavitated in Figure 5.3) .................70 
Figure 5.3: Oscilloscope recording from subject (Left L1/L2 cavitation).  
                   A= complete time scale. B= expanded timeline of Figure A. 
                   C=expanded timeline of Figure B ...................................................................72 
Figure 5.4: Microsoft Windows Live Meeting, showing number of persons in  
                   the meeting (2) and the content to be shared in the meeting ..........................74 
xi 
Figure 5.5: A and B shows LabVIEW data being shared using 
                  Windows Live Meeting....................................................................................75 
Figure 5.6: A, B and C show cavitations occurring at the Left L3/L4 Z joint  
                   from three different subjects with high, medium, and low levels  
                  of cavitation amplitude, respectively ...............................................................78 
Figure 5.7: Accelerometer placed at SP L1 responding alone. This was one type of 
artifact identified during the study ...................................................................80 
Figure 5.8: Accelerometers placed at SP L1 and SP L5 responding at the same  
                   time, indicating that this was an artifact rather than a cavitation ....................80 
Figure 5.9:  The flexible spine (or loose superficial fascia) subject artifact ......................83 
Figure 5.10: The hyper-muscular subject artifact ..............................................................85 
Figure 5.11: Resisting or not relaxed subject?s artifact .....................................................87 
Figure 5.12: Unstable Hand Contact (UHC) (i.e., clinician?s hand contact) artifact.89 
Figure 5.13: Miscellaneous artifacts ..................................................................................90 
Figure 5.14: Complete time cycle for subject ID 0035 ......................................................96 
Figure 5.15: Artifact 1 for subject ID 0035 .......................................................................96 
Figure 5.16: Artifact 2 for subject ID 0035 .......................................................................97 
Figure 5.17: Artifact 3 for subject ID 0035 .......................................................................97 
Figure 5.18: Artifact 4 for subject ID 0035 .......................................................................98 
Figure 5.19: Artifact 5 for subject ID 0035 .......................................................................98 
Figure 5.20: Artifact 6 for subject ID 0035 .......................................................................99 
Figure 5.21: Artifact 7 for subject ID 0035 .......................................................................99 
Figure 5.22: Artifact 8 for subject ID 0035 .....................................................................100 
xii 
Figure 5.23: Cavitation 1, full view for subject ID 0035 .................................................101 
Figure 5.24: Expanded (zoomed) view of figure 5.23 .....................................................101 
Figure 5.25: Cavitation 2, full view for subject ID 0035 .................................................102 
Figure 5.26: Expanded (zoomed) view of figure 5.25 .....................................................102 
Figure 5.27: Complete time scale for subject ID 0046 ....................................................105 
Figure 5.28: Artifact 1 for subject ID 0046 .....................................................................105 
Figure 5.29: Artifact 2 for subject ID 0046 .....................................................................106 
Figure 5.30: Artifact 3 for subject ID 0046 .....................................................................106 
Figure 5.31: Artifact 4 for subject ID 0046 .....................................................................107 
Figure 5.32: Artifact 5 for subject ID 0046 .....................................................................107 
Figure 6.1: Frequency chart for inter-rater agreement (Kappa) between  
                   accelerometers and subjects ..........................................................................116 
Figure 6.2: Frequency chart for inter-rater agreement (Kappa) between  
                  clinician and subjects .....................................................................................118 
Figure 6.3: Frequency chart for inter-rater agreement (Kappa) between  
                  accelerometers and clinician ..........................................................................119 
Figure 6.4: Frequency chart for reliability study of analysis of  
                   accelerometry recordings ..............................................................................122 
Figure 6.5: Percentage of cavitations recorded at each joint ...........................................126 
Figure 6.6: Percentage of cavitation in Group 1 and Group 2 .........................................128 
Figure 6.7: Percentage of cavitations at targeted and non targeted Z joints ....................129 
Figure 6.8: Percentage of cavitations recorded on left side and right side ......................130 
Figure 6.9: Posterior view of the spine with the number of cavitations recorded 
xiii 
                   at each joint on the left and right side for Group 1 (SMT group) .................131 
Figure 6.10: Posterior view of the spine with the number of cavitations recorded  
                    at each joint on the left and right side for Group 2 (non-SMT group).........132 
Figure A.1: Geometric dimensioning of Bruel & Kjaer Delta Tron4507  
                   Accelerometer ...............................................................................................154 
Figure A.2: Bruel & Kjaer 2693 NEXUS signal conditioning amplifier ........................157 
Figure A.3: NI PXIe-6356 Simultaneous X Series Data Acquisition Card .....................161 
Figure A.4: NI BNC-2090 rack-mounted BNC terminal block .......................................161 
Figure A.5: NI PXIe-1073 chassis ...................................................................................162 
Figure A.6: NI PXIe-express card 8360 ..........................................................................164 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
xiv 
List of Tables 
 
 
Table 4.1: Assigned number for each accelerometer and the designated  
                 placement location on the subject .....................................................................61 
Table 5.1: Assessment of Group 1 accelerometer data-Artifacts ......................................94 
Table 5.2: Assessment of Group 1 accelerometer data- Cavitations .................................95 
Table 5.3: Amplitude of cavitations Group 1 ....................................................................95 
Table 5.4: Assessment of Group 2 accelerometer data- Artifacts ...................................103 
Table 5.5: Assessment of Group 2 accelerometer data- Cavitations ...............................103 
Table 5.6: Amplitude of cavitations Group 2 ..................................................................103 
Table 5.7: Presence of cavitations as identified by the accelerometers,  
                 clinician, and the subject .................................................................................109 
Table 5.8: Assignment of numeric values to Z joints for reliability study 
                 data entry .........................................................................................................112 
Table 6.1: Inter-rater agreement (Kappa) between accelerometers and subjects ............116 
Table 6.2: Inter-rater agreement (Kappa) between clinician and subjects .......................117 
Table 6.3: Inter-rater agreement (Kappa) between accelerometers and clinician ...........119 
Table 6.4: Interpretation of cavitations by PB and Dr. Ross ...........................................121 
Table 6.5: Inter-rater agreement (Kappa) for analysis of accelerometry  
                 recordings between two observers ..................................................................122 
Table 6.6: Number and percentage of cavitations recorded at each joint ........................125 
xv 
Table 6.7: Distribution of cavitations detected by accelerometers ..................................134 
Table 6.8: Amplitude of cavitation signals ......................................................................135 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1 
1. INTRODUCTION 
 
 
Therapeutic value of cavitation and gapping is discussed in this chapter along 
with the future scope of diagnostic methods. The review of literature in each section is 
arranged in chronological order. 
 
1.1 Review of Literature 
Zygapophysial (Z) Joint Gapping  
Z joint gapping is explained as separation of the surfaces of a Z joint and has 
therapeutic value in the clinical studies. Janse (1976) explains that the chiropractic 
adjustment is an effective way of restoring motion to hypomobile vertebral and pelvic 
segments: ?If, to a segment held in partial articular fixation [Z joints], a quick deft thrust 
is applied, not sufficient in force to produce trauma but sufficient to break the fixation, 
there will result a definite tendency for the articular surfaces to find the normal 
relationship in full freedom of motion?. Our research study was part of a larger study to 
determine the relationship between cavitation and Z joint gapping and its potential 
therapeutic benefits as discussed here. If a relationship is found between cavitation and 
gapping, then cavitation can be used to locate the specific Z joint gapped. 
McFadden and Taylor (1990) performed a study on the effect of axially rotating 
the lumbar spine and gapping Z joints. Lumbosacral spines (L3-S2) were removed, along 
with all muscles and tissues, from 12 unfixed cadavers ranging in age from 14 to 75 
 2 
years. They fixed one end of each spine while a weight and pulley system applied 
torque/twist to the opposite end. They concluded that lumbar motion segments are too 
stiff to undergo pure axial rotation in healthy spines, producing gapping only in the Z 
joints of spines showing significant pathology with degenerative changes. Healthy 
subjects were sought for our research study as they can be helpful to see if Z joint 
gapping takes place when followed by spinal manipulative therapy (SMT). 
Cramer et al., (2000; 2002) performed two studies to test if lumbar side-posture 
SMT produces gapping in Z joints. Volunteers were randomized into 4 different groups: 
Group 1, (neutral position, followed by side-posture positioning); Group 2 (neutral 
position, followed by side posture spinal adjusting on left side, followed by neutral 
position); Group 3 (neutral position, followed by side posture spinal adjusting on the left 
side, followed by side posture positioning) and Group 4 (neutral position followed by 
neutral position). First, a magnetic resonance imaging (MRI) scan was taken in neutral 
position, followed by a second scan either in neutral or side posture position, depending 
in which group the patient was randomized. MRI scans were taken of the L3-S1 region. 
Patients with side-posture adjusting produced more gapping than the control group or the 
side-posture positioning without spinal adjusting group. The left side, in particular 
produced the most gapping, as the patients underwent SMT of the left side. Similar 
methods were used in our study by first taking an MRI scan in neutral position followed 
by side-posture positioning to determine cavitation in Z joints followed by SMT or side-
posture positioning. 
 
 
 3 
 Cavitation 
Unsworth et al., (1971) conducted a study of cavitation in the 
metacarpophalangeal (MCP) joint. A machine was designed and built to apply loading on 
the MCP joint in 17 males. X-rays were taken during the loading before cavitation of the 
joint and simultaneously when cavitation was heard, the amount of load applied was 
recorded. A crescent shaped area of high contrast was seen in the space between the 
articular surfaces on X-rays for patients who produced cavitation and was absent from the 
X-rays of patients who did not produce cavitation. They concluded that low pressures 
develop in the synovial fluid during the separation of joints and caused gas to liberate and 
vaporize from the fluid. When these vapor-filled bubbles move into the higher pressure 
areas, instant collapse takes place with a very high energy release called cavitation. 
Theory behind cavitation and its effect on the MCP joint is discussed in this study and a 
similar theory could be applied to Z joint cavitation and gapping. 
Meal and Scott (1986) performed a study to see if cavitation had therapeutic 
value, to see when the joint went into the paraphysiological space (the expanded range of 
motion following cavitation) and to understand when the sound is produced in relation to 
separation of the joint surfaces. They used a method of simultaneously recording the 
sound produced by the joint while measuring the tension across the MCP joint. A cubicle 
was used to isolate extraneous sounds from the actual MCP joint sound. An ultraviolet 
recorder was used instead of a pen recorder for greater accuracy, wider choice and 
broader range of speed. A tension transducer was positioned directly on the joint using 
adhesive tape. The experiments showed that the cavitation played an important role as an 
indicator that the MCP joint had separated and moved into the paraphysiological space. 
 4 
Joint cavitation was seen on oscilloscope recordings as a double sound wave, with the 
majority of joint surface separation occurring between the two sound waves. Sound 
recordings have been used to observe cavitation following gapping in the MCP joint and 
can be used to determine cavitation in Z joints and SIJs. 
Brodeur (1995) reviewed the literature on cavitation associated with SMT. While 
previous studies had concluded that cavitation occurred during SMT as joints separated 
suddenly, the mechanisms behind this separation remained unclear. Brodeur?s review led 
him to propose that cavitation occurs by elastic recoil of the synovial capsule. This recoil 
in Z joints initiates beneficial neurological reflex actions (decreased pain and muscle 
relaxation). Therefore, if cavitation was required for the recoil of the synovial capsule 
that led to its reflex actions, cavitation would be essential to the therapeutic benefit of 
joint manipulation. Consequently, studying cavitation in Z joints and SIJ is important in 
understanding the therapeutic value of gapping. 
Reggars and Pollard (1995) conducted a study to record cavitation using 
microphones in the cervical region. They placed a microphone on the left and right side 
of the neck immediately in front of (anterior) the transverse process of C2. Interestingly, 
in 94% of the subjects, the cavitation was recorded on the same side as head and neck 
rotation (i.e., head and neck rotated to the right was associated with cavitation on the 
right). They were unable to identify the segmental (vertebral) level of cavitation, since 
only two microphones were used. Our study could determine if higher percentages of 
cavitations occur in the same side in which force is applied during SMT or side-posture 
positioning. Reggars and Pollard?s research is discussed further as microphones were an 
important part of recording cavitation. 
 5 
Beffa and Mathews (2004) conducted a study to locate cavitation sounds during 
?L5 spinous hook? and lower SIJ SMT. Eight microphones were taped to the subject?s 
back at respective Z joints using adhesive tape, and their locations were confirmed by 
taking a radiographic image. After SMT Fast Fourier Transformation (FFT) analysis was 
done, the maximum amplitude of each microphone was identified to determine the 
location of cavitation. No correlation was found to exist between the anatomical location 
of cavitation sounds and the respective technique used for SMT. New accelerometry 
techniques used in our study could be used for locating level of cavitation to help develop 
methods to improve SMT techniques. Beffa and Mathews? research is included in the 
microphone section of this chapter because microphones were important part of their 
study. 
Ross et al., (2004) used accelerometers to determine if cavitation occurs at the 
targeted joints following SMT. They used three accelerometers (same as used in our 
study) positioned at the twelfth thoracic (T12), third lumbar (L3), and first sacral 
segments (S1). While the patient was lying in the side-posture position, they measured 
the distance between the accelerometers. Since they found that the velocity of vibration 
waves from cavitations traveled through the body at 1400m/sec, to get distance 
discrimination of 0.5cm the sampling rate was set to 320,000Hz. A grease pencil marked 
the targeted area where a cavitation was supposed to occur. Therefore, the location of 
each cavitation was determined from the distances above or below each accelerometer 
based on the time of the accelerometers? response. The side where cavitations occurred 
(left or right) could not be determined, however this study helped us in determining the 
sampling rate of data acquisition from accelerometers and deciding additional 
 6 
accelerometers on the left and right side of the spine to determine on which side the 
cavitation occurred. Our study will help in finding cavitation at specific lumbar Z joints 
or SIJs and also side (left or right) of cavitation. Ross?s study is also included in the 
accelerometers section of this chapter, because accelerometers were an important part of 
this study. 
Flynn et al., (2006) conducted a study to determine the relationship between 
cavitation associated with spinal manipulation and improvement in pain and functions of 
patients suffering from low back pain. For this study, 70 patients were randomly assigned 
to get physical therapy for a total of 5 sessions. Therapists recorded if they heard 
cavitations during high velocity thrust manipulation and outcomes were measured on an 
11-point rating scale called the Oswestry Disability Questionnaire. They concluded that 
cavitation was heard in 84% of the patients and no difference was recorded in pain during 
baseline visits or any of the follow-up therapies related with the cavitation. Results also 
showed that cavitation may not relate to possible outcomes from spinal manipulation in 
terms of the improvement for patients with low back pain. This study used a specific 
technique for manipulation of SIJ that is very rarely used by clinicians. 
Teodorczyk-Injeyan et al., (2006) performed a study to examine the effect of 
SMT on the in vitro production of proinflammatory cytokines (chemicals involved in 
producing inflammation) in relation to the systematic levels of neurotransmitter substance 
P. Sixty-four subjects were sought to complete this study. It was found that subjects who 
cavitated after SMT had reduced production of inflammatory cytokines as compared to 
the subjects in which no cavitation occurred after SMT. Cavitation has been associated 
 7 
with positive outcomes and our study can be used to locate level of cavitation at lumbar Z 
joints and SIJs.  
Teodorczyk-Injeyan et al., (2008) performed a study to determine changes in the 
production of the immunoregulatory cytokine interleukin 2 (IL-2) followed by SMT and 
determined changes that might differ with or without cavitation. Seventy-six 
asymptomatic subjects were sought to complete the study. It was concluded that in the 
presence or absence of cavitation followed by SMT there was increase in the presence of 
IL-2 synthesis compared with controls and hence it is important to study localization of 
cavitation.  
Cramer et al., (2008) conducted a study to find out if the method proposed in this 
study was feasible or not. First, an MRI scan of the subject was taken in the neutral 
position to find out if there was any pathology in the subject and also to determine the 
accuracy with which high signal markers (filled with mineral oil) could be taped on the 
T12, L3 and S1 locations. This was used for accurate placement of the accelerometers as 
the high signal markers were easily identified on MRI. The markers also left a slight 
indentation on the subject?s back. After this, the subject was held in the side-posture 
position with accelerometers placed at the desired location; the distance between each of 
the accelerometers was measured, and the subject was given SMT with the left side up. 
This system was capable of finding the location of cavitation but not the side of 
cavitation, because there were no accelerometers taped on the left and right side of the 
spine. Methods used in our study were similar to the methods discussed above, except the 
number of accelerometers used in our study was greater, which determined cavitations at 
targeted and non-targeted area. This study is also discussed in the ?accelerometers? 
 8 
section of this chapter, because of the importance of using accelerometers in locating 
cavitation in Z joints. 
 
Accelerometers:  
Nokes et al., (1984) recorded vibration response of bone from a skin mounted 
accelerometer; preloading of the accelerometers was done to overcome the damping 
effect from the interposed soft tissue. Measurements were taken on left legs of randomly 
chosen cadavers ranging in age from 68-75 years. Two accelerometers weighing15g were 
used. One accelerometer was positioned on the skin, perpendicular to its surface, while 
the second was mounted adjacent to the first, directly on the bone by means of a pin that 
was screwed to its base and put through the skin of the left leg tibia. An impulse was 
applied by a 28g steel ball dropped from a constant height down a vertical Perspex tube. 
They concluded that excessive preload in accelerometers caused distortion of the 
recorded signal by a high frequency component. The required preload is proportional to 
the soft tissue thickness, but for any given soft tissue thickness, the range of allowable 
preload was fairly wide. Strong adhesive tape was used in our study to apply sufficient 
preload on accelerometers and help in keeping accelerometers in firm contact with skin. 
Ross et al., (2004) performed a study as discussed earlier using three 
accelerometers to find location of cavitation followed by SMT. Accelerometers were an 
important part of the study and helped in locating cavitation some distance away from the 
accelerometers.  
Cramer et al., (2008) conducted a feasibility study using three accelerometers to 
find out if the accelerometery methods proposed in our study was feasible or not. This 
 9 
study was also discussed in the previous section because of its importance in locating 
cavitation, but not the side (left or right) of cavitation. 
 
Microphone:  
Heffez and Blaustein (1986) used modern sonographic techniques for recording 
and interpreting joint sounds from the temporomandibular joints (TMJ). They discussed 
modern sonographic techniques and presented preliminary observations on the wave 
patterns and power spectral analysis of TMJ sounds. They used a digital DSP-200 
stethoscope and specially designed external auditory canal listening device for the 
purpose of auscultation and recording of the TMJ sounds, in a soundproof room. They 
found that TMJ sounds are a composite of sinusoidal waves. This method can be used to 
find unique pattern of waveform signals of cavitation as compared to noise signals in our 
study. 
Gay et al., (1987) performed a study to record and analyze sound emitted from the 
left and right TMJs during simple motions for diagnosing disorder of the joint from a 
total of 79 patients. Sound was detected from each joint simultaneously using two 
separate vibration transducers (small conventional condenser microphones). They found 
three different categories of TMJ disorders: internal derangement, degenerative disease 
and extra capsular disorders depending on acoustic waveform pattern. They also found 
that amplitude (or changes in sound amplitude) might have a relation to the severity of 
the disease. However it was not possible to quantify this parameter because amplitude or 
changes in sound amplitude was affected by the impedance of the intervening tissue, the 
distance between the transducer and the joint, and the acoustical fit of the transducer at 
 10 
the placement site. But one important thing was shown: that although the amplitude of 
the signal is conditioned by transmission line factors, the duration and the shape of the 
signal are not. Considering these factors, it will be difficult to estimate gapping 
depending on the amplitude of cavitation signals in other study not reported here. 
As discussed earlier, Reggars and Pollard (1995) determined the side of cavitation 
in the cervical region, but not the vertebral level of cavitation by using two microphones. 
Reggars and Pollard?s research is included in this section because microphones were 
important part of their study. 
Siffert and Kaufman (1996) conducted a study to determine the capabilities and 
limitations of auscultatory percussion (the act of striking a body part with short, sharp 
blows as an aid in diagnosing the condition of the underlying parts by the sounds 
obtained) technique for assessing the healing of bone fractures. A portable vibration 
analysis device was used and experiments were conducted to objectively evaluate the 
capabilities of auscultatory percussion techniques. Technique was based on observations 
of the ability of bone to conduct sound, which required simply tapping a bony 
prominence distal to the suspected or known fracture and listening over the opposite end 
of that bone. They found that decrease in volume and pitch of vibration energy wave 
indicates discontinuity at a fracture site or an established nonunion. They concluded that 
vibration assessment is, however, subject to systematic and random errors, and thus 
cannot always discriminate between the stages of healing in a fractured bone. In addition; 
various artifacts can lead to significant uncertainty in diagnosis, and more sophisticated 
methods could be used in the future to improve accuracy of vibration techniques, as well 
as placement of transducers. Accelerometers were placed on spinous processes (SPs) in 
 11 
our study because bone is discussed as a good conductor for transmitting vibration waves 
and an accurate placement of transducer on bone is necessary to overcome dampening 
effects. 
As discussed earlier, Beffa and Mathews (2004) conducted a study to locate the 
cavitation sounds using microphones, during an ?L5 spinous hook? and lower SIJ SMT. 
Microphones were an important part of this study in locating cavitation hence, included 
in this section. 
 
Diagnosis 
Accelerometers and microphones can be used in future studies to develop new 
techniques to deepen the understanding of the mechanics of the spine.  
Chu et al., (1976) performed a study to analyze the sound emitted by normal and 
pathological knee joints during active movements. Two identical microphones were 
attached to the knee by means of a holding assembly. One microphone was directed 
towards the knee and other was rotated 180 degrees opposite to the first microphone (i.e., 
away from the knee) to hear the surrounding noise and implement the noise cancellation 
method. They concluded that the recordings covering normal, rheumatoid, arthritic, and 
degenerative knees showed respective waveforms, spectral patterns and statistical 
property. Also, auto correrelation appeared to be unique and therefore may well prove to 
be a promising non-invasive tool for early detection of the type and extent of knee joint 
damage. This noise cancellation method can be used in future studies to eliminate 
unwanted surrounding noise by using a differential amplifier for microphones and 
develop new methods in detecting degenerative disease in spine. 
 12 
Gallo et al., (1993) attempted to record the sounds emitted by healthy TMJs with 
and without mandibular movements for determination of baseline spectra. TMJ sounds 
were recorded bilaterally from 40 subjects with healthy joints by means of a self 
developed recording system using miniature capacitor microphones inserted into the 
earpieces of a medical sthethoscope. The recordings were performed with no mandibular 
movements and during three consecutive opening and closing movements. Acoustic 
recordings conducted for patients without mandibular movements were below 800Hz and 
were significantly less than frequencies with mandibular movements. The baseline 
frequency spectrum could be developed in future studies to differentiate healthy subjects 
and subjects with different pathologies. 
Widmalm et al., (1992) attempted to better understand the cause of different 
sounds from the TMJ joint in 27 fresh autopsy specimens. They displayed the time 
frequency distribution of sound as a three dimensional graph and correlated the sound 
character to morphologic observations at subsequent dissection. It was also shown that all 
joints with sounds had different degrees of intra-articular changes and the reciprocal 
clicking occurred both in joints with disc displacement with and without reduction, as 
well as in joints with arthrotic changes. Crepitation occurred only in joints with arthrosis 
and perforation. A high-frequency sound appeared to be associated with arthrosis of the 
articular surfaces. It was concluded that joint sounds indicate  abnormality, but the 
absence of joint sound does not exclude intraarticular pathology. New methods can be 
developed in the future to record the sound of crepitus in the spine, which can help in 
early detection of degenerative disease using non-invasive tools. 
 13 
Reddy et al., (2001) developed the accelerometry technique to characterize 
various types of arthritis and spondyloarthropathy in knee joint. An ultra-miniature 
accelerometer was placed on the patella, and the subject was asked to rhythmically rotate 
the knee from 90 degree flexion to full extension. They concluded that mean power of 
acceleration signal in the range of 100?500 Hz is significantly different for 
spondyloarthropathy patients when compared to rheumatoid arthritis patients. 
Degenerative disease in the spine could be detected and diagnosed using new methods 
depending on the unique waveform patterns. 
Brown (1997) used a clinical heart sounds stethoscope and an electronic 
stethoscope to identify difference in sound frequency of muscles between healthy patients 
and patients with untreated Parkinson?s disease. Sound signals were analyzed 
spectrographically on a computer. Measurement of muscle discharge was made 6-18 
hours after withdrawal of antiparkinsonian medication and again 1-2 hours after 
treatment was restarted on the same day. Microphone signals from the forearm of 
untreated parkinsonian patients maximally extending the wrist showed discrete bursts of 
sound at a frequency of about 10 Hz, while increase in frequency was seen for healthy 
patients at around 40 Hz.  
Jaskolska and Madeleine (2007) conducted a study to compare mechanomyogram 
(MMG) recording using a condenser microphone (MIC) and an accelerometer (ACC), 
during submaximal isometric, concentric and eccentric contractions in 14 males. The 
reason for using ACC and MIC together was for the reason that each of the devices 
respond differently and can be used together in diagnosis in the future. The maximal 
voluntary force (MVC) of the biceps brachii was measured, and they concluded that 
 14 
during isometric, concentric and eccentric contractions of increased intensity, the results 
found from accelerometer, microphone, and MMG signals were different despite of the 
similar trends. 
 
1.2 Main Objective: 
The main objective of this research project is to locate the level of zygapophysial 
(Z) joints in the lumbar spine (L1/L2-L5/S1) and the sacroiliac joints (SIJs) that cavitate 
during side-posture positioning and spinal manipulative therapy (SMT). 
The specific goals that will contribute to achieve the objective are: 
1. To refine previously designed accelerometry techniques used to evaluate cavitations 
following SMT. 
2. To conduct recordings in a clinical investigation using the refined accelerometry 
methods. 
3. To report the results of accelerometry recordings during SMT (Group 1) and side-
posture positioning alone (Group 2) and make comparisons of cavitations 
between: Group 1 vs. Group 2; left side (up-side during SMT) vs. right side 
(down-side during SMT); and ?target? Z joints (L3/L4, L4/L5, and L5/S1) vs. 
?non-target? Z joints (L1/L2, L2/L3, and SIJs). 
4. To assess the accelerometry methods used in the clinical study to determine the 
strengths and challenges (weaknesses) of the methods. 
5. To propose ideas to further refine the accelerometry methods and to add acoustics 
methods for future studies. 
 15 
Each specific goal is discussed in the future chapters to achieve main objective of this 
research project. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 16 
2. ANATOMY AND PHYSIOLOGY OF LUMBAR SPINE 
 
 
This chapter describes the basic anatomy of the spine. A basic knowledge of 
spinal anatomy is needed to understand the research described in future chapters. The 
spine is a very important part of the body, as it helps to give the body its structure and 
support. The spine is also designed to protect the spinal cord and the nerves that carry 
information from the body to the brain and from the brain to the body. 
 
2.1 Spinal Column 
The adult spine is composed of the vertebral column and the muscles, ligaments 
and other connective tissues that attach to the vertebral column. The vertebral column is 
made up of 24 small bones called vertebrae that are stacked on each other (Figure 2.1) 
(Taiwanspinecenter, no date). In addition, the sacrum and the coccyx (found below the 
sacrum) are part of the vertebral column (Cramer and Darby, 2005). 
The spine is made up of 5 regions (Rodts, 2008):  
 Cervical / (C1- C7) 
 Thoracic / (T1- T12) 
 Lumbar / (L1-L5) 
 Sacrum 
 Coccyx 
 17 
The cervical region is the upper part of the spine and is made up of seven 
vertebrae. The thoracic region is the middle portion of the spine and is made of 12 
vertebrae. The lumbar region is the lower portion of the spine and is made of 5 vertebrae, 
but some people can have 6 lumbar vertebrae. These 3 areas are the main regions of the 
spine. The sacral region (sacrum) is situated below the lumbar spine and is also made of 
specialized vertebrae that help to connect the vertebral column with the pelvis. The 
coccyx is the lowest part of spinal column and is made of approximately 4 fused, 
undeveloped vertebrae (Cramer and Darby, 2005). The normal spine has an ?S? like curve 
when seen from side. This ?S? shape curve helps the spine to withstand all kinds of 
stresses. 
 
 
 
 18 
 
Figure 2.1: Three views of the vertebral column showing: 
A) Anterior view 
B) Lateral view  
C) Posterior view showing cervical, thoracic, lumbar, sacral, and coccyx regions 
 
Between each vertebra there is a gel-like cushion called the intervertebral disc 
(Figure 2.3) (University of Maryland Spine Center, 2007). The intervertebral discs allow 
 19 
motion to occur between vertebral segments and absorb various forces placed on the 
vertebrae. 
 
 2.2 Vertebrae 
The individual bones of the spine are called the vertebrae (Cramer and Darby, 
2005). These are the building blocks of the vertebral column. A vertebra supports and 
protects the spinal cord and also bears the majority of the weight added on the spine. A 
vertebra is made up of the following parts: (Figure 2.2) (University of Maryland Spine 
Center, 2007). The body of each vertebra is the large front portion of the vertebra.  The 
body is attached to a bony ring (posterior arch).The vertebral foramen is a hole found 
between the body and the posterior arch of each vertebra. When vertebrae are stacked on 
each other, the vertebral foramina creates a hollow tube through which the spinal cord 
passes. The posterior arch consists of several parts. The left and right pedicles and 
laminae (Figure 2.2) extend from the body to cover the vertebral foramen on each side 
and behind the spinous process (SP) is the bony portion opposite the body of the 
vertebra. One can feel the SPs as projections in the midline of the back. Several large 
processes extend from the junction of the pedicles and laminae. Two (left and right) 
transverse processes project to each side. Finally, superior articular process (top side) 
and inferior articular process (bottom side) project from the junction of the pedicle and 
lamina on the left and right side. The articular processes help to form the zygapophysial 
(Z) joints. 
 
 
 20 
 
Figure 2.2: Parts of lumbar vertebrae 
 
2.3 Zygapophysial (Z) Joints  
Zygapophysial joints are also called as facet joints. The facets are the smooth 
articulating surfaces of the superior and inferior articular processes. The facets of 
adjacent superior and inferior articular processes meet to form the zygapophysial joints 
(Figure 2.3) (University of Maryland Spine Center, 2007). The left and right Z joints 
(along with the intervertebral disc) join to adjacent vertebrae together. Without the Z 
joints, there would not be motion in the vertebral column.  
The Z joints are a type of synovial joint. The synovial joints, such as those found 
in the knees or elbows, are joints that allows movement between two bones. In a synovial 
joint, the ends of the bones are covered with a material called articular cartilage. This is a 
spongy material that allows the bones to glide against each other without much friction. 
Synovial fluid is found in the synovial joints and is important for reducing friction.  
 
 21 
 
Figure 2.3: Lumbar zygapophysial joint 
 
In addition, the Z joints help to carry loads placed on the spine, particularly during 
extension and rotation.  
 
Figure 2.4: Showing motion in Z joints during flexion and 
extension 
 22 
Two vertebrae, and the tissues that connect them, comprise the smallest working 
unit of the spine. This unit is referred to as the ?spinal motion unit?. This motion unit 
(Figure 2.4) (All About Back and Neck Pain , 2009) allows for normal flexion, extension, 
lateral bending (lateral flexion), and rotation to occur between adjacent vertebrae.  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 23 
3. ACCELEROMETERS AND MICROPHONES 
 
 
Accelerometers are an important part of the research described in the following 
chapters. Microphones are being considered for future studies. Both are described in this 
chapter. 
 
3.1 Accelerometers 
An accelerometer is an electromechanical device that measures acceleration 
forces (Bruel & Kjaer, 2010). These forces may be static, like the constant force of 
gravity pulling at your feet, or they could be dynamic, caused by moving or vibrating the 
accelerometer. 
3.2 Types of Accelerometers 
1. Piezoelectric accelerometers 
2. Piezoresistive accelerometers 
3. Capacitive accelerometers 
 
 
 
 24 
3.2.1 Piezoelectric Accelerometers  
Piezoelectricity is the ability of some materials (notably crystals and certain 
ceramics) to generate an electric field or electric potential in response to an applied 
mechanical strain (ratio of change in volume to original volume) (Gautschi, 2002). The 
effect is closely related to a change of polarization density within the material's volume. 
If the material is not short-circuited, the applied stress/strain induces a voltage across the 
material (Figure 3.1) (Wikipedia, no date). However, if the circuit is closed, the energy 
will be quickly released. So, in order to run an electric load (such as a light bulb) on a 
piezoelectric device, the applied mechanical stress must oscillate back and forth. The root 
?piezo? is derived from the Greek piezo or piezein, which means to squeeze or press 
(Avallone and Baumeister, 1996). The piezoelectric effect is reversible. Materials 
exhibiting the direct piezoelectric effect (the production of an electric potential when 
stress is applied) also exhibit the reverse piezoelectric effect (the production of stress 
and/or strain when an electric field is applied). 
 
Figure 3.1: A piezoelectric disk generates voltage when 
deformed 
 25 
A piezoelectric accelerometer (such as those used in this study) utilizes the 
piezoelectric effect of certain materials to measure dynamic changes in mechanical 
variables (E.g. acceleration, vibration, and mechanical shock).  
Piezoelectric accelerometers convert one form of energy into another and provide 
an electrical signal in response to a quantity, property, or condition that is being measured 
(Figure 3.2) (Wikipedia, no date). In this instance, acceleration acts upon a seismic mass 
that is restrained by a spring or suspended on a cantilever beam, and converts a physical 
force into an electrical signal (Dimension Engineering, no date). Before the acceleration 
can be converted into an electrical quantity, it must first be converted into either a force 
or displacement. 
Piezoelectric accelerometers rely on piezoceramics (e.g. lead zirconate, titanate, 
etc) or single crystals (e.g. quartz, tourmaline, etc). They are unmatched in terms of their 
upper frequency range, low packaged weight and high temperature range. They are 
frequently used in aerospace, ballistics, engine testing and other applications (Aszkler, 
2005). 
       
          A) Before acceleration.                              B) After acceleration. 
Figure 3.2: Piezoelectric accelerometer showing no voltage when acceleration is not 
applied and voltage generated with acceleration 
 26 
3.2.2 Piezoresistive Accelerometers 
The piezoresistive effect describes the changing electrical resistance of a material 
under applied mechanical stress (Avallone and Baumeister, 1996). In contrast to the 
piezoelectric effect, the piezoresistive effect only causes a change in resistance and does 
not produce an electric potential. 
The sensitivity of piezoresistive devices is characterized by the gauge factor 
(Avallone and Baumeister, 1996): 
 
Where dR is the change in resistance due to deformation, R is the undeformed resistance 
and is the strain 
Piezoresistive accelerometers may be fabricated from metal strain gauges, 
piezoresistive silicon, or as a MEMS (Micro Electro Mechanical Systems) device. In such 
designs, a resistive material is typically bonded to a cantilever beam that bends under the 
influence of acceleration. This bending causes deformation of the resistor, leading to a 
change in its resistance (Figure 3.3) (PCB Piezotronics, 2009). The resistors are normally 
configured into a Wheatstone bridge circuit which provides a change in output voltage 
proportional to acceleration. Piezoresistive accelerometers are capable of measuring 
constant, transient, and periodic acceleration. Piezoresistive accelerometers are preferred 
in high shock applications. 
 
 27 
 
Figure 3.3: Piezoresistive accelerometer 
 
3.2.3 Capacitive Accelerometers 
Capacitive accelerometers utilize the properties of an opposed plate capacitor for 
which the distance between the plates vary proportionally to applied acceleration, thus 
altering capacitance (Figure 3.4) (PCB Piezotronics, 2009). This variable is used in a 
circuit to deliver a voltage signal that is proportional to acceleration. Capacitive 
accelerometers are capable of measuring constant as well as slow transient and periodic 
acceleration. These accelerometers are suited for measuring low frequency vibration, 
motion, and steady-state acceleration. 
 
 
Figure 3.4: Capacitive accelerometer 
 28 
 Capacitive accelerometers typically use a silicon micro-machined sensing 
element. Their performance is superior in the low frequency range, and they can be 
operated in servo mode to achieve high stability and linearity. 
 
3.3 Applications  
3.3.1 Medical Applications 
Within the last several years, Nike, Polar and other companies have produced and 
marketed sports watches for runners that include foot pods that contain different kinds of 
accelerometers that determine the speed and distance for the runner wearing the unit. 
3.3.2 Transport 
One of the most common uses for MEMS accelerometers is in airbag deployment 
systems for modern automobiles. In this case, the accelerometers are used to detect the 
rapid negative acceleration of the vehicle to determine when a collision has occurred and 
the severity of the collision (Bruel & Kjaer, 2010). Piezoresistive accelerometers are also 
commonly used in detecting crashes. Another common automotive use is in electronic 
stability control systems, which uses a lateral accelerometer to measure cornering forces. 
The widespread use of accelerometers in the automotive industry has pushed their cost 
down dramatically. Another automotive application is the monitoring of noise, vibration 
and harshness (NVH), conditions that cause discomfort for drivers and passengers and 
may also be indicators of mechanical faults. 
 
 29 
3.3.3 Consumer Electronics 
Accelerometers are increasingly being incorporated into personal electronic 
devices such as smart phones, audio players, video game consoles, camcorders etc. 
 Motion Input 
Some smart phones, digital audio players and personal digital assistants contain 
accelerometers for user interface control; often the MEMS accelerometer is used to 
present landscape or portrait views of the device's screen, based on the way the device is 
being held. 
Nintendo's Wii video game console uses a controller called a Wii Remote that 
contains a three-axis MEMS accelerometer (Analog Devices, 2010) and was designed 
primarily for motion input. Users also have the option of buying an additional motion-
sensitive attachment, the Nunchuk, so that motion input could be recorded from both of 
the user's hands independently. 
The Nokia 5500 sport features a 3D accelerometer that can be accessed from 
software. It is used for step recognition (counting) in a sport application, and for tap 
gesture recognition in the user interface. Tap gestures can be used for controlling the 
music player and the sport application, for example to change to next song by tapping 
through clothing when the device is in a pocket (Nokia, 2010).  
 30 
Orientation Sensing 
A number of modern notebook computers use 3 axis MEMS accelerometers to 
automatically align the screen depending on the direction the device is held, i.e. switching 
between portrait and landscape modes. For example, Apple uses an LIS302DL 
accelerometer in the iPhone, iPod Touch and the 4th&5th generation iPod Nano allowing 
the device to know when it is tilted on its side (Stmicroelectronics, 2010).  
 Image Stabilization 
Camcorders use accelerometers for image stabilization. Still cameras use 
accelerometers for anti-blur capturing. The camera holds off snapping the CCD "shutter" 
when the camera is moving. When the camera is still (if only for a millisecond, as could 
be the case for vibration), the CCD is "snapped".  
3.4 Microphones 
A microphone (mic) is an acoustic-to-electric transducer or sensor that converts 
sound into an electrical signal (PCB Piezotronics, 2009). In 1876, Emile Berliner 
invented the first microphone used as a telephone voice transmitter. 
The sensitive transducer element of a microphone is called the element or capsule. 
A complete microphone also includes a housing, which brings the signal from the 
element to other equipment. 
 
 
 
 31 
3.5 Types of Microphones 
1. Condenser Microphones 
2. Dynamic Microphones 
3. Piezoelectric Microphones 
4. Fiber Optic Microphones 
 
3.5.1 Condenser Microphones 
Condenser microphone is also called a capacitor microphone or electrostatic 
microphone. A capacitor is an electronic component which stores energy in the form of 
an electrostatic field.  
A capacitor has two plates with a voltage between them. In the condenser 
microphone, one of these plates is made of very light material and acts as the diaphragm 
(Figure 3.5) (Media College, no date). The diaphragm vibrates when struck by sound 
waves; changing the distance between the two plates and therefore changing the 
capacitance (Sessler and West, 1962). When the plates are closer together, capacitance 
increases and a charge current occurs. When the plates are further apart, capacitance 
decreases and a discharge current occurs. Nearly all cell-phone, computer, PDA and 
headset microphones are condenser microphones. 
 32 
 
Figure 3.5: Cross section of condenser microphone 
 
3.5.2 Dynamic Microphones 
Dynamic microphones are versatile and ideal for general usage. They use a simple 
design with few moving parts. They are relatively sturdy and resilient to rough handling. 
They are also better suited for handling high volume levels, such as those from certain 
musical instruments or amplifiers. They have no internal amplifier and do not require 
batteries or external power. 
When a magnet is moved near a coil of wire, an electrical current is generated in 
the wire. Using this electromagnet principle, the dynamic microphone uses a wire coil 
and magnet to create the audio signal (Figure 3.6) (PCB Piezotronics, 2009). The 
diaphragm is attached to the coil. When the diaphragm vibrates in response to incoming 
sound waves, the coil moves backwards and forwards past the magnet. This creates a 
current in the coil which is channeled from the microphone by wires. 
 33 
 
Figure 3.6: Cross section of Dynamic Microphone 
 
3.5.3 Piezoelectric Microphones 
Piezoelectric microphones work on the principle of the piezoelectric effect as 
described earlier in piezoelectric accelerometers (Section 3.2.1). Piezoelectric 
microphones use ceramic or quartz crystals (or other piezoelectric materials) linked to a 
diaphragm or directly exposed to acoustic waves (Figure 3.7) (PCB Piezotronics, 2009). 
Stresses in the crystals, resulting from a sound field, generate an output proportional to 
the acoustic pressure. Many designs incorporate a built-in preamplifier next to the crystal. 
This arrangement reduces the electrical noise and output impedance. Piezoelectric 
microphones are commonly used in medical and aero acoustic applications because of 
their high sensitivity (Horowitz et al., 2006). 
 
 34 
 
Figure 3.7: Cross section of Piezoelectric Microphone 
Where: 
AP: Acoustic pressure. 
Uo: Output voltage. 
1) Diaphragm. 
2) Piezoelectric material. 
3) Built in preamplifier. 
4) Case. 
 
3.5.4 Fiber Optic Microphones 
A fiber optic microphone converts acoustic waves into electrical signals by 
responding to changes in light intensity, rather than changes in capacitance or magnetic 
field (Paritsky and Kots, 1997).  
Fiber optic microphones use light from a laser source travelling through an optical 
fiber to illuminate the surface of a tiny, sound-sensitive reflective diaphragm. Sound 
causes the diaphragm to vibrate, thereby minutely changing the intensity of the light it 
reflects. The modulated light is then transmitted over a second optical fiber to a photo 
detector which transforms the intensity-modulated light into analog or digital audio for 
transmission or recording.  
 35 
Fiber optic microphones do not react with any electrical, magnetic, electrostatic or 
radioactive fields. The fiber optic microphone design, therefore, is ideal for use in areas 
where conventional microphones are ineffective or dangerous, such as inside industrial 
turbines or in magnetic resonance imaging (MRI) equipment environments. 
 
3.6 Applications of Microphones in Research and Industry 
3.6.1 Research and Product Design 
Since excessive sound pressure can cause damage to products or human hearing, 
microphones are used to measure the pressure level exerted on a surface. Sound 
measurement is used in a variety of applications including the study of door slams, clutch 
engagements, starter impact and sunroof noise. Engine noise in a cabin or car interior and 
sound exhibited from consumer appliances is tested and analyzed to extend the lifespan 
of the product and minimize external noise (Bruel & Kjaer, 2010). 
Increased sound levels or changes in frequency can indicate that a product is not 
working to its capacity. Motors, gears, bearings, blades, and other industrial components 
can all experience changes in decibel level or frequency shift when not working properly. 
High precision microphones can be utilized to confirm that a product is experiencing a 
problem, or can be used to predict failure of a component (PCB Piezotronics, 2009). 
 
 
 
 36 
3.6.2 Compliance 
Microphone tests can be performed and recorded for verification of pressure 
levels on products, and can be utilized in legal situations. Companies can use high 
precision microphone tests for proof of sound pressure levels during design. Microphones 
are used on sound level meters to ensure compliance with national standards for shop 
noise (Extech Instruments, 2010). 
3.6.3 Environmental Noise Analysis 
There are certain sound pressure levels that the human ear can only be subjected 
to for specific amounts of time before ear damage occurs, such as industrial shop, airport, 
and automotive highway noise. Acoustic testing is performed to better understand the 
sound levels that are experienced in these surroundings and make adjustments to provide 
greater personal protection (Extech Instruments, 2010).  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 37 
4. INSTRUMENTATION AND METHODOLOGY 
 
 
This chapter describes the instrumentation for accelerometry equipment used in 
this study and our methodology. The methods begin with a description of the process 
used to refine previous methods, followed by descriptions of the clinical study for Group 
1 and Group 2. 
 
4.1 Instrumentation 
Nine accelerometers (Bruel & Kjaer Delta Tron4507, Naerum, Denmark) (Figure 
4.11) were used to pick up vibration data created by cavitations during side-posture 
positioning and SMT. Direction of acceleration (Figure A.1., Appendix section A.1) was 
also considered for all the accelerometers during clinical study. These piezoelectric 
accelerometers were used because cavitations stimulated them at their natural frequency 
of 21 KHz. Three signal conditioning amplifiers (Bruel & Kjaer 2693, Naerum, 
Denmark) (Figure A.2, Appendix section A.2) (Bruel & Kjaer, 2010) were used to 
amplify the signal from accelerometers. Each amplifier was capable of handling data 
from 4 different channels/accelerometers (Figure 4.1). The analog signal was converted 
into digital information at a maximum rate of 10MHz (10 million samples per second) by 
two NI PXIe 6356 (National Instruments, Austin, Texas, USA) (Figure A.3, Appendix 
section A.3) (National Instruments, 2010) data acquisition (DAQ) devices. The velocity 
 38 
of vibration waves from cavitations travelling through the body is at 1400m/sec (Ross et 
al., 2004). To get distance discrimination of 0.5cm the sampling rate was set to  
320,000Hz. Given that Z joints between adjacent vertebrae and on the left to right sides 
are separated by a distance of 3 to 4.5 cm, there was no reason to sample at a rate any 
greater than this, even though the equipment had the potential to do so.  Since the study 
required nine analog channels, we used two data acquisition devices, each capable of 
handling data from eight simultaneous channels. Each device was capable of handling a 
maximum sampling rate of 1.25 MS/s/channel (million samples per second per channel). 
These devices included multithreaded NI-DAQmx (National Instruments, Austin, Texas) 
driver software, which was compatible with the following versions (or later) of NI 
application software: LabVIEW 8.2, Lab Windows?/CVI 7.x, or Measurement Studio 
7.x; LabVIEW Signal Express 1.x; and LabVIEW with the LabVIEW Real-Time Module 
8.2. 
LabVIEW 9.0 was used in this study to write different programs for data 
acquisition and storage and data analysis. 
Two NI BNC 2090 connector blocks (Figure A.4, Appendix section A.4) 
(National Instruments, 2010) were used to connect data from the amplifier to DAQ 
devices (Figure 4.1). A NI PXIe 1073 (Figure A.5, Appendix section A.5) (National 
Instruments, 2010) chassis was used to house both high speed DAQ devices. This chassis 
was capable of accepting PXI Express modules in two slots and supporting standard PXI 
hybrid-compatible modules in up to three slots. This chassis was also used to supply 
 39 
power for the system. A NI PXIe-Express Card 8360 (Figure A.6, Appendix section A.6) 
(National Instruments, 2010) was used to connect to a Dell Precision M6400 laptop via 
an Express Card MXI cable connected to slot 1 of the chassis. All equipments were 
placed on a portable cart (Figure 4.2). This configuration provided the portability to re-
locate the entire system and the flexibility to unplug the components. This way, for 
example, the computer could be used in a different location for data analysis, etc. 
The system was powered up and powered off in the following sequential manner: 
Powering up 
1. Amplifiers 
2. PXIe chassis 
3. Laptop 
 
Powering off 
1. Laptop 
2. PXIe chassis 
3. Amplifiers 
 
 
 40 
 
Figure 4.1: Block diagram of the system 
Where: 
Connector Block 1, 2: NI BNC 2090 
DAQ 1, 2: NI PXIe 6356 
Chassis: NI PXIe 1073 
Express Card: NI PXIe- Express Card 8360 
 
 
 41 
 
Figure 4.2: Instrument located on a portable cart/trolley 
 
 
 
 
 
 
 
 
 
 
 42 
4.2 Refining Previously Designed Accelerometry Techniques Used to Evaluate 
Cavitations Following Side-Posture Positioning and SMT 
 
4.2.1 Previously Designed Accelerometry Methods 
Accelerometers were previously used to identify cavitations originating from the 
left and right L3/L4, L4/L5, and L5/S1 Z joints during side-posture positioning and SMT. 
The original plan for the current study was to tape seven accelerometers to the subject?s 
skin. Three accelerometers would be placed over the spinous processes (SPs) of the third 
(L3), fourth (L4), and fifth (L5) lumbar segments (Figure 4.3). Another would be placed 
3 cm to the left and another 3 cm to the right mid-distance between the SP L3 and SP L4 
midline accelerometers. Finally, one more would be placed 3 cm to the left and another 3 
cm to the right mid-distance between the SP L4 and SP L5 midline accelerometers 
(Figure 4.3).  
 
 
 
 43 
 
Figure 4.3: Position of accelerometers as per the original 
(previous) design 
 
 
 
 
 
 44 
4.2.2 Re-Design of the Accelerometry Methods 
Several problems were identified during initial testing. Differentiating cavitations 
originating in segments immediately above the SP L3 accelerometers or SIJ cavitations 
below the SP L5 accelerometer from L3/L4 or L5/S1 cavitations was very difficult. A 
specific configuration of nine accelerometers resolved these challenges. Seven 
accelerometers were placed over the spinal column of a subject?s back at the levels of the 
SPs of the first (L1), second (L2), third (L3), fourth (L4), fifth (L5) lumbar segments, and 
the S1 and S2 (sacral tubercles) (Figure 4.4). Another was placed 3 cm to the left and 
another 3 cm to the right mid-distance between the SP L4 and SP L5 midline 
accelerometers. This configuration was designed to identify the specific Z joints from 
which cavitations originated by identifying the order in which the recordings of 
individual accelerometers deviated from the baseline. 
 45 
 
Figure4.4: Position of accelerometers in the re-designed 
method 
 
 
 
 
 
 
 
 
 46 
4.3 Testing the System 
 
4.3.1 Re-Designed Method 
Accelerometers were taped on top of a table in the same configuration as if they 
were to be taped on the spine at SPs L1, L2, L3, L4, L5, S1 and S2. These accelerometers 
were separated by 3cm. Two accelerometers were also placed lateral to the midline 
accelerometers, one 3 cm to the left and one 3 cm to the right of the mid-distance 
between the SPs L4 and L5 midline accelerometers (Figure 4.5). 
A target area was marked and a steel ball was dropped in between accelerometers 
at SPs L4, L5, and Left L4L5 and the data was collected and analyzed. In this 
experiment, the accelerometer at SP L4 or SP L5 would respond first, then Left L4L5, 
Right L4L5 and then accelerometers sitting above and below SPs L4 and L5 (i.e., SP L3 
and S1) would respond. The order of the response would be based upon how the vibration 
wave reached the individual accelerometers. The experiment was repeated several times 
at different locations. 
 
 
 
 
 
 
 
 
 47 
 
 
Figure 4.5: Accelerometers taped on the table according to the re-
designed system. Each accelerometer was separated by a 
distance of 3cm 
 
 
 48 
 
Figure 4.6: The complete time scale of the collected data during dropping of the ball 
at the Left L4L5 location for redesigned system. The legend above the graph 
shows the different colors assigned to each accelerometer 
 
 
 49 
 
Figure 4.7: The timeline and amplitude (X-axis and Y-axis) of Figure 4.6 have been 
expanded (zoomed in during the time period of 3.101734 to 3.10224 sec) 
showing the sequence in which each of the accelerometer responded for 
redesigned system 
  
Accelerometers responded in the following order: Left L4L5, SP L4, SP L5, Right 
L4L5, SP L3, SP L2, S1, SP L1 and S2 (Figure 4.7). As Figure 4.7 shows, the experiment 
successfully determined that the ball was dropped in between Left L4L5, SP L4, and SP 
L5 accelerometers. The experiments reported at other locations were equally successful. 
 
 
 
 
 
 50 
 
4.3.2 Comparison with Previously (Originally) Designed Method 
Accelerometers were taped on top of a table in the same configuration as if they 
were to be taped on the spine at SPs L3, L4, and L5 (Figure 4.8). One accelerometer was 
taped 3 cm to the left and another 3 cm to the right mid-distance between the SP L3 and 
SP L4 midline accelerometers. Finally, one more accelerometer was taped 3 cm to the left 
and another 3 cm to the right mid-distance between the SP L4 and SP L5 midline 
accelerometers (Figure 4.8).  
A target area was marked and a steel ball was dropped in between accelerometers 
at SPs L4, L5, and Left L4L5 and the data was collected and analyzed. According to the 
hypotheses, if cavitation occurred, or a ball was dropped as in our example, in the 
designated area then the accelerometer at SP L4 or SP L5 should respond first (or both 
would respond approximately at the same time), followed by the Left L4L5 and then 
other accelerometers. Figures 4.8-4.10 show the methods and results of this experiment. 
Several trials at different locations were also performed. 
Y axis for all the collected cavitation data shows the change of voltage generated 
in the piezoelectric accelerometers used. 
 51 
  
Figure 4.8: Accelerometers taped on the table according to the previously designed 
method. SPs L3, L4 and L5 accelerometers were separated by a distance of 3cm 
 52 
 
Figure 4.9: The complete time scale of the collected data during dropping of the ball 
at the Left L4L5 location for the previous (original) designed system. The 
legend above the graph shows the different colors assigned to each 
accelerometer. 
 
 
 53 
 
Figure 4.10: The timeline and amplitude (X-axis and Y-axis) of Figure 4.9 have 
been expanded (zoomed in during the time period of 2.784192 to 2.78455 
sec), showing the sequence in which each of the accelerometer responded 
for previous (original) design 
 
Accelerometers responded in the following order: SP L5, SP L4, Left L4L5 and 
Left L3L4 (both at same time), SP L3, and then Right L4L5 and Right L3L4 (both at 
same time) (Figure 4.10). It was difficult to determine where the ball was dropped based 
on the graph. The graphs of other trials were equally difficult to interpret. 
Consequently, the new re-designed nine accelerometers method was used in the 
clinical study. The new method also allowed identification of cavitations at targeted 
(L3/L4, L4/L5, L5/S1) and non-targeted Z joints. The targeted Z joints were used in 
another clinical study (not reported here) in which MRI scans were taken of L3/L4, 
L4/L5 and L5/S1 Z joints. Therefore, being able to identify cavitations from targeted and 
 54 
non-targeted Z and SIJ joints was important. The new design was capable of detecting 
cavitations at Left L1/L2, Right L1/L2, Left L2/L3, Right L2/L3, Left L3/L4, Right 
L3/L4, Left L4/L5, Right L4/L5, Left L5/S1, Right L5/S1, Left SI joint and Right SI 
joints. 
  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 55 
 
4.4 Clinical Study 
The new methods were used in the clinical study. A total 40 subjects were 
included in the IRB approved study and randomized into one of the two groups. These 
groups were:  
Group 1, SMT Group (total of 30 subjects were randomized into this group) who 
received SMT; and  
Group 2, Control Group (total of 10 subjects were randomized into this group) 
who were just held in side-posture position but didn?t receive SMT. Data was collected 
from accelerometers with subjects in Group 2, because it was assumed and found that 
some Z joints cavitate when subjects are held in the side-posture position. 
Specific programs written using LabVIEW were used for collecting and analyzing 
data. Frank Balester from NUHS wrote these programs with the aid of technical support 
from National Instruments. Writing separate programs added flexibility to the process of 
collecting and storing data into the computer RAM in the approximate format and 
analyze it at anytime. 
The procedures used in the clinical trial were as follows: 
1. After initial screening (phone screening), consenting subjects were 
scheduled for examination. Examination was performed against the inclusion and 
exclusion criteria for this research project. Eligible subjects were scheduled for an MRI 
appointment.  
2. Location of SP L4 was marked using grease pencil before the first MRI 
scan and high signal marker was taped to the subject?s skin, at marked location by trained 
 56 
clinician, while the subject was lying on their stomach. Targeted area of the MRI scan 
was L3/L4, L4/L5 and L5/S1 area. 
3. The subject was asked to lay on his or her back and first MRI scan was 
taken in neutral position. During the first MRI scan, location of SP L4 was identified by 
the high signal marker (distinct bright white appearance) on MRI. This procedure was 
used to verify the SP L4. The SP L4 was used as the primary landmark from which all of 
the accelerometers were placed at their designated locations. 
4. A radiologist checked first MRI scan of the subject for any pathology 
(required for other study not discussed here) against the inclusion, exclusion criteria. 
Accelerometry and second MRI scan was proceeded only for eligible subjects and those 
who were ineligible were excluded from the study. 
5. After the first MRI scan, the eligible subject was pulled out of the MRI 
unit (subject remained on the MRI gantry table) and asked to lay on his/her stomach 
again. A trained research assistant (tenth trimester chiropractic intern) then used a grease 
pencil to mark the remaining levels of the SPs, SIJs and the left and right locations (mid-
distance between SPs L4 and L5) where accelerometers were to be placed. 
6. The Auburn University Graduate Student Preetam Bora (PB), who was 
running the accelerometry equipment, helped the research assistant tape accelerometers 
to the subject?s back starting with the levels of the first (L1), second (L2), third (L3), 
fourth (L4), fifth (L5) SPs of the lumbar segments, and S1 and S2 (sacral tubercles). 
Another was placed 3 cm to the left and lastly another one 3 cm to the right mid-distance 
between the SPs L4 and L5 midline accelerometers (Figures 4.11-4.15).  
7. Each accelerometer was specifically identified (i.e., the same 
 57 
accelerometer was always used for the SP L1, the left accelerometer between SPs L4 and 
L5, etc.). For this, each accelerometer was specified with a unique number to make sure it 
was placed correctly at the desired location (Figure 4.11). 
 
 
Figure 4.11: Unique number of the accelerometer marked and the same 
accelerometer placed in its box (notice corresponding label) for error-
resistant methodology 
 
Hooks were placed behind the cart to roll wires for each accelerometer and 
labeled according to their number (Figure 4.12). This reduced the amount of time 
required to tape the accelerometers to the subject?s back and lessened the chance of a 
mistake. The same number was marked on the box of each accelerometer, ensuring the 
correct accelerometer was used for each location on the subject?s back. This correct 
labeling also ensured the accelerometers were placed back in the correct box after data 
 58 
collection (Figure 4.11, 4.13). The same number was also marked on the amplifier?s input 
and output ports (Figure 4.14). 
 
 
Figure 4.12: Hooks with respective accelerometer number marked on 
each, with properly rolled up wires attaching to corresponding 
accelerometers 
 
 59 
 
Figure 4.13: Accelerometers placed in marked boxes 
 
Figure 4.14: Amplifiers marked with respective accelerometer number 
attached to input (I/P) and output (O/P) ports 
 
 60 
Table 4.1 was prepared and posted in the MRI room where the study was 
conducted. The table included the assigned number for each accelerometer and the 
designated location where it should be placed on the subject. This sequence was followed 
while taping accelerometers on the subject?s back for error-proof methodology 
throughout the study. The table also shows the color assigned to each accelerometer on 
the LabVIEW program. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 61 
Table 4.1: Assigned number for each accelerometer and the designated 
placement location on the subject 
 
ACCELEROMETER 
 
POSITION 
 
COLOR 
 
1 
 
SP L1 
 
WHITE 
 
2 
 
SP L2 
 
RED 
 
3 
 
SP L3 
 
GREEN 
 
4 
 
SP L4 
 
LIGHT 
BLUE 
 
5 
 
SP L5 
 
YELLOW 
 
6 
 
S1 
 
PURPLE 
 
7 
 
S2 
 
ORANGE 
 
8 
 
LEFT L4L5 
 
DARK 
BLUE 
 
9 
 
RIGHT L4L5 
 
PINK 
 
 
 
 62 
8. After all the accelerometers were taped on the subject?s back (Figure 
4.15), he/she was asked to slowly lay on his/her right side slowly without cavitating their 
back (it was assumed and found that cavitation occurs when subjects are held in side-
posture positioning). At the same time, the subject was randomized into either Group 1 
(subject was placed on right side and received SMT on left side) or Group 2 (subject was 
placed on right side and held in side-position without receiving SMT). 
A unique Subject ID number was given to each subject, and data was collected 
and stored into computer RAM using this ID number as file name. 
 
 
Figure 4.15: All nine accelerometers taped in the assigned position on a 
subject?s back 
 
9. Data was collected and stored in computer memory for analysis at a later 
time. 
 63 
During methods testing procedures on mock subjects, we found that collecting 
data at the high sampling rate (320,000 Hz) from the nine accelerometers created a buffer 
issue in the computer, leading to a 5 sec limit on data collection. Collecting data in this 
time window required a great amount of understanding between the clinician (Dr. Scott 
Selby); PB and the trained research assistant (Adam Habeck), who was responsible for 
recording the time of first thrust using a stop watch.  
Two different procedures were practiced and used, depending into which group 
the subject was randomized (see next section).  
 
 
4.5 Data Acquisition for Subjects Randomized into Group 1 
 Unique Subject ID number was used as file name for data collected from 
each subject. 
 When the clinician was ready to give side-posture SMT to the subject, he 
would say ?ready? loudly. 
 PB would click on the ?run? button (Shown as 1 in figure 4.16) and then 
click the ?collect? button (Shown as 2 in figure 4.16) and say ?go? simultaneously. The 
run button was used to see if the system was running perfectly and the actual collection of 
data started after clicking on the ?collect? button. 
 The clinician would apply two rapid thrusts using a common SMT 
procedure (hypothenar ilium) and then say ?stop? when he was done with the SMT. PB 
would then click the ?stop? button to stop collecting data (Shown as 3 in figure 4.16). 
Two thrusts were applied to increase the likelihood of producing a cavitation. 
 64 
 During the procedure, the research assistant (AH) recorded time using a 
stop watch, beginning when PB said ?go? and stopping when he saw the clinician starting 
to apply the first thrust on the subject?s body. This time was recorded to get the time of 
initiation of the first thrust of the SMT. 
 
 
 
Figure 4.16: LabVIEW data acquisition window 
 65 
 
Figure 4.17: Spinal manipulative therapy (SMT) performed while recording from 
accelerometers 
 
4.6 Data Acquisition for Subjects Randomized into Group 2 
Procedure was slightly modified to accommodate for the unique attributes of 
Group 2. 
 Unique Subject ID number was used as file name for data collected from 
each subject. 
 When the clinician was ready to hold the subject in side-posture position, 
he would say ?ready? loudly. 
 PB would click on the ?run? button (Shown as 1 in figure 4.16) and then 
click the ?collect? button (Shown as 2 in figure 4.16) and say ?go? simultaneously. The 
 66 
run button was used to see if the system was running perfectly and the actual collection of 
data started after clicking on the ?collect? button. 
 The research assistant recorded time using a stop watch, beginning when 
PB said ?go? and stopping when he saw the time of three seconds on the stop watch. He 
would say ?stop? loudly and then PB would click on the stop button (Shown as 3 in 
figure 4.16) to stop collecting data.  
 
These procedures ensured that data collection did not exceed 5 seconds for any 
subject (i.e., the recording time was approximately the same for all subjects in both 
groups). 
 When the collection of data was completed for subjects either in Group 1 
or Group 2, accelerometers were removed from the subject. The research assistant asked 
the subject and clinician whether each heard a cavitation during side-posture positioning, 
first thrust, and second thrust, and their responses (blinded to each others) were recorded. 
The second MRI scan was taken in side-posture position with the left side up (the second 
MRI scan was part of the other study).  
To save the subject?s time, all the accelerometers were taken outside the MRI 
room after they were removed from the subject?s back, so the second MRI scan could 
start immediately. Alcohol wipes were used to clean the accelerometers, the wires were 
rolled, and the accelerometers were arranged on the cart for the next subject. 
 
 
 67 
4.7 Information Collected for Subject Data Base during the Accelerometry 
Procedures 
The following points were noted in each subject?s file for later references (after 
data analysis was completed). 
 Randomization in Group 1 or Group 2 
 Time  
 Time of first thrust for subjects randomized into Group 1  
OR 
     Total time of side posture positioning for subjects randomized into Group 2. 
Responses to questions asked by research assistant to the subject and clinician regarding 
whether or not each heard a cavitation during side-posture positioning, first thrust, and 
second thrust.  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 68 
5. DATA ANALYSIS 
 
 
The accelerometry data was analyzed with a program written by Frank Balester in 
LabVIEW 2009 using technical support from National Instruments (Figure 5.1). This 
program provided the flexibility of analyzing collected data (stored in computer memory) 
at any time. 
 
 
Figure 5.1: LabVIEW data analyzing window before accelerometry data was loaded
 69 
Data was stored in the computer using Subject ID numbers and opened in the data 
analysis program. Data was acquired from the computer RAM, and displayed for all the 
nine accelerometers simultaneously, but the recording of each accelerometer was 
identified by a unique color. The colors assigned to each of the accelerometers were 
shown on the color pallet to the left of the waveform graph window (Figure 5.1). This 
color pallet also showed the relative position of the accelerometer?s placement on the 
subject?s spine. 
 
5.1 Training in Data Analysis 
Dr. Kim Ross (an expert in accelerometry as it applies to SMT) (Ross et al., 2004) 
trained the Auburn University Graduate Student (PB) to detect cavitations from noise or 
artifacts during several visits to NUHS. Data was also collected from mock subjects (test 
subjects) during this time, and we analyzed the collected data together to identify 
cavitations. 
Cavitations were recognized by their unique ?non-continuous? (Macquarie 
University, 2008) waveform pattern (Figure 5.3B) that differentiated them from noise or 
artifacts created by different sources which will be discussed later in this chapter. 
Accelerometers responded sequentially to cavitations depending on the location of its 
origin (Figure 5.3C). For example, if the cavitation occurred at the Left L1/L2 Z joint 
(Figure 5.2); accelerometers responded or ?fired? in following sequence: 
 SP L1 or SP L2 
 SP L2 or SP L1 
 SP L3 
 70 
 SP L4 
 SP L5 
 SP S1 
 SP S2 
 Left L4/L5 and Right L4/L5 accelerometer may respond or ?fire? at any 
instance between the accelerometers placed on SP L1 to S2. But if the cavitation occurred 
at the left Z joint then the Left L4/L5 accelerometer responded first, followed by the 
Right L4/L5 accelerometer. 
 
Figure 5.2: Cavitation at Left L1/L2 Z joint as 
demonstrated in the recordings of Figure 5.3 
 71 
 
 
 72 
 
Figure 5.3:  Figure A shows the complete time scale of data collected during side-posture 
positioning and SMT with the subject laying on their ride side. Figure B is an 
expanded timeline of Figure A (zoomed in during the time period of 1.5380 to 
1.5387 sec), showing the unique waveform pattern of a cavitation. Figure C 
shows a timeline that is further expanded from Figure B (zoomed in during the 
time period of 1.538055 to 1.5381 sec) with accelerometers responding in the 
following sequence: SP L1, SP L2, SP L3, LEFT L4L5, SP L4, SP L5, RIGHT 
L4L5, S1 and S2. As there is only one accelerometer placed on left and right side 
of the SP, the sequence of responses from the accelerometers indicates that the 
cavitation occurred at the Left L1/L2 Z joint 
 
 73 
Analyzing the data for the presence of cavitations (Figure 5.3B) required 
evaluating the entire time line from start to end (Figure 5.3A). Once a cavitation was 
identified, the sequential response or ?firing? of the accelerometers was determined 
(Figure 5.3C). 
 
5.2 Data Analysis by Two Observers 
PB analyzed the data of each subject and made notes on cavitations. These notes 
included the time of occurrence of each cavitation, the sequential response or ?firing? of 
accelerometers and PB?s hypothesis of the Z joint where the cavitation occurred. 
Microsoft Windows Live Meeting software was used to discuss the analyzed data 
with Dr. Kim Ross, who was in Toronto, Canada. This software had the capability of 
sharing data from each other?s desktop. The program also had audio and video capability 
and could add as many attendees or presenters as needed (Figure 5.4).   
Windows Live Meeting helped in analyzing and discussing the data with Dr. 
Ross, who was able to see the same screen and data PB was viewing even though they 
were separated by several hundred miles. 
 
Data analyzed by PB was discussed with Dr. Ross using Microsoft Windows Live 
Meeting in the following way: 
 Live Meeting was scheduled with Dr. Ross. 
 A subject?s data was opened (using the Subject ID number) from the 
LabVIEW program and shared using Windows Live Meeting (Figure 5.5 B and C). Dr. 
 74 
Ross was able to see the shared data from LabVIEW on his screen even though he didn?t 
have LabVIEW installed on his computer.  
 Control was handed to Dr. Ross using a tool in Windows Live Meeting, so 
he and PB both had full control of the shared content. This allowed Dr. Ross to zoom in, 
zoom out etc., and both observers were able to see each other?s cursor moving and the 
changes (e.g. Zooming) taking place on each other?s screen. 
 PB would zoom in on the area where a cavitation occurred and compare 
opinions with Dr. Ross. Together they would decide whether or not the recordings 
indicated a cavitation, and if so, Z joint where the cavitation occurred. 
 
 
 
Figure 5.4: Microsoft Windows Live Meeting, showing number of persons in the 
meeting (2) and the content to be shared in the meeting 
 75 
 
 
 
 
Figure 5.5: A and B shows LabVIEW data being shared using Windows Live Meeting. A 
show shared data viewed by PB and B shows shared data viewed by Dr. Ross in 
Canada. 
 
 76 
5.3 Amplitude of Cavitation Waves 
Each analyzed cavitation was categorized under one of the following amplitude 
levels. 
1. High 
2. Medium  
3. Low 
 
Confidence level of cavitation was based on the amplitude of its recorded waves. 
If the intensity of cavitation itself was high, then the cavitation was assumed to have 
created stronger vibration waves. Accelerometers responding to these stronger vibration 
waves created signals of larger amplitude and responded sequentially and clearly.  
Amplitude of the signals decreased for cavitations categorized as medium or low. 
Quantifying the amplitudes of the nine accelerometers to assign an overall 
amplitude level was not considered feasible because of the normal variation of 
amplitudes in a typical recording. Consequently, the high, medium, and low amplitude 
levels subsequent decision made by consensus of Dr. Ross and PB. The decisions were 
relatively easy because the cavitations naturally divided into three general levels of 
amplitude as discussed below. 
Figure 5.6 A, B, C shows three different Left L3L4 Z joint cavitations with high, 
medium, and low amplitudes, respectively (Y axis shows the signal amplitude). 
 
 
 
 77 
 
 
 
 
 
 78 
 
Figure 5.6: A, B and C show cavitations occurring at the Left L3/L4 Z joint from three 
different subjects with high, medium, and low levels of cavitation amplitude, 
respectively. All the graphs are shown with the same scale (Y axis, which shows 
amplitude of signal in voltage). Figure A shows the accelerometers responding 
sequentially and very clearly. But in order to get the same visualization of 
response from accelerometers, the scale of Figure B would have to be expanded 
more than Figure A and Figure C expanded more than Figure B. The program 
easily allowed for such expansion 
 
 
 
 
 
 
 
 
 79 
5.4 Artifacts 
The identification and categorization of artifacts was important to the overall 
success of the project. However, clearly identifying artifacts allowed for efficient analysis 
of cavitations, which were the primary focus of this project. Some accelerometer signals 
that were not caused by cavitation were known as artifacts or noise. The wave forms for 
artifacts were irregular with larger spikes and did not fit the smooth ?non-continuous? 
wave form pattern found in cavitations. In addition, accelerometers responded 
dramatically out of sequence when recording artifacts, whereas the accelerometers 
responded in a very predictable pattern when recording cavitations. If only one 
accelerometer responded, (Figure 5.7) or if two-three responded or ?fired? but in non-
sequential order, (Figure 5.8) this was considered an artifact, generally caused by a 
clinician accidently touching the tape or the accelerometers. Other causes of artifacts 
were movement in wires, excessive movement in skin and muscles, movement of the 
subject?s body while applying force to cavitate the joints, and the subject excessively 
resisting movement. 
 
 80 
 
Figure 5.7: Accelerometer placed at SP L1 responding alone. This was one type of 
artifact identified during the study 
 
 
Figure 5.8: Accelerometers placed at SP L1 and SP L5 responding at the same time, with 
no other accelerometers responding, indicating that this was an artifact rather than 
a cavitation 
 
 81 
5.5 Types of Artifacts 
Based on notes taken during data acquisition and assessment for all the subjects 
several times, combined with the unique pattern and nature of recorded waveform 
signals, artifacts were identified and categorized into the following groups:  
1. Flexible Spine (or loose superficial fascia) subjects 
2. Hyper-muscular subjects (e.g., body builders) 
3. Resisting or Not Relaxed Subjects 
4. Unstable hand contact (UHC) (i.e., clinician?s hand contact) 
5. Miscellaneous 
The following subsections discuss each type of artifact. 
 
5.5.1 Flexible Spine (or Loose Superficial Fascia) Subjects 
The spine of some subjects was very flexible and the large amount of loose 
adipose tissues in other subjects made their superficial tissues (superficial fascia) very 
flexible. Similar artifacts were found in both types of subjects. The artifacts were spikes 
from the superior most and inferior most accelerometers. These artifacts can be explained 
with an analogy of the spine as a rope. Consider a flexible rope, and suppose someone 
tried to twist it by holding it at both ends. Then imagine that accelerometers were taped 
along the rope. The accelerometers taped nearest to the ends (superior most and inferior 
most) (SP L1 and S2 in our case) will tend to pop off (i.e., the rope will not hold) because 
of the greater amount of rotation at their locations. Likewise, the accelerometers taped to 
SP L1 and S1, and S2 tend to partially pop off with these subjects. This tendency was 
enhanced by the accelerometer being placed in the spinal gutter (the grooved area created 
 82 
by the muscles surrounding the SPs). This resulted in artifacts in which large spikes 
(Figure 5.9A) were seen throughout the time the thrust was applied. The accelerometer 
placed on S2 also tried to pop off, but the spikes were smaller (Figure 5.9B) than those 
from the SP L1 accelerometer, because there was usually more fat from the sacral area 
underneath, and supporting the accelerometer. Also, the majority of response was seen 
from these accelerometers compared with the other types of artifact (Figure 5.9C). To 
minimize this artifact, more tape was used to apply the accelerometers on the subject?s 
back in side-posture position during clinical study. 
 
 
 
 83 
 
 
 
Figure 5.9:  The flexible spine (or loose superficial fascia) subject artifact, A shows large 
spikes from the accelerometer placed at SP L1.  B shows spikes from the 
accelerometer placed at S2 (zoomed in during the time period of 0.984454 to 
0.998 sec). Note the pattern of waveform for both the accelerometers. C shows the 
complete time of collected data from all the accelerometers 
 84 
5.5.2 Hyper-Muscular Subjects (e.g., Body Builders) 
Subjects in this category were much more muscular than other subjects in the 
study. So the tape for all the accelerometers were placed on the larger curved area created 
by the paraspinal muscles. Also all the seven midline accelerometers were placed in the 
deeper groove area (spinal groove or gutter) created by the hypertrophic paraspinal 
muscles. Contrast this with subjects with a flexible spine where only accelerometers 
placed near to the top and bottom end (SP L1 and S2) were positioned in a deep groove 
area (as discussed earlier, Section 5.5.1). When a thrust or force was applied to the 
muscular subject?s body during SMT, almost all the accelerometers responded (Figure 
5.10A) because they were placed deeper in the groove area, which caused additional 
movement of the tape used for their application to the subject?s skin. So compared to the 
subjects with flexible spine (or loose superficial fascia) subjects in which only 
accelerometers placed at SP L1 and S2 responded, in muscular subjects most of the 
accelerometers responded with large spikes. Cavitations could be easily distinguished 
from this type of artifact. 
 
 
 
 
 85 
 
 
Figure 5.10: The hyper-muscular subject artifact. A shows complete time of collected 
data. Figure B shows most of the accelerometers responded 
 
 
 
 
 
 
 
 86 
5.5.3 Resisting or Not Relaxed Subjects  
Subjects in this category were not completely relaxed (subjects were ?jittery?), 
causing the muscles under the treatment area, or the area where the accelerometers were 
placed, to be excessively stiff or contract irregularly. For example, often the clinician will 
ask the subject to relax. If a subject is relaxed, the clinician should be able to lift and 
release the subject?s arm and have it fall down freely. By contrast, if the subject tries to 
control the arm and prevent it from falling down, the muscles in that area will be stiff. 
Similarly, if the subject tries to resist the clinician at the time when data was collected 
during side-posture positioning and/or side-posture SMT, a force (applied by clinician) 
and a reaction (resisting force by the subject) will occur. This makes almost all the 
accelerometers placed on the subject?s skin respond because of the reacting force and 
stiffness or resisting contraction occurring in the subject?s muscles (Figure 5.11). This 
type of artifact was not compatible with assessing cavitations, and subjects with this type 
of artifact were not included in the study. To overcome resisting, or not-relaxed subject 
artifacts for the remaining part of the study, subjects were asked to relax and not to resist 
the study clinician. 
 
 
 
 
 
 
 
 87 
 
 
Figure 5.11: Resisting or not relaxed subject?s artifact. Figure A shows complete time of 
collected data. Figure B shows most of the accelerometers responded and went off 
from the baseline because of stiffness and resistance of the not-relaxed subject 
(zoomed in during the time period of 1.223231 to 1.490796 sec of Figure B) 
 
 
 
 
 
 88 
5.5.4 Unstable Hand Contact (UHC) (i.e., Clinician?s Hand Contact)  
Subjects in this category were wearing shorts made of a slippery type of fabric 
(basketball shorts made of slippery nylon or a similar fabric, for example). This could 
prevent the clinician from getting a firm contact on the subject?s body, making it slip 
during application of the thrust during SMT. This could cause jiggling of the muscles 
and/or skin, causing most of the accelerometers to respond (Figure 5.12). This type of 
artifact was sometimes compatible with assessing cavitations, but other times caused too 
much ?noise? to adequately assess the data for cavitations. This artifact was identified 
relatively early in the project. As a result, cotton surgical scrub pants were purchased in 
several sizes and all subjects were required to wear the scrubs during the experiments.  
 
 
 
 
 
 
 
 
 
 
 89 
 
 
Figure 5.12: Unstable Hand Contact (UHC) (i.e., clinician?s hand contact) artifact. Figure 
A shows the complete time of the collected data. Figure B shows that most of the 
accelerometers responded simultaneously because of the jiggling of tissues due to 
the clinician?s unstable hand contact caused by ?slippery shorts? (zoomed in 
during the time period of 1.828613 to 1.982901 sec of Figure A) 
5.5.5 Miscellaneous  
Some artifacts were generated because of reasons other than flexible spine (or 
loose superficial fascia) subjects, hyper-muscular subjects (e.g., body builders), resisting 
or not relaxed subjects, and UHC (i.e., clinician?s hand contact). These subjects were 
 90 
included in the miscellaneous category. The number of artifacts in this category was less 
than the others. In general, these artifacts were caused by an irregular movement of the 
subject?s body while applying SMT. Frequently; these artifacts were localized and did 
not interfere with the assessment of cavitations. Figure 5.13 shows an example of this 
type of artifact. 
 
 
 
Figure 5.13: Miscellaneous artifacts. Figure A shows the complete time of the collected 
data. Figure B shows the accelerometer placed at SP L5 was the only one that 
responded during this brief time period 
 91 
5.6 Data Sheets 
Data sheets were made specifically for this project and were completed for all the 
subjects that finished the project (n=40). 
Data sheets (with no personal identifiers) were prepared, burned on compact disc 
(CD) and printed to keep electronic and hard copy records of all the subjects included in 
the study. This also helped the ?clinical studies office? to efficiently enter the analyzed 
data into their system. Two separate binders of printed data sheets were made, one for 
keeping record of all the subjects with the Principal Investigator (Dr. Gregory Cramer) 
and the other for Clinical Studies Office. Dr. Cramer also kept the CDs in a locked 
fireproof safe in his office. 
The following information was noted for inclusion on the data sheet for each 
subject (identified by ID number during the accelerometry appointment): 
 Subject ID number 
 Date 
 Time  
 Time of first thrust for subjects randomized into Group 1  
OR 
            Time of side-posture positioning for subjects randomized into Group 2. 
 Questions asked by research assistant to the subject and clinician about 
report of cavitation heard and/or felt during side-posture positioning, first thrust, and 
second thrust were recorded in the subject?s research file and entered into the master data 
base.  
 
 92 
The following information was noted on the data sheet for each subject during 
data analysis: 
Assessment of accelerometer data: Artifacts 
 Sequential number (by time) for each artifact. 
 Time of occurrence of each artifact. 
 Comments/ notes on each artifact. 
 
Assessment of accelerometer data: Cavitations 
 Sequential number (by time) for each cavitation. 
 Time of occurrence of each cavitation. 
 Comments/notes on each cavitation. 
 
Amplitude of cavitations 
 Sequential number for each cavitation. 
 Order of response of accelerometers placed on SP. 
 Order of response of accelerometers placed Left and Right of L4/L5 
interspinous space. 
 Level (grade: high, medium, low) of amplitude for each cavitation. 
 
Comments/ notes for each subject. 
 
Printouts were made of the oscilloscope recordings for each artifact, cavitation, 
and any other finding of interest were attached in sequential order as entered into the data 
 93 
sheet and were a part of the electronic and hard copy record for each subject. These 
supporting documents for each subject ID number were printed on a white background 
using the ?Export Simplified Image? option in LabVIEW. This procedure produced a 
clear image. 
Examples of data sheets and supporting documents (printouts of artifacts and 
cavitations) for representative subject from Group 1 and Group 2 are included in the 
following pages. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 94 
DATA SHEET (from subject in Group 1 with many artifacts) 
 
 
 SUBJECT ID: 0035                     DATE: 03/11/2010 
 
 TIME OF FIRST                                                                     TIME OF MRI: 8:45am 
THRUST (SEC) 
Or 
POSITIONING TIME:   1.19 
      
 
Table 5.1: ASSESSMENT OF GROUP 1 ACCELEROMETER DATA-- 
ARTIFACTS 
ARTIFACTS TIME 
(SEC) 
COMMENTS/ NOTES 
 
1) 
 
 
2) 
 
 
 
 
3) 
 
 
4) 
 
 
 
5) 
 
 
 
6) 
 
 
 
7) 
 
 
8) 
 
0.7936 
 
 
0.9 
 
 
 
 
0.9436 
 
 
1.0148 
 
 
 
1.1395 
 
 
 
1.408356 
 
 
 
1.8262 
 
 
2.0478 
 
Only L2 responded, which shows no sign of 
cavitation (Miscellaneous). 
 
L5, L3 and then L2 responded but L4 hasn?t 
responded after L5, hence it?s not a cavitation 
(Hyper-muscular subject). 
 
L3, L2 and then L5 responded but L4 haven?t 
responded before L5, hence it?s not a cavitation 
(Hyper-muscular subject). 
 
L3, L2 and then L5 responded but L4 hasn?t 
responded before L5, hence it?s not a cavitation 
(Hyper-muscular subject). 
 
L3, L2 and then L5 responded but L4 hasn?t 
responded before L5, hence it?s not a cavitation 
(Hyper-muscular subject). 
 
Only L1 and Left L4L5 responded, which shows no 
sign of cavitation (Miscellaneous). 
 
L5 and then L3 responded but L4 haven?t 
responded after L5, hence it?s not a cavitation 
(Miscellaneous). 
 
 Only L1 and Left L4L5 responded, which shows no 
sign of cavitation (Miscellaneous). 
 
 
 95 
Table 5.2: ASSESSMENT OF GROUP 1 ACCELEROMETER DATA?
CAVITATIONS 
CAVITATIONS TIME NOTES 
 
1) 
 
 
1.2923 
 
Left L4L3 
 
2) 
 
 
1.35686 
 
Left L4L3 
 
 
Table 5.3: AMPLITUDE OF CAVITATIONS GROUP 1 
CAVITATION NO ORDER OF SP 
FIRING 
ORDER (LEFT 
OR RIGHT) 
LEVEL OF 
CONFIDENCE 
 
1) 
 
L4 
L3 
L5 
L2 
S1 
S2 
L1 
 
 
Left L4L5 
Right L4L5 
 
High 
 
2) 
 
L3 
L4 
L2 
L5 
L1 
 
 
Left L4L5 
Right L4L5 
 
High 
 
 
 
COMMENTS/NOTES:  
 
This subject was in group one (adjustment) and interestingly two cavitations were 
recorded at the same joint. 
 
 
 
 96 
 
Figure: 5.14: Complete time cycle for subject ID 0035 
 
Figure 5.15: Artifact 1 for subject ID 0035 
 97 
 
Figure 5.16: Artifact 2 for subject ID 0035 
 
Figure 5.17: Artifact 3 for subject ID 0035 
 
 98 
 
Figure 5.18: Artifact 4 for subject ID 0035 
 
Figure 5.19: Artifact 5 for subject ID 0035 
 
 
 99 
 
Figure 5.20: Artifact 6 for subject ID 0035 
 
 
Figure 5.21: Artifact 7 for subject ID 0035 
 100 
 
Figure 5.22: Artifact 8 for subject ID 0035 
 
 
 
 
 
 
 
 
 
 
 101 
 
Figure 5.23: Cavitation 1 full view for subject ID 0035 
 
Figure 5.24: Expanded (zoomed), view of figure 5.23 (1.292286 to 1.29244 sec) 
showing sequential response of accelerometers 
 102 
 
Figure 5.25: Cavitation 2 full view for subject ID 0035 
 
Figure 5.26: Expanded (zoomed), view of figure 5.25 (1.356843 to 1.356944 sec) 
showing sequential response of accelerometers 
 
 103 
DATA SHEET (for subject in Group 2) 
 
SUBJECT ID: 0046                     DATE: 03/12/2010 
 
TIME OF FIRST THRUST                                                       TIME OF MRI: 11:00am 
or    
 POSITIONING TIME: 2.84                                                     
 (SEC)                     
      
 
Table 5.4: ASSESSMENT OF GROUP 2 ACCELEROMETER DATA--
ARTIFACTS 
ARTIFACTS TIME 
(SEC) 
COMMENTS/ NOTES 
 
1) 
 
 
2) 
 
 
3) 
 
 
4) 
 
5) 
 
0.50585  
 
 
0.546 
 
 
0.9159 
 
 
1.3382 
 
1.5966 
 
L5 and then L1 responded, which shows no sign of 
cavitation (Miscellaneous). 
 
S1 and then L1 responded, which shows no sign of 
cavitation (Miscellaneous). 
 
Only S2 responded, which shows no sign of 
cavitation (Miscellaneous). 
 
Only L4 responded, which shows no sign of 
cavitation (Miscellaneous). 
 
Only S2 responded, which shows no sign of 
cavitation (Miscellaneous). 
 
 
 
 
Table 5.5: ASSESSMENT OF GROUP 2 ACCELEROMETER DATA?
CAVITATIONS 
CAVITATIONS TIME NOTES 
None   
 
 
 
 
Table 5.6: AMPLITUDE OF CAVITATIONS GROUP 2 
CAVITATION NO ORDER OF SP 
FIRING 
ORDER (LEFT 
OR RIGHT) 
LEVEL OF 
CONFIDENCE 
None    
 
 104 
COMMENTS/NOTES: 
 
 This subject was in Group two (side-posture positioning) and no cavitation was 
recorded. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 105 
 
Figure 5.27: Complete time scale for subject ID 0046 
 
Figure 5.28: Artifact 1 for subject ID 0046 
 106 
 
Figure 5.29: Artifact 2 for subject ID 0046 
 
Figure 5.30: Artifact 3 for subject ID 0046 
 
 107 
 
Figure 5.31: Artifact 4 for subject ID 0046 
 
Figure 5.32: Artifact 5 for subject ID 0046 
 
 
 108 
5.7 Reliability Studies 
 
Two types of reliability studies were conducted in this research project.  
 
5.7.1 Inter-Rater Agreement between Computer (Accelerometers), Clinician, and 
Subject 
In this reliability study, agreement between the computer (accelerometers), 
clinician, and subject report in determining any cavitation during side-posture positioning 
or SMT was determined. As discussed earlier in the ?clinical study? section, a trained 
research assistant noted the responses of a clinician and subject (both blinded to each 
other) and whether or not they heard or felt cavitation during SMT or side posture-
positioning. Inter-rater agreements (Kappa) were calculated using Medcalc. Agreement is 
quantified by the Kappa (K) statistic: 
 K is 1 when there is perfect agreement between the observers using the 
classification system (when they agreed on the presence of a cavitation, in our case). 
 K is 0 when there is no agreement better than chance. 
 K is negative when agreement is worse than chance. 
 
 
 
 
 109 
Table 5.7: Presence of cavitations as identified by the accelerometers, clinician, and the 
subject 
*SUBJECT ID ACCELEROMETERS CLINICIAN SUBJECT 
6 1 1 1 
9 0 0 0 
10 0 0 0 
11 1 1 1 
12 1 1 1 
13 1 1 1 
14 1 1 1 
16 0 1 1 
18 1 1 1 
21 1 1 1 
24 1 1 1 
30 1 1 1 
32 1 1 1 
34 1 1 1 
35 1 1 1 
37 1 1 1 
40 1 1 1 
41 0 0 0 
45 1 1 1 
 110 
46 0 0 0 
47 1 1 1 
58 1 1 1 
59 1 1 1 
61 1 1 1 
65 1 1 1 
69 1 1 1 
71 1 1 1 
72 1 1 1 
74 1 1 1 
75 1 1 1 
80 0 0 0 
82 1 1 1 
84 0 1 0 
85 1 1 1 
86 1 1 1 
87 0 0 0 
90 1 1 1 
92 1 1 1 
93 1 1 1 
97 1 1 1 
* Subjects are listed in ascending order by their Subject ID number (each was assigned a 
random Subject ID number 1-100). 
 111 
5.7.2 Reliability Study for Agreement Assessing Accelerometry Data 
This study was conducted between two observers, (PB and Dr. Kim Ross DC, 
PhD), each blinded to the other?s results, to identify the vertebral level of cavitations  
For this study, initial cavitations for ten subjects were chosen using a ?Random 
Number Generator? program. PB assessed the cavitations, recorded his results and gave 
them to Dr. Cramer. A Windows Live Meeting was then scheduled with Dr. Ross, and 
each cavitation was shared with him one at a time. Control was given to Dr. Ross, so that 
he could scroll through the accelerometer recording time and magnify the scales in any 
manner he chose. Dr. Ross recorded his results and sent them to Dr. Cramer via email. 
PB and Dr. Ross were blinded to each others results. 
Inter-rater agreement (Kappa) was calculated using Medcalc software. The Z 
joints identified as having cavitated were individually and sequentially entered into the 
database for each observer (PB and Dr. Ross). Since there were 12 joints included in the 
study in total, (Left L1L2, Right L1L2, Left L2L3, Right L2L3, Left L3L4, Right L3L4, 
Left L4L5, Right L4L5, Left L5S1, Right L5S1, Left SIJ and Right SIJ) a number was 
assigned to each of them sequentially from 1 to 12. For example, 1 was assigned to the 
Left L1L2 Z joint, 2 was assigned to the Right L1L2 Z joint, etc. So, if PB and Dr. Ross 
concluded cavitation occurred at the same Z joint for a particular joint, the same number 
would be entered for each of them in the data entry table of Medcalc. The weighted 
Kappa statistic was used to analyze the data. 
 
 
 
 112 
Table 5.8: Assignment of numeric values to Z joints for reliability study data entry 
 
DATA ENTRY 
 
JOINTS 
1 LEFT L1L2 
2 RIGHT L1L2 
3 LEFT L2L3 
4 RIGHT L2L3 
5 LEFT L3L4 
6 RIGHT L3L4 
7 LEFT L4L5 
8 RIGHT L4L5 
9 LEFT L5S1 
10 RIGHT L5S1 
11 LEFT SIJ 
12 RIGHT SIJ 
 
 
 
 
 
 
 
 
 113 
5.8 Summary of Cavitation Assessment 
Data analysis in this research project included the identification of the vertebral 
level where cavitation occurred (i.e., identification of cavitation at specific Z joints from 
L1-L5 SPs and SIJs). Inferential methods were used to compare cavitations in the 
following manner: Group 1 (SMT group) vs. Group 2 (side-posture positioning group), 
targeted area (L3/L4, L4/L5, L5/S1) vs. non-targeted area (L1/L2, L2/L3, SIJ), left side 
vs. right side.  
 
5.9 Statistical Analysis 
Statistical analysis was conducted between Group 1 and Group 2 using the chi-
square method because of the independent location of cavitations in two groups. 
McNemar?s test was done for statistical analysis for cavitations in left side vs. right side 
and targeted vs. non-targeted areas. Mcnemar?s test was done because two related 
measurements on the same sample group were considered (Snedecor and Cochran, 1967). 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 114 
6. RESULTS 
 
 
Forty-nine subjects were randomized into Group 1 or Group 2. Nine of these were 
excluded from the study immediately after side-posture positioning or SMT because of 
many artifacts incorporated during accelerometry data collection (Three subjects were 
excluded because of UHC (i.e., clinician?s hand contact) artifact.  Five subjects were 
excluded because of resisting or not relaxed subjects artifact. One subject was excluded 
from the study because of incomplete data collection). Consequently, a total of 40 
subjects (20 males and 20 females) completed the study. As planned, 30 subjects were 
randomized into Group 1, and 10 subjects were randomized into Group 2. Group 1 was 
the SMT group (a total of 30 subjects were randomized into this group) and Group 2 was 
the Control Group (a total of 10 subjects were randomized into this group). These 
subjects were held in side-posture position, but didn?t receive side-posture SMT. Data 
was collected from accelerometers with subjects in Group 2 to test the hypothesis, that 
cavitation can occur when subjects are held in side-posture position. 
 
 
 
 
 
 
 115 
6.1 Reliability Studies 
As discussed earlier, two types of reliability studies were conducted. One study 
attempted to find agreement between the computer (accelerometers), clinician report and 
subject report in determining any cavitation during side-posture positioning or SMT. The 
other attempted to identify the vertebral level of cavitations by two observers (PB and Dr. 
Kim Ross DC, PhD) blinded to each others results. Results of reliability studies were 
calculated using Medcalc. 
6.1.1 Agreement between Accelerometers and Subjects 
There was agreement between the accelerometer recordings and subjects for 39 of 
the 40 instances, and they both disagreed for one instance as seen in Tables 5.8, 6.1 and 
in Figure 6.1. When no cavitation was identified from accelerometer recordings (n=8) 
then also, subject said ?no? (0 on Table 5.8) 7 times (agreement) and ?yes? 1 time 
(disagreement) (Tables 5.8, 6.1 and in Figure 6.1). When the accelerometers identified a 
cavitation (n=32) the subjects said yes (1 on Table 5.8) all 32 times (complete agreement) 
(Tables 5.8, 6.1 and in Figure 6.1).  
Accelerometer recordings of cavitations were in ?almost perfect agreement? 
(Cohen, 1960; Landis and Koch, 1977; Gottman, 1995; Fleiss et al., 2003) with the 
subjects (K=0.918, 95% CI= 0.759 to 1.077) (Table 6.1).  
 
 
 
 
 116 
Table 6.1: Inter-rater agreement (Kappa) between accelerometers and subjects 
Inter-rater agreement (kappa) 
Observer A Accelerometers 
Observer B Subjects 
 Observer A  
Observer B 0 1  
0 7 0 7 (17.5%) 
1 1 32 33 (82.5%) 
 8 
(20.0%) 
32 
(80.0%) 
40 
Kappa 0.918 
Standard error 0.0809 
95% CI 0.759 to 1.077 
 
 
Figure 6.1: Frequency chart for Inter-rater agreement (Kappa) between 
accelerometers and subjects 
 
 117 
6.1.2 Agreement between Clinician and Subjects 
Of 40 subjects, there was agreement between clinician and subjects for 39 
instances and they disagreed in one instance as seen in Tables 5.8, 6.2 and Figure 6.2. 
When the clinician reported ?no cavitation? (0 on Table 5.8, n=6), the subjects reported 
?no? all 6 times (complete agreement) (Table 6.2). When the clinician reported ?yes, 
there was a cavitation? (1 on Table 5.8, n=34) the subjects reported ?no? 1 time 
(agreement in all but 1 instance) (Tables 5.8, 6.2 and Figure 6.2). 
Clinician report of cavitations were in ?almost perfect agreement? (Cohen, 1960; 
Landis and Koch, 1977; Gottman, 1995; Fleiss et al., 2003) with the subjects (K=0.918, 
95% CI= 0.731 to 1.086) (Table 6.2).  
 
Table 6.2: Inter-rater agreement (Kappa) between clinician and subjects 
Inter-rater agreement (kappa) 
Observer A  Clinician 
Observer B  Subjects 
  Observer A    
Observer B  0 1   
0 6 1 7 (17.5%) 
1 0 33 33 (82.5%) 
  6 
(15.0%) 
34 
(85.0%) 
40 
Kappa  0.908 
Standard error  0.0906 
95% CI  0.731 to 1.086 
 118 
 
Figure 6.2: Frequency chart for Inter-rater agreement (Kappa) between 
clinician and subjects 
 
6.1.3 Agreement between Accelerometers and Clinician 
Of 40 subjects, there was agreement between accelerometers and the clinician for 
38 instances and disagreement for two instances as seen in Tables 5.8, 6.3 and Figure 6.3. 
When the accelerometers did not identify a cavitation (0 in Table 5.8, n=8) the clinician 
reported no 6 times and yes 2 (1 in Table 5.8) times (Table 6.3). When accelerometers 
identified a cavitation, (1 in Table 5.8, n=32) the clinician reported yes all 32 times 
(complete agreement).  
Accelerometer recordings of cavitations were in ?almost perfect agreement? 
(Cohen, 1960; Landis and Koch, 1977; Gottman, 1995; Fleiss et al., 2003) with the 
clinician (K=0.828, 95% CI= 0.595 to 1.060) (Table 6.3).  
 119 
Table 6.3: Inter-rater agreement (Kappa) between accelerometers and clinician 
Inter-rater agreement (kappa) 
Observer A  Accelerometers 
Observer B  Clinician 
  Observer A    
Observer B  0 1   
0 6 0 6 (15.0%) 
1 2 32 34 (85.0%) 
  8 
(20.0%) 
32 
(80.0%) 
40 
Kappa  0.828 
Standard error  0.119 
95% CI  0.595 to 1.060 
 
 
Figure 6.3: Frequency chart for Inter-rater agreement (Kappa) between 
accelerometers and clinician 
 
 
 120 
6.1.4 Agreement Assessing Accelerometry Data  
This reliability study was conducted between two observers (PB and Dr. Kim 
Ross DC, PhD), blinded to each others results, to identify the vertebral level of 
cavitations. Initial cavitations for ten subjects were chosen using a ?Random Number 
Generator? program and the level of cavitations was identified by each observer using a 
previously discussed method. 
Of ten cavitations, there was agreement between PB and Dr. Ross in nine 
instances identifying the level of cavitation (Cavitation number 1-9 Table 6.4) (complete 
agreement) and disagreement in one instance (cavitation number 10 Table 6.4). PB and 
Dr. Ross identified level of cavitation at Left L1/L2 Z joint in 2 instances (1 in Table 6.5 
and Figure 6.4), Left L2/L3 Z joint in one instance (3 in Table 6.5 and Figure 6.4), Left 
L3/L4 Z joint in 5 instances (5 in Table 6.5 and Figure 6.4), and Left SIJ in one instance 
(11 in Table 6.5 and Figure 6.4). Only in one instance PB identified level of cavitation at 
the Left L3/L4 Z joint and Dr. Ross identified the same cavitation as the Left L2/L3 Z 
joint (Tables 6.4, 6.5 and Figure 6.4). PB and Dr. Ross agreed on all but one Z joint and 
were only off by 1 level in that instance, i.e., they agreed in 9 of 10 instances (Table 6.4). 
The reliability study in assessing accelerometry data between PB and Dr. Ross 
was in ?almost perfect agreement? (Cohen, 1960; Landis and Koch, 1977; Gottman, 
1995; Fleiss et al., 2003) (K=0.841, 95% CI = 0.546 to 1.136). 
 
 
 
 
 121 
Table 6.4: Interpretation of cavitations by PB and Dr. Ross 
 
CAVITATION NUMBER 
 
PB RESULTS 
 
DR. ROSS RESULTS 
1 LEFT L3L4 LEFT L3L4 
2 LEFT L3L4 LEFT L3L4 
3 LEFT L1L2 LEFT L1L2 
4 LEFT S1S2 LEFT S1S2 
5 LEFT L3L4 LEFT L3L4 
6 LEFT L2L3 LEFT L2L3 
7 LEFT L1L2 LEFT L1L2 
8 LEFT L3L4 LEFT L3L4 
9 LEFT L3L4 LEFT L3L4 
10  *LEFT L3L4 *LEFT L2L3 
 
* Single disagreement between observers.  
 
 
 
 
 
 
 
 122 
Table 6.5: Inter-rater agreement (Kappa) for analysis of accelerometry recordings 
between two observers 
Inter-rater agreement (kappa) 
Observer A PB 
Observer B Dr. Ross 
 
 Observer A  
Observer B 1 3 5 11  
1 2 0 0 0 2 (20.0%) 
3 0 1 1 0 2 (20.0%) 
5 0 0 5 0 5 (50.0%) 
11 0 0 0 1 1 (10.0%) 
 2 
(20.0%) 
1 
(10.0%) 
6 
(60.0%) 
1 
(10.0%) 
10 
Kappa 0.841 
Standard error 0.151 
95% CI 0.546 to 1.136 
 Figure 6.4: Frequency chart for reliability study of analysis of accelerometry 
recordings 
 123 
6.1.5 Summary of Reliability Studies 
 
Kappa findings of both reliability studies are summarized in this section. 
 
6.1.5.1 Agreement between Accelerometers, Clinician, and Subjects 
 
 For 38 (No=6; Yes=32) of 40 subjects, all three sources (accelerometers, 
clinician and subjects) agreed. 
 For the remaining 2 subjects, accelerometers detected nothing. 
 The doctor reported cavitation in both instances. 
 The subject reported cavitation in one of these instances. 
 Inter-rater agreement (Kappa) between accelerometers and subjects was 
0.918 (almost perfect) (Cohen, 1960; Landis and Koch, 1977; Gottman, 1995; Fleiss et 
al., 2003). 
 Inter-rater agreement (Kappa) between clinician and subjects was 0.908 
(almost perfect) (Cohen, 1960; Landis and Koch, 1977; Gottman, 1995; Fleiss et al., 
2003). 
 Inter-rater agreement (Kappa) between accelerometers and clinician was 
0.828 (almost perfect) (Cohen, 1960; Landis and Koch, 1977; Gottman, 1995; Fleiss et 
al., 2003). 
 
6.1.5.2 Agreement between Two Observers in Assessing Accelerometry Data 
Both observers, PB and Dr. Ross agreed on all but one Z joint and were only off 
by 1 level in that instance, i.e., they agreed in 9 of 10 instances. Inter-rater agreement 
(Kappa) between two observers was 0.841 (almost perfect) (Cohen, 1960; Landis and 
Koch, 1977; Gottman, 1995; Fleiss et al., 2003). 
 124 
6.2 Number and Percentage of Cavitations (CAV) Detected By the Accelerometers 
at Each Joint 
The main focus of this research project was to find cavitations at SP L1-SIJ and, 
for this reason; each cavitation recorded by accelerometers is reported in this project. A 
total of 56 cavitations were recorded from 40 subjects. 53 cavitations were recorded from 
subjects randomized into Group 1 and 3 from subjects randomized into Group 2 (Table 
6.6). The maximum number of cavitations was recorded at Left L3/L4 Z joint (Table 6.6, 
Figure 6.5). No cavitations were recorded at the Right L1/L2, L2/L3, and L3/L4 Z joints 
(Table 6.6). 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 125 
Table 6.6: Number and percentage of cavitations recorded at each joint 
 
JOINT 
 
GROUP 1 
 
GROUP 2 
 
PERCENTAGE 
CAV/ JOINT 
LEFT L1L2 7 0 12.50 
RIGHT L1L2 0 0 0 
LEFT L2L3 6 0 10.71 
RIGHT L2L3 0 0 0 
LEFT L3L4 24 2 46.43 
RIGHT L3L4 0 0 0 
LEFT L4L5 8 0 14.28 
RIGHT L4L5 1 0 1.78 
LEFT L5S1 4 1 8.93 
RIGHT L5S1 1 0 1.78 
LEFT SIJ 1 0 1.78 
RIGHT SIJ 1 0 1.78 
TOTAL 53 3 100 
 
 126 
 
 
Figure 6.5: Percentage of cavitations recorded at each joint. 
Notice the maximum number of cavitations was 
recorded at Left L3/L4 Z joint and that Right L1/L2, 
L2/L3, and L3/L4 did not cavitate at all 
 
 
 
 
 
 
 127 
6.3 Cavitations in Group 1 and Group 2 
Comparisons of cavitations between Group 1 and Group 2 can be summarized as 
follows:  
 53 cavitations (53/56 = 94.64%) were recorded from subjects randomized 
into Group 1 (n=30 subjects) (Table 6.6, and Figures 6.6, 6.9). 38 of 53 cavitations were 
recorded at the targeted area (L3/L4, L4/L5, and L5/S1) and 15 of 53 cavitations were 
recorded at the non-targeted area (L1/L2, L2/L3, and SIJ) from subjects randomized into 
Group 1 (Figure 6.9). 
 3 (3/56 = 5.36%) cavitations were recorded from subjects randomized into 
Group 2 (n=10 subjects) (Table 6.6, and Figures 6.6, 6.10). None of the cavitations were 
recorded on right side for the subjects randomized into Group 2 (Figure 6.10). 
 29 out of 30 subjects cavitated in Group 1 (29/30 = 97%), which was 
significantly more vs. 3 out of 10 subjects that cavitated in Group 2 (3/10 = 30%) (2 
sided test, p< 0.0001, 95%CI=34%-86%) (Refer chi-square test). 
 Average cavitations / subject in Group 1 was 1.77 (53 / 30). 
 Average cavitations / subject in Group 2 was 0.3 (3 / 10). 
 
 
 
 
 
 
 
 128 
 
 
Figure 6.6: Percentage of cavitation in Group 1 and Group 2 
 
 
6.4 Cavitations at Targeted and Non-targeted Area 
Forty-one cavitations (41/56 = 73.21%) (Figures 6.9, 6.10) were recorded at the 
targeted Z joints (Left L3/L4, Right L3/L4, Left L4/L5, Right L4/L5, Left L5/S1, and 
Right L5/S1). Of these 41 cavitations, 39 were recorded on the left side, and only 2 were 
recorded on the right side (Figures 6.9, 6.10). Fifteen cavitations (15/56 = 26.80%) 
(Figures 6.9, 6.10) were recorded at non-targeted Z joints (Left L1/L2, Right L1/L2, Left 
L2/L3, Right L2/L3, Left SIJ and Right SIJ). Of the15 cavitations, 14 were recorded from 
 129 
left side non-targeted Z joints, and only 1 cavitation was recorded on the right side of 
non- targeted Z joints (Figures 6.9, 6.10). 
Thirty-eight of 53 cavitations were recorded at the targeted area, and 15 of 53 
cavitations were recorded at the non-targeted area for the subjects randomized into Group 
1 (Figure 6.9). All 3 cavitations were recorded at the target area and, interestingly, on the 
left side for subjects randomized into Group 2 (Figure 6.10). 
Cavitation occurred significantly more frequently at the targeted Z joints vs. non-
targeted Z joints (p<0.001). 
 
 
Figure 6.7: Percentage of cavitations at targeted and non targeted Z 
joints 
 
 
 
 130 
6.5 Cavitations at Left Side and Right Side 
Fifty-three cavitations (53/56 = 94.63%) (Figure 6.8) were recorded from joints 
located on the left side, 39 of which were recorded at targeted Z joints and 14 at non-
targeted Z joints (Figures 6.9, 6.10). Three cavitations (3/56 = 5.43%) (Figure 6.8) were 
recorded from joints located on the right side, 2 of which were recorded at targeted Z 
joints and 1 at non-targeted Z joints (Figures 6.9, 6.10). Cavitation occurred significantly 
more frequently at left-side vs. right-side (p<0.001). 
Fifty of 53 cavitations (Figure 6.9) were recorded at the left side and 3 of 53 
cavitations (Figure 6.9) were recorded at the right side from subjects randomized into 
Group 1. All 3 cavitations were recorded at the left side from subjects randomized into 
Group 2 (Figure 6.10).  
 
Figure 6.8: Percentage of cavitations recorded on left side and right 
side 
 131 
 
Figure 6.9: Posterior view of the spine with the number of cavitations 
recorded at each joint on the left and right side for Group 1 (SMT 
group). Red line shows the targeted Z joints (L3/L4, L4/L5, and 
L5/S1). Significantly more cavitations were recorded on left side 
as compared to right side (p<0.001). Interestingly, more 
cavitations were recorded at targeted Z joints than non-targeted Z 
joints (p<0.001) 
 132 
 
Figure 6.10: Posterior view of the spine with the number of cavitations 
recorded at each joint on the left and right side for Group 2 
(non-SMT group). Red line shows the targeted Z joints (L3/L4, 
L4/L5 and L5/S1). Interestingly, all the cavitations in this 
group were recorded on the left side and in the targeted area 
 
 
 
 
 
 
 133 
6.6 Multiple Cavitations 
The accelerometry methods used in this study showed the capability of recording 
multiple cavitations from different joints during a single SMT. Also interestingly; 
multiple cavitations from the same joint were recorded.  
Fifty-six cavitations were recorded from a total of 40 subjects randomized into 
Group 1 or Group 2 (Table 6.6). Out of these 56, single cavitations were recorded in 38 
joints (Table 6.7) (35 in Group 1 and 3 in Group 2), double cavitations in 7 joints and 4 
cavitations in 1 joint (Table 6.7). Multiple cavitations from the same joint were detected 
only on the left side in Group 1; none were detected on the right side or in Group 2 
(Table 6.7). 
The range of cavitations for subjects included in Group 1 was 0-6 and was 0-1 for 
subjects in Group 2.  
The maximum of six cavitations were recorded from a subject randomized in 
Group 1, interestingly with 4 cavitations at the same joint (Left L3/L4) and 2 cavitations 
at another joint (Left L2/L3). This unique subject?s data was collected in the late evening, 
the only evening recording in the study. 
 
 
 
 
 
 
 
 134 
Table 6.7: Distribution of cavitations detected by accelerometers 
Joints Number of Cavitations Reported for a Single Joint 
 0 1 2 4 
LEFT L1L2 49 5 1 0 
RIGHT L1L2 0 0 0 0 
LEFT L2L3 50 4 1 0 
RIGHT L2L3 0 0 0 0 
LEFT L3L4 30 14 4 1 
RIGHT L3L4 0 0 0 0 
LEFT L4L5 48 8 0 0 
RIGHT L4L5 55 1 0 0 
LEFT L5S1 51 3 1 0 
RIGHT L5S1 55 1 0 0 
LEFT SIJ 55 1 0 0 
RIGHT SIJ 55 1 0 0 
 
 
 
 
 
 
 
 
 
 135 
6.7 Amplitude of Cavitations 
 
Table 6.8 shows the amplitude grade for each cavitation. This data will be used in 
an ongoing study (not reported here) to see if there is any relation between amplitude of 
cavitation signals (high, medium, and low) and the amount of gapping in the respective Z 
joints. 
 
Table 6.8: Amplitude of cavitation signals 
SUBJECT ID NUMBER 
 
LEVEL OF 
CAVITATION 
AMPLITUDE OF 
CAVITATION 
6 LEFT L3L4 MEDIUM 
9 NONE NONE 
10 NONE NONE 
11 RIGHT L5S1 
 
RIGHT L4L5 
HIGH 
 
HIGH 
12 LEFT L3L4 
 
LEFT L4L5 
MODERATE 
 
HIGH 
13 LEFT L3L4 
 
RIGHT SIJ 
HIGH 
 
HIGH 
14 LEFT L4L5 
 
LEFT L2L3 
MEDIUM 
 
HIGH 
16 NONE NONE 
18 LEFT L4L5 
 
LEFT L2L3 
HIGH 
 
HIGH 
21 LEFT L3L4 HIGH 
24 LEFT L3L4 
 
LEFT L4L5 
LOW 
 
MEDIUM 
30 LEFT L3L4 HIGH 
32 LEFT L3L4 
 
LEFT L3L4 
HIGH 
 
HIGH 
34 LEFT L5S1 
 
LEFT L5S1 
HIGH 
 
MODERATE 
 136 
 
LEFT L4L5 
 
HIGH 
35 LEFT L3L4 
 
LEFT L3L4 
HIGH 
 
HIGH 
37 LEFT L3L4 
 
LEFT L4L5 
MEDIUM 
 
LOW 
40 LEFT L3L4 
 
LEFT L3L4 
HIGH 
 
HIGH 
41 NONE NONE 
45 LEFT L5S1 HIGH 
46 NONE NONE 
47 LEFT L3L4 
 
LEFT L3L4 
HIGH 
 
HIGH 
58 LEFT L4L5 
 
LEFT L5S1 
HIGH 
 
HIGH 
59 LEFT L1L2 MEDIUM 
61 LEFT L4L5 LOW 
65 LEFT L3L4 HIGH 
69 LEFT L3L4 HIGH 
71 LEFT L1L2 
 
LEFT SIJ 
 
LEFT L1L2 
HIGH 
 
HIGH 
 
HIGH 
72 LEFT L3L4 LOW 
74 LEFT L5S1 LOW 
75 LEFT L2L3 
 
LEFT L1L2 
HIGH 
 
HIGH 
80 NONE NONE 
82 LEFT L1L2 HIGH 
84 NONE NONE 
85 LEFT L3L4 HIGH 
86 LEFT L3L4 HIGH 
87 NONE NONE 
90 LEFT L1L2 HIGH 
92 LEFT L3L4 
 
LEFT L2L3 
HIGH 
 
HIGH 
93 LEFT L3L4 HIGH 
 137 
 
LEFT L1L2 
 
HIGH 
97 LEFT L3L4 
 
LEFT L3L4 
 
LEFT L3L4 
 
LEFT L2L3 
 
LEFT L3L4 
 
LEFT L2L3 
HIGH 
 
HIGH 
 
HIGH 
 
HIGH 
 
MEDIUM 
 
HIGH 
* Subjects are listed in ascending order by their Subject ID number (each was assigned a 
random Subject ID number 1-100)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 138 
7. DISCUSSION  
 
 
The main objective of this research project was to locate the level of Z joints in 
the lumbar spine that cavitated following SMT and side-posture positioning. Specific 
goals were set to refine previous methods (Goal 1 of Section 1.2), conduct recordings 
using the re-designed methods (Goal 2 of Section 1.2), and report the results of 
accelerometry recordings (Goal 3 of Section 1.2). The re-designed instrumentation and 
methodology were successful in accomplishing the main objective. Cavitations were 
recognized by their unique waveform pattern (?non-continuous? or ?damped? waveform 
pattern) (Macquarie University, 2008) and by the sequential response of accelerometers 
(Ross et al., 2004).  
The reliability study comparing the results of two observers (blinded to the results 
of one another) in identifying cavitations from the accelerometer recordings showed 
?almost perfect agreement? (Cohen, 1960; Landis and Koch, 1977; Gottman, 1995; Fleiss 
et al., 2003). Accelerometer recordings of cavitations were also in ?almost perfect 
agreement? with verbal reporting by the subjects and clinician. The clinician and subject 
were also in ?almost perfect agreement? (Cohen, 1960; Landis and Koch, 1977; Gottman, 
1995; Fleiss et al., 2003). 
Assessing the relationship between amplitude of cavitations and Z joint gapping 
in the ongoing study may be challenging, because the amplitude was affected by the 
impedance of the intervening tissue and the distance between the transducer and the joint. 
 139 
This is similar to the difficulty other authors have had in interpreting the severity of 
degenerative disease by waveform amplitude (Gay et al., 1987).  
Fifty-six cavitations were recorded from 40 subjects. Fifty-three cavitations were 
recorded from subjects randomized in Group 1 (SMT group) and 3 cavitations were 
recorded from subjects randomized in Group 2 (Control group in which subjects were 
only held in the side-posture position but did not received SMT). The range of cavitations 
for subjects included in Group 1 was 0-6 and in Group 2 was 0-1. Therefore, the 
hypothesis that subjects can cavitate during side-posture positioning was confirmed. 
Group 1 subjects (SMT) cavitated more than Group 2 subjects (side-posture position), 
putatively because of the additional force received by the Z joints and SIJs during SMT. 
 Fifty-three cavitations (53/56 = 94.63%) were recorded from joints located on the 
left side and 3 cavitations (3/56 = 5.43%) were recorded from the joints located on the 
right side. Reggars and Pollard (1995) conducted a study to determine the side of audible 
release in the cervical region using two microphones. Interestingly, they found that in 
94% of the subjects, audible release was on the same side as head and neck rotation (i.e., 
head and neck rotated to the right associated with cavitation on right side). Our findings 
agree with Reggars and Pollard, as the rotation in the lumbar spine during side-posture 
positioning and SMT in this study was the same as theirs in the cervical region (Reggars 
and Pollard, 1995). Subjects in our study were held in side-posture position, and SMT 
was applied on left side as subjects were laying on their right side. That is to say, a higher 
percentage of cavitations were recorded on the same side in which force was applied. 
A robust cavitation response was found in the targeted Z joints. Forty-one 
cavitations (41/56 = 73.21%) were recorded at targeted Z joints (Left L3/L4, Right 
 140 
L3/L4, Left L4/L5, Right L4/L5, Left L5/S1, and Right L5/S1) and 26.80% cavitations 
were recorded at non-targeted Z joints (Left L1/L2, Right L1/L2, Left L2/L3, Right 
L2/L3, Left SIJ and Right SIJ). Interestingly, a maximum of 46.43% cavitations were 
recorded at Left L3/L4 Z joint. These results indicate that although SMT is not precise 
(Ross et al., 2004) at producing cavitation at a specific single segment, it was precise up 
to 3 segments in our study. 
Multiple cavitations were recorded from the same joint. Double cavitations were 
recorded in 7 joints, and interestingly, 4 cavitations in 1 joint. There might be several 
reasons behind this. First, methods used in this study always included two thrusts, 
increasing the chances of getting multiple cavitations. Second, the resistance within the Z 
joint to the force of SMT may have been unique in these joints. Third, the specific 
mechanism of gapping in these joints may have been unique. That is, some joints may 
gap evenly along the joint surfaces while others may gap in stages, first at the medial or 
lateral border and then at the opposite side of the joint. Fourth, the shape of the Z joints 
associated with multiple cavitations might differ from the shape of the Z joints that 
cavitated once or did not cavitate. A future study is planned that will assess this last 
potential mechanism associated with multiple cavitations. In the study, the shape of the Z 
joints that cavitate multiple times will be compared with the shape of Z joints cavitating 
once, as well as those that did not cavitate at same segmental level and in the subjects of 
same gender. 
A maximum of 6 cavitations were recorded from a single subject randomized in 
Group 1, with 4 cavitations at one joint and 2 cavitations at another joint. As mentioned 
previously, this interesting case was the only subject on which data was collected in the 
 141 
late evening. There might be several reasons for getting multiple cavitations at this time. 
First, the disc is less hydrated in the evening as compared to the morning, increasing 
motion of the Z joints (Paajanen et al., 1994). Also, because of various activities 
performed by the subject throughout the day, the Z joints might be more lubricated in the 
evening, providing increased freedom of motion. Further study of relative lubrication 
(synovial fluid volume) in the morning vs. evening is needed to determine if this 
hypothesis is biologically plausible. 
Some accelerometer signals that were not caused by cavitation, known as artifacts 
or noise, provided a challenge to this study. The waveforms for artifacts were irregular 
with larger spikes and did not fit the smooth ?non-continuous? or ?damped? waveform 
pattern found in cavitations. In addition, accelerometers responded dramatically out of 
sequence when recording artifacts, whereas the accelerometers responded in a very 
predictable pattern when recording cavitations. Artifacts were generated by the clinician 
accidently touching the tape or accelerometers, movement in wires, movement in skin 
and muscles, movement of the subject?s body while applying force to cavitate joints, a 
subject not relaxing or opposing the clinician, and other yet undetermined 
(miscellaneous) reasons. Based on notes taken during data acquisition and assessment, 
combined with the unique pattern and nature of recorded waveform signals, artifacts were 
categorized into the following groups: flexible spine (or loose superficial fascia) subjects, 
hyper-muscular subjects (e.g., body builders), resisting or not relaxed subjects, unstable 
hand contact (UHC) (i.e., clinician?s hand contact), and miscellaneous.  It was not 
possible to assess cavitations from data that included resisting or not relaxed subjects or 
UHC artifacts. Consequently, subjects with these artifacts were excluded from data 
 142 
analysis for cavitations. To reduce the artifacts due to flexible spine (or loose superficial 
fascia) subjects, additional tape was applied to the accelerometers on all subjects? backs 
in side-posture position during the clinical study. In order to prevent resisting or not 
relaxed subject artifacts, subjects were asked to relax and not resist the study clinician. 
UHC artifacts were prevented by requiring subjects to wear cotton surgical scrub pants 
(rather than ?slippery shorts?) during the clinical study.   
Goal 4 of this study (discussed in Section 1.2) was: ?To assess the accelerometry 
methods used in the clinical study to determine the strengths and challenges (weaknesses) 
of the re-designed methods.? Several strengths and challenges of the re-designed methods 
used in this study were identified and are listed immediately below. 
Primary Strengths ? The new methods were able to: 
 
1. Differentiate cavitations from artifacts. 
2. Identify cavitations reliably (two reliability studies). 
3. Capture cavitations at the targeted Z joints and non-targeted Z joints during 
side-posture positioning and SMT. 
4. Record, archive cavitations, and artifacts using screenshots of the complete 
time scale. 
The primary challenge of the re-designed methods was the presence of artifacts. 
The challenge of differentiating between cavitations and artifacts was overcome by 
assessing the accelerometry data. Cavitations had a smooth, non-continuous (damped) 
waveform pattern and responded sequentially, whereas artifacts had a waveform pattern 
characterized by many large irregular spikes and the accelerometers responded in a 
 143 
chaotic sequence. Artifacts were able to be categorized based on the notes taken during 
data acquisition and the unique pattern and nature of waveform signals.    
Goal 5 (discussed in Section 1.2) was set to propose ideas for future studies that 
would further refine accelerometry methods and add acoustics methods. Future spine 
research combining microphones with accelerometers should use a surface microphone in 
addition to accelerometers. Recordings should be made during full ranges of motions 
(full flexion, extension, left and right lateral flexion, and left and right rotation), side-
posture positioning and SMT. This would allow for the recording of additional joint 
sounds known as ?crepitus?. Data should be collected from accelerometers and the 
microphone simultaneously, because each captures unique and complementary 
information (Jaskolska and Madeleine, 2007). Subjects in younger and older age groups 
should be sought to find unique information related to changes in joints with age. Lumbar 
x-rays of the subjects should be taken, and the amount of spinal degeneration should be 
graded from the x-rays (Cramer et al., 2008). Fast Fourier Transformation (FFT) should 
be performed to determine the unique characteristics of the sound waves recorded during 
the ranges of motions and SMT. Such frequency analysis of sound waves has been used 
to assess sounds associated with temporomandibular joint dysfunction (Heffez and 
Blaustein, 1986; Widmalm et al., 1992; Gallo et al., 1993). The amount of degeneration 
graded from the x-rays in the proposed study should be used to explore the possibility of 
unique waveform patterns associated with different amounts of degeneration. A 
relationship between degeneration or other joint pathology and analysis of sound waves 
has been successfully investigated in the knee (Chu et al., 1976). The processing capacity 
of the computer can be increased for longer-duration sampling at a higher sampling rate 
 144 
to prevent the computer from overloading the buffer during simultaneous data collection 
from 10 channels (9 accelerometers and 1 surface microphone). Kinesio tape (Kinesio 
tape, 2010) could be used to place accelerometers and the microphone on the subject?s 
body to prevent the tape from ?popping off? during data collection of the subject?s full 
ranges of motion. Kinesio tape allows for increased flexibility during ranges of motions 
compared with the strong adhesion and non-flexible tape used in this study. Ultrasound 
gel should be applied to the location where the microphone is placed, and the microphone 
could be placed on top of rubber pads. These procedures could minimize frictional noise 
between the skin and the microphone (Chu et al., 1976). The results of such a study could 
potentially initiate a line of investigation deepening the understanding of the mechanics 
of the spine and SMT by assessing differences found during various motions, SMT, and 
in spines with varying degrees of degeneration. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 145 
8. CONCLUSIONS  
 
 
The accelerometry methods used in this research project effectively identified the 
segmental level of cavitation, which was the main objective of this research project.  
The reliability study showed ?almost perfect agreement? (Cohen, 1960; Landis 
and Koch, 1977; Gottman, 1995; Fleiss et al., 2003) between the results of two observers 
(blinded to the results of one another) in identifying cavitations in the accelerometer 
recordings. 
Accelerometer recordings of cavitations were in ?almost perfect agreement? 
(Cohen, 1960; Landis and Koch, 1977; Gottman, 1995; Fleiss et al., 2003) with the 
subject and clinician. The clinician and subjects were also in ?almost perfect agreement? 
(Cohen, 1960; Landis and Koch, 1977; Gottman, 1995; Fleiss et al., 2003). 
Cavitations were recognized by their unique ?non-continuous? or ?damped? 
waveform pattern combined with the sequential response of accelerometers. Subjects 
randomized into Group 1 who received SMT cavitated significantly more often (96.67% 
of Group 1 subjects) than those who were just held in side-posture-positioning (30% of 
Group 2 subjects) (p< 0.0001). Most cavitations were recorded on the left-side (up-side, 
94.63%) and at targeted Z joints (73.20%). Instruments used in this study showed the 
capability of recording multiple cavitations from different joints and also (interestingly) 
multiple cavitations from the same joint. The range of cavitations for subjects in Group 1 
 146 
was 0-6 and Group 2 was 0-1. A maximum of 6 cavitations were recorded from a single 
Group 1 subject, with 4 cavitations at one Z joint and 2 cavitations at another Z joint. 
Artifacts had waveform patterns distinctly different from cavitations. The 
waveforms for artifacts were irregular with larger spikes and did not fit the smooth ?non-
continuous? or ?damped? waveform pattern found in cavitations. In addition, 
accelerometers responded dramatically out of sequence when recording artifacts, whereas 
the accelerometers responded in a very predictable pattern when recording cavitations. 
Based on notes taken during data acquisition and assessment,  combined with the unique 
pattern and nature of recorded waveform signals, artifacts were categorized into the 
following groups: flexible spine (or loose superficial fascia) subjects, hyper-muscular 
subjects (e.g., body builders), resisting or not relaxed subjects, unstable hand contact 
(UHC) of clinician (i.e., clinician?s hand contact) and miscellaneous. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 147 
REFERENCES 
 
 
All About Back and Neck Pain, (image on the internet), 2009 (cited 2010 Jul 1), 
Available from: 
http://www.allaboutbackandneckpain.com/understand/anatomy.asp 
 
Analog Devices, (ADXL330 datasheet on the internet), 2010 (cited 2010 Jul 7), 
Available from: http://www.analog.com/en/sensors/inertial- 
sensors/adxl330/products/product.html 
 
Avallone, E. A., and Baumeister, T., 1996, Marks' Standard Handbook for Mechanical 
Engineers, McGraw-Hill, Columbus, OH. 
 
Aszkler, C., 2005 (cited 2010 Jul 8), ?The Principles of Acceleration, Shock, and 
Vibration Sensors,? Available from: 
http://www.sensorsmag.com/sensors/acceleration-vibration/the-principles-
acceleration-shock-and-vibration-sensors-574 
 
Beffa, R., and Mathews, R., 2004, ?Does the Adjustment Cavitate the Targeted Joint? An 
Investigation into the Location of Cavitation Sounds,? J. Manipulative Physiol. 
Ther., 27(2), pp. 118-122. 
 
Brodeur, R., 1995, ?The Audible Release Associated with Joint Manipulation,? J. 
Manipulative Physiol. Ther., 18(3), pp. 155-164. 
 
Brown, P.  L., 1997, ?Muscle Sounds in Parkinson's Disease,? Elsevier., 349(9051), pp. 
533-535 
 
Bruel & Kjaer, 2010 (cited 2010 Jul 8), ?Choose the Right Accelerometer,? Available 
from: 
http://www.bksv.com/products/transducersconditioning/vibrationtransducers/acce
lerometers/choosingaccelerometers.aspx 
 
Bruel & Kjaer, 2010 (cited 2010 Jul 8), ?Engine Noise,? Available from: 
http://www.bksv.com/markets/aerospace/aero%20engines/predictnoiseperfomanc
e.aspx 
 
Bruel & Kjaer, 2010 (cited 2010 Aug 26), ?Product Data Sheet: Miniature Delta Tron 
Accelerometers,? Available from: http://www.bksv.com/doc/bp1841.pdf 
 
 148 
Bruel & Kjaer, 2010 (cited 2010 Aug 26), ?Product Data,? Available from: 
http://www.bksv.com/doc/bp1853.pdf  
 
Chu, M. L., Gradisar, I. A., and Railey, M. R., 1976, ?Detection of Knee Joint Diseases 
Using Acoustical Pattern Recognition Technique,? J. Biomech. Eng., 9(3), pp. 
111-114. 
 
Cohen, J., 1960, ?A Coefficient of Agreement for Nominal Scales,? Educational and 
Psychological Measurement., 20, pp. 37-46. 
 
Cramer, G. D., Tuck, N. R., Knudsen, J. T., Fonda, S. D., Schliesser, J. S., Fournier, J. T., 
and Patel, P.,  2000, ?Effects of Side-Posture Positioning and Side-Posture 
Adjusting on the Lumbar Zygapophysial Joints as Evaluated by Magnetic 
Resonance Imaging: A Before and After Study with Randomization,? J. 
Manipulative Physiol. Ther., 23(6), pp. 380-94. 
 
Cramer, G. D., Gregerson, D. M., Knudsen, J. T., Hubbard, B. B., Ustas, L. M., and 
Cantu, J. A., 2002, ?The Effects of Side-Posture Positioning and Spinal Adjusting 
on the Lumbar Z joints: A Randomized Controlled Trial with Sixty-Four 
Subjects,? Spine., 27(22), pp. 2459-2466. 
 
Cramer, G. D., and Darby, S. A., 2005, Basic and Clinical Anatomy of the Spine, Spinal 
Cord, And Ans, Elsevier Mosby, St. Louis, Chap. 2, pp. 18-34. 
 
Cramer, G. D., Henderson, C. R., Pocius, J., Little, J.W., and Grieve, T. J., 2008, 
?Grading System for Osteophytic Changes in Zygapophysial (Z) Joints Following 
Induced Segmental Hypomobility in the Rat,? Abstract 773.2, Experimental 
Biology Annual Meeting, San Diego, California. 
 
Cramer, G. D., Ross, K., Pocius, J., Cantu, J. A., Laptook, E., Fergus, M., Gregerson, D., 
and Selby, S., 2008, ?Assessing Cavitation and Z Joint Gapping Following Side-
Posture Spinal Adjusting: A Feasibility Case Series,? Thirteenth Annual Research 
Agenda Conference, Washington, DC, Journal of Chiropractic Education., pp. 22-
54. 
 
Dimension Engineering, No Date (cited 2010 Jul 8), ?A Beginner?s Guide to 
Accelerometers,? Available from: 
http://www.dimensionengineering.com/accelerometers.htm  
 
Extech Instruments (homepage on the internet), 2010 (cited 2010 Jul 8), Available from: 
http://www.extech.com/instruments  
 
Fleiss, J. L., Levin, B., and Paik, M. C., 2003, Statistical Methods for Rates and 
Proportions, John Wiley & Sons, Hoboken, NJ. 
 
 149 
Flynn, T. W., Fritz, J. M., Wainner, R. S., and Whitman, J. M., 2003, ?The Audible Pop 
is Not Necessary for Successful Spinal High-Velocity Thrust Manipulation in 
Individuals with Low Back Pain,? Arch. Phys. Med. Rehabil., 84(7), pp. 1057-
1060. 
 
Gallo, L. M., Airoldi, R., Ernst, ?., and Palla, S., 1993, "Power Spectral Analysis of 
Temporomandibular Joint Sounds in Asymptomatic Subjects,? J. Dent. Res., 
72(5), pp. 871-875. 
 
Gautschi, G., 2002, Piezoelectric Sensors: Force, Strain, Pressure, Acceleration and 
Acoustic Emission Sensors, Materials and Amplifiers, Springer, Berlin. 
 
Gay, T., Bertolami, C. N., , Donoff, R. B., Keith, D. A., and Kelly, J. P., 1987, ?The 
Acoustical Characteristics of the Normal and Abnormal Temporomandibular 
Joint,?  J. Oral and Maxillofacial Surgery, 45(5), pp. 397-407. 
 
Gottman, J. M., 1995, The Analysis of Change, Lawrence Earlbaum Associates, 
Mahwah, NJ. 
 
Heffez, L., and Blaustein, D., 1986, ?Advances in Sonography of the 
Temporomandibular Joint,? Oral Surg. Oral Med. Oral Pathol.,  62(5), pp.486-
495. 
 
Horowitz, S.,  Nishida, T.,  Cattafesta, L., and  Sheplak, M., 2006, ?A Micromachined 
Piezoelectric Microphone for Aeroacoustics Applications,? Solid-State Sensor, 
Actuator and Microsystems Workshop, Hilton Head, SC. 
 
Janse, J., 1976, Principles and Practice of Chiropractic: An Anthology, Kjellberg & Sons, 
Inc., Wheaton, IL, Chap.6. 
 
Jaskolska, A., and Madeleine, P., 2007, ?A Comparison Between Mechanomyographic 
Condenser Microphone and Accelerometer Measurements During Submaximal 
Isometric, Concentric and Eccentric Contractions,? J. Electromyogr. Kinesiol., 
17(3), pp. 336-347. 
 
Kinesio tape (homepage on the internet), 2010 (cited 2010 Jul 24), Available from: 
http://www.kinesiotaping.com  
 
Landis, J. R., and Koch, G. G., 1977, ?The Measurement of Observer Agreement for 
Categorical Data,? Biometrics., 33(1), pp. 159-174. 
 
Dimension Engineering, No Date (cited 2010 Jul 8), ?A Beginner?s Guide to 
Accelerometers,? Available from: 
http://www.dimensionengineering.com/accelerometers.htm  
 
 150 
Macquarie University, 2008 (cited 2010 Aug 24), ?Non-Continuous Waveform,? 
Available from: http://clas.mq.edu.au/acoustics/waveforms/damping.html 
 
McFadden, K. D., and Taylor, J. R., 1990, ?Axial Rotation in the Lumbar Spine and 
Gaping of the Zygapophysial Joints,? Spine 15(4), pp. 295-299. 
 
Meal, G. M., and Scott, R. A., 1986, ?Analysis of the Joint Crack by Simultaneous 
Recording of Sound and Tension,? J. Manipulative Physiol. Ther., 9(3), pp. 189-
95. 
 
Media College (image on the internet), No Date (cited 2010 Jul 8), Available from: 
http://www.mediacollege.com/audio/microphones/condenser.html 
 
National Instruments, 2010 (cited 2010 Aug 26), ?NI PXIe-6356,? Available from: 
http://sine.ni.com/nips/cds/view/p/lang/en/nid/207418 
 
National Instruments, 2010 (cited 2010 Aug 26), ?NI BNC 2090,? Available from: 
http://sine.ni.com/nips/cds/view/p/lang/en/nid/1177 
 
National Instruments, 2010 (cited 2010 Aug 26), ?NI PXIe-1073,? Available from: 
http://sine.ni.com/nips/cds/view/p/lang/en/nid/207401 
 
National Instruments, 2010 (cited 2010 Aug 26), ?NI PXIe- Express Card 8360,? 
Available from: http://sine.ni.com/nips/cds/view/p/lang/en/nid/202665  
  
Nokes, L., Fairclough, J. A., Mintowt-Czyz, W. J., Mackie, I., and Williams, J., 1984, 
?Vibration Analysis of Human Tibia: The Effect of Soft Tissue on the Output 
From Skin-Mounted Accelerometers,? J. Biomed. Eng., 6(3), pp. 223-6. 
 
Nokia, 2010 (cited 2010 Jul 8), ?Nokia 5500 Sport Support,? Available from: 
http://europe.nokia.com/support/product-support/nokia-5500-sport 
 
Paajanen, H., Lehto, I., Alanen, A., Erkintalo, M., and Komu, M., 1994, ?Diurnal Fluid 
Changes of Lumbar Discs Measured Indirectly By Magnetic Resonance 
Imaging,? J. Orthop. Res., 12, pp. 509-514. 
 
Paritsky, A., and Kots, A., 1997, ?Fiber Optic Microphone as a Realization of Fiber Optic 
Positioning Sensors,? Proc. of International Society for Optical Engineering 
(SPIE), 3110, pp.  408?409. 
 
PCB Piezotronics, Inc, (image on the internet), 2009 (cited 2010 Jul 8), Available from: 
http://www.pcb.com/Accelerometers/Sensing_Technologies.asp 
 
PCB Piezotronics, Inc, (image on the internet), 2009 (cited 2010 Jul 8), Available from: 
http://www.modalshop.com/techlibrary/Microphone%20handbook%20PCB.pdf 
 
 151 
PCB Piezotronics, Inc, 2009 (cited 2010 July 8), ?Acoustic Measurement Products?, 
Available from: http://www.pcb.com/auto/Acoustic_Measurement_Products.asp  
 
Reddy, N. P., Bruce, M. R., Verrall, E., and Joshi, A., 2001, ?Noninvasive Measurement 
of Acceleration at the Knee Joint in Patients with Rheumatoid Arthritis and 
Spondyloarthropathy of the Knee,? Annals of Biomedical Engineering, 29, pp. 
1106?1111. 
 
Reggars, J. W., and Pollard, H. P., 1995, ?Analysis of Zygapophysial Joint Cracking 
During Chiropractic Manipulation,? J. Manipulative Physiol. Ther., 18(2), pp. 65-
71. 
 
Rodts, M., 2008 (updated 2010 Dec 1; cited 2010 Jul 1), ?Your Healthy Spine,? 
Spineuniverse, Available from: http://www.spineuniverse.com/anatomy/your-
healthy-spine 
 
Ross, K. J., Bereznick, D. E., and McGill, S. M., 2004, ?Determining Cavitation Location 
During Lumbar and Thoracic Spinal Manipulation. Is Spinal Manipulation 
Accurate and Specific?,? Spine, 29(13), pp. 1452?1457. 
 
Sessler, G. M., and West, J. E., 1962, ?Self-Biased Condenser Microphone with High 
Capacitance,? Journal of the Acoustical Society of America, 34, pp. 1787?1788. 
 
Siffert, R. S., and Kaufman, J. J., 1996, ?Acoustic Assessment of Fracture Healing, 
Capabilities and Limitations of a Lost Art,? Am. J. Orthop., 25(9), pp. 614-618. 
 
Snedecor, W. G., and Cochran, W. G., 1967, Statistical Methods, The Iowa State 
University Press, Ames, Iowa. 
 
Stmicroelectronics, 2010 (cited 2010 Jul 8), ?MEMS motion sensor,? Available from: 
http://www.st.com/stonline/products/literature/ds/12726/lis302dl.htm 
 
Taiwanspinecenter, (image on the internet), No Date (cited 2010 Jul 1), Available from: 
http://www.taiwanspinecenter.com  
 
Teodorczyk-Injeyan, J. A., Injeyan, H. S., and Ruegg, R., 2006, ?Spinal Manipulative 
Therapy Reduces Inflammatory Cytokines But Not Substance P Production in 
Normal Subjects,? J. Manipulative Physiol. Ther., 29, pp. 14-21. 
 
Teodorczyk-Injeyan, J. A., Injeyan, H. S., McGregor, M., Harris, G. M., and Ruegg, R., 
2008, ?Enhancement of In Vitro Interleukin-2 Production in Normal Subjects 
Following a Single Spinal Manipulative Treatment,? J. Chiropr. Osteopat., 16(5). 
 
University of Maryland Spine Center, 2007 (updated 2009 Jul 21; cited 2010 Jul 1), 
Available from: www.umm.edu/spinecenter/education 
 
 152 
Unsworth, A., Dowson, D., and Wright, V., 1971, ?Cracking Joints, A Bioengineering 
Study of Cavitation in the Metacarpophalangeal Joint,? Ann Rheum Dis., 30(4), 
pp. 348?358. 
 
Widmalm, S. E., Westesson, P. L., Brooks, S. L., Hatala, M. P., and Paesani, D., 1992, 
?Temporomandibular Joint Sounds: Correlation to Joint Structure in Fresh 
Autopsy Specimens,?  Am. J. Orthod. Dentofacial Orthop., 101(1), pp. 60-69. 
 
Wikipedia, (image on the internet), No Date (updated 2007 Oct 22; cited 2010 Jul 8), 
Available from: http://en.wikipedia.org/wiki/File:SchemaPiezo.gif  
 
Wikipedia, (image on the internet), No Date (updated 2008 Jan 25; cited 2010 Jul 8), 
Available from: http://en.wikipedia.org/wiki/File:PiezoAccelTheory.gif 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 153 
APPENDICES 
 
 
A. Technical Specifications 
 
1) Bruel & Kjaer Delta Tron4507 Accelerometer 
Bruel & Kjaer Delta Tron4507 accelerometer was used in our research project as shown 
in figures 4.11, A.1 (Bruel & Kjaer, 2010) because of its size and high sensitivity. 
Reference Sensitivity: 9.766mV/ms-2 
Frequency Range: 0.3Hz to 6 kHz 
Mounted Resonance Frequency: 18 kHz 
 
Electrical 
Bias Voltage: At full temperature and current range: +12V +-1V 
Power Supply requirements:  Constant Current: +2 to +20mA 
                                                Unloaded Supply Voltage: +24V to +30 V 
Output Impedance: Less than 2 ohms 
Recommended cables: AO 1382 
                                     AO 0531 
                                    AO 0463 
Environmental 
Temperature Range: -54 to +121?C 
Temperature Coefficient of Sensitivity: +0.09 %/?C 
 154 
Magnetic Sensitivity: 3ms-2/T 
Maximum Non-destructive Shock: 50 kms-2 peak (5000 g peak) 
Humidity: 90% RH non-condensing 
Mechanical 
Case Material: Titanium ASTM Grade 2 
Sensing Element: Piezoelectric, Type PZ 23 
Construction: Theta Shear 
Sealing: Welded 
Weight: 4.8 gram 
Electrical Connector: 10-32 UNF-2A 
Mounting Surface Flatness: < 3?m 
 
Geometric Dimensioning: 
 
Figure A.1: Geometric dimensioning of Bruel & Kjaer Delta Tron4507 accelerometer 
Polarity of the electrical signal is positive for acceleration in the direction of the arrow on 
the drawing. 
 155 
2) Bruel & Kjaer 2693 NEXUS Signal Conditioning Amplifier  
Connector: TNC 
Grounding: Single-ended or floating 
Differential Charge: 10 nC (peak);  
Common Mode Voltage: 4.2V (peak) 
Amplifier gain: 0.1mV/pC to 10 V/pC (.20 to +80 dB gain with 1 nF transducer 
capacity); 
Transducer sensitivity range: 10.19 to 10.6 C/MU 
(MU = mechanical units: m/s2; g, N, lb., Pa) 
Frequency range: Acceleration: 0.1Hz to 100 kHz (transducer cable length <10m) 
Magnetic Field: <0.2 fC/ (A/m) 
Electromagnetic Field: <20 fC/ (V/m) or <4 fC/V 
Vibration (10 to 500Hz): <30 fC/ (m/s2) 
 
DeltaTron? Input 
Connector: BNT 
Grounding: Single-ended or floating 
Input impedance: 1M? || 100 pF (AC coupled) 
Differential Voltage: ?31.6V (peak) 
Common Mode Voltage: 4.2V (peak) 
Amplifier gain: .20 to +60 dB gain (80 dB with reduced specs.) 
Transducer sensitivity range: 10.12V/MU to 103V/MU 
 
 156 
General Specifications:  
Power Supply: Internal battery  
Each amplifier is capable of supporting four transducers as shown in figure A.2 (Bruel & 
Kjaer, 2010), displayed on the first column. Amplifier uses Nickel-Metal Hydride 
rechargeable battery supporting SMBus and on-battery charge level meter. It provides 
typically 15 hours of continuous use with a single channel and 4 hours with four channels 
without backlighting and without optional filters. With backlighting on, and with optional 
filters, battery provides typically 3 hours of continuous use. 
External dc power input: Complies with ISO 7637.1 (12 V) and 7637.2 (24V) 
Input Range: 10 to 33VDC 
Mains supply: Supported via supplied Mains Adaptor ZG-0426 (included), 90 .264VAC, 
40.65Hz 
 
Display Interface 
Display: 64 ? 128 pixel graphical display with back-lighting on/off  
Overload detection: On both common-mode and differential signals applied before filters. 
LED overload indication at the front panel and overload indication via RS.232 control 
interface. 
Dynamic range: .30 to +10 dB V (peak) 
Resolution: 1 dB 
 157 
 
Figure A.2: Bruel & Kjaer 2693 NEXUS signal conditioning amplifier (an accelerometer 
is shown to the right of the microphones, microphones were not used in this study) 
 
 
3) NI PXIe-6356 Simultaneous X Series Data Acquisition Card  
  8 simultaneous analog inputs at 1.25 MS/s/channel with 16-bit resolution (National 
Instruments, 2010). 
  Two analog outputs, 3.33 MS/s, 16-bit resolution, ?10 V. 
  24 digital I/O lines (8 hardware-timed up to 10 MHz). 
  Four 32-bit counter/timers for PWM, encoder, frequency, event counting, and more. 
  Analog and digital triggering and advanced timing with NI-STC3 technology. 
  Support for Windows 7/Vista/XP/2000. 
 
 158 
Overview 
NI X Series multifunction data acquisition (DAQ) devices provide a high-throughput PCI 
Express bus, excellent NI-STC3 timing and synchronization technology, and multicore-
optimized driver and application software. 
High-Throughput PCI Express Bus 
PCI express offers dedicated bandwidth of up to 250 MB/s in each direction to each 
device, and X series devices feature a native PCI express interface with optimizations for 
high throughput and low latency. 
NI-STC3 Technology 
NI-STC3 timing and synchronization provides advanced timing functionality, including 
independent analog and digital timing engines, retriggerable measurement tasks, and four 
counter/timers with more functionality than previous technologies. 
Technical Specifications 
PXI Bus Type: PXI express, PXI hybrid compatible 
Operating System/Target: Real-Time, windows 
LabVIEW RT Support: Yes 
DAQ Product Family: X Series 
Measurement Type: Digital, frequency, quadrature encoder, voltage 
RoHS Compliant: Yes 
 
 
 159 
Analog Input 
Channels: 0, 8 
Single-Ended Channels: 0 
Differential Channels: 8 
Resolution: 16 bits 
Sample Rate: 1.25MS/s 
Throughput: 10 MS/s 
Max Voltage: 10 V 
Maximum Voltage Range: -10 V, 10 V 
Maximum Voltage Range Accuracy: 2.566 mV 
Minimum Voltage Range: -1 V, 1 V 
Minimum Voltage Range Accuracy: 307 ?V 
Number of Ranges: 4 
Simultaneous Sampling: Yes 
 
Analog Output 
Channels: 2 
Resolution: 16 bits 
Max Voltage: 10 V 
Maximum Voltage Range: -10 V, 10 V 
Maximum Voltage Range Accuracy: 3.066 mV 
Minimum Voltage Range: -5 V, 5 V 
Minimum Voltage Range Accuracy: 1.526 mV 
 160 
Digital I/O 
Bidirectional Channels: 24 
Input-Only Channels: 0 
Output-Only Channels: 0 
Number of Channels: 24, 0, 0 
Timing: Software, hardware 
Clocked Lines: 8 
Max Clock Rate: Sinking, sourcing 
Logic Levels: TTL 
Input Current Flow: Sinking, sourcing 
Output Current Flow: 10 MHz 
Programmable Input Filters: Yes 
Current Drive Single: 24 mA 
Current Drive All: 576 mA 
Watchdog Timer: Yes 
Maximum Input Range: 0 V, 5 V 
Maximum Output Range: 0 V, 5 V 
 
Counter/Timers 
Counters: 4 
Number of DMA Channels: 8 
Buffered Operations: Yes 
GPS Synchronization: No 
 161 
Maximum Range: 0 V, 5 V 
Max Source Frequency: 100 MHz 
Minimum Input Pulse Width: 10 ns 
Pulse Generation: Yes 
Resolution: 32 bits 
Time base Stability: 50 ppm 
Logic Levels: TTL 
 
Figure A.3: NI PXIe-6356 Simultaneous X Series Data Acquisition Card  
 
4) NI BNC-2090 Rack-Mounted BNC Terminal Block 
 
 
Figure A.4: NI BNC-2090 rack-mounted BNC terminal block 
The National Instruments BNC-2090 as shown in figure A.4 (National Instruments, 
2010) is a shielded rack-mountable connector block having 22 signal-labeled BNC 
connectors and 28 spring terminal blocks to simplify connections to I/O signals from NI 
 162 
data acquisition devices. It connects to any 68-pin M series, S series, E series, or B series 
DAQ device. The NI BNC-2090 also includes component locations for passive signal 
conditioning. 
  Shielded, rack-mountable BNC adapter chassis 
  22 BNC connectors for analog, digital, and timing signals 
  28 spring terminals for digital/timing signals 
 
5) NI PXIe-1073 CHASSIS 
  3 hybrid slots and 2 PXI express slots - every slot accepts PXI express modules 
  150 W total power available from 0 to 50 ?C 
  Integrated MXI-express controller included 
  Up to 250 MB/s per-slot dedicated bandwidth and 250 MB/s system bandwidth 
  Compatibility with PXI, PXI express, compact PCI, and compact PCI express modules 
 
Figure A.5: NI PXIe-1073 chassis 
The NI PXIe-1073 chassis with integrated controller is designed for variety of 
applications. It accepts PXI express modules in every slot and supports standard PXI 
hybrid-compatible modules in up to three slots. The chassis features compact, rugged 
 163 
packaging as well as acoustically quiet operation, which makes it ideal for both portable 
and desktop systems. The sturdy design of the chassis enables it to operate in 
environments from 0 to 50 ?C. The chassis also incorporates all of the latest PXI 
specification features including PXI and PXI express module support and built-in timing 
and synchronization. 
 
Specifications 
General 
Form Factor: PXI platform 
PXI Bus Type: PXI express, PXI hybrid compatible 
Operating System/Target: Any 
LabVIEW RT Support: Yes 
 
Chassis 
Power Supply: AC 
Number of Slots: 5 
Number of PXI Express Peripheral Slots: 2 
Number of Hybrid Peripheral Slots: 3 
Maximum System Bandwidth: 250 
Accepts both 3U PXI and Compact PCI Modules: Yes 
Integrated Controller: Yes 
Remote Power Inhibit Control and Voltage Monitoring: No 
Total Available Power: 150 
 164 
Input Voltage Range: 100 V, 240 V 
Input Frequency Range: 47 Hz, 63 Hz 
Field-replaceable Power Supply: No 
Auto Fan Sound Pressure Level: 43.3 
Auto Fan Sound Power: 51.3 
High Fan Sound Pressure Level: 58.3 
High Fan Sound Power: 64.6 
Auto/High Fan Selector: Yes 
Number of Fans: 1 
 
6) NI PXIe-Express Card 8360 
 
 
Figure A.6: NI PXIe-express card 8360 
 
  Sustained throughput up to 214 MB/s 
  Software-transparent link that requires no programming 
  Cabling up to 7 m with rugged screw-in connectors 
  Ability to use the same PXI express module (NI PXIe-8360) and cable as MXI-express 
for PXI express 
 165 
  Ideal for mobile applications such as field tests, in-vehicle data logging, NVH, NDT, 
and RF testing (and movable accelerometry cart used in this study figure 4.2) 
  Direct laptop control of PXI Express/Compact PCI Express 
 
Specifications 
General 
Form Factor: PXI platform 
PXI Bus Type: PXI express 
Operating System/Target: Windows, real-time 
LabVIEW RT Support: Yes 
 
Controller 
Controller Type: Remote 
Communication Technology: PCI express 
Sustained Performance: 214 MB/s 
Cable Material: Copper 
Maximum Cable Length: 7 m 
Maximum Links per Host Card: 1 
Slot Requirement: 1 
 
 
 
 
 166 
B. Auburn University Institutional Review Board (IRB) Approval  
 
  
 
 
 167 
  
 
 
 
 
 168 
  
 
 
 
 
 
 169 
  
 
 
 
 
 
 170 
  
 
 
 
 
 
 171 
  
 
 
 
 
 
 172 
 
 
 
 
 
 
 
 173 
  
 
 
 
 
 
 174 
  
 
 
 
 
 
 175 
 
 
 
 
 
 
 
 176 
  
 
 
 
 
 
 177 
  
 
 
 
 
 
 178 
 
 
 
 
 
 
 
 179