3.4.1 Participant Testing; Study 1 and Study 2
Isometric ILEX strength was tested twice, on separate days (at least 72 hours apart in order to avoid the effects of residual fatigue or soreness) both before and after the intervention. Each test using the lumbar extension machine involved maximal voluntary
isometric contractions at various angles through the participant’s full ROM. Details of the
full test protocol using the lumbar extension machine and details of the restraint mechanisms have been documented previously elsewhere (Graves et al., 1990a) and the
equipment manufacturer’s operation instructions for conducting testing are included in Appendix 7.4. During the first and second to last visit to the laboratory, participants were required to complete the VAS and the ODI, and also to have their standing ROM measured using the modified Schober’s test. Gait data were collected using the Vicon system during the third visit to the laboratory, and also during the participant’s final visit to the laboratory after the intervention period.
3.4.1.1 Three Dimensional Motion Analyses
Due to the lumbar spine’s capacity to rotate about three orthogonal axes, a three
dimensional approach was used for data collection. Ten cameras were set up and angled in a manner so as to reduce hidden spots that might obscure data collection. Figure 11 shows the setup of the cameras relative to the participant during walking trials as viewed using the Nexus software. The cameras identified reflective markers attached to the participant and output three dimensional coordinates for each marker. Data were recorded for 5 walking trials both pre and post intervention. Participants walked barefoot from one end of a marked runway to the other that was 8 metres in length at their free walking speed. The first full gait cycle captured where the participants entered the calibration volume during each walking trial was used.
154 | P a g e
Figure 11. Three dimensional motion camera set up
3.4.1.2 Biomechanical Model
The body of interest for the current study was the lumbar spine considered from S1 to T12 relative to the pelvis. For the purpose of analysis the lumbar spine was modelled as a rigid segment. The reasoning for not considering intervertebral segment movements was due to the small segments ranging from S2 to T10 always bending laterally toward the support leg with little variation between segments (Syczewska et al., 1999). Lumbar spine data were collected through three axes using the same model previously described by Schache et al. (2002a; 2002b), which has been shown to have high overall repeatability of
angular parameters (Schache et al., 2002b).
3.4.1.3 Marker Set Up
All markers were placed by the same investigator for all gait trials. Reflective markers (Figure 12) were placed over anatomical landmarks on the pelvis at both anterior superior iliac spines (ASIS) and at the midpoint of the posterior superior iliac spine (PSIS) using double sided adhesive tape. Reflective markers were also used upon a thoraco-lumbar marker cluster similar to that used by Schache et al., (2002a; 2002b). As with the
biomechanical model, this marker set up has been previously described elsewhere (Schache et al., 2002a; 2002b). The only alteration in this present study was the use of a
Participant
155 | P a g e flexibly-based wand marker for the thoraco-lumbar cluster. Two additional markers were secured equidistant either side of the midpoint of the wand markers base. This was placed in the same position over the 12th thoracic spinous process with the mid-point of the base
located over T12. The base was secured also using double sided adhesive tape. This removed the requirement for an elastic thoracic strap. T12 was first located using the technique suggested in Gray’s Anatomy for Students (Drake et al., 2008). This location was confirmed whilst the participant was in a flexed standing position, supporting themselves upon a stool, by palpation and counting of the spinous processes from this marked point down to the sacrum, and then double checked by counting back up to the marked spinous process.
Figure 12. Marker arrangement
3.4.1.4 Kinematic Data
Variability of angular kinematics of the lumbar spine about the three described axes relative to the pelvic segment was of primary interest (i.e. movement of the thoraco- lumbar marker cluster with respect to the pelvic markers). The Vicon Nexus software was used to run a Bodybuilder (Vicon, Oxford) code pipeline to calculate joint angles as outputs using Cardan (Euler) angles. The angles were calculated in the following order; 1) sagittal, 2) frontal, and 3) transverse. As with the biomechanical model, the Bodybuilder code used was the same as used by Schache et al. (2002a; 2002b). Data were filtered
using a low pass Butterworth filter (fourth order, cutoff frequency determined for each individual participant as sum of residuals closest to zero using 4Hz, 6Hz, 8Hz, 10Hz, and
156 | P a g e 12Hz) and normalised to percentage gait cycle corresponding to initial right heel contact (0%) and subsequent right heel contact (100%). Heel contacts were identified as the lowest vertical displacement of a right heel marker. Stride duration and length was also calculated using the horizontal displacement of the right heel marker from initial right heel contact and subsequent right heel contact. Mean values for angular displacements, stride- to-stride intra-subject variability using CVp and CVo, were calculated for lumbar spine
kinematics relative to the pelvis across all three planes of movement.
Intra-subject variability in the mean ensemble average has been typically calculated using
Winter’s (1983) coefficient of variation (Winter’s CV) in studies of lumbar kinematic variability in chronic LBP (Vogt et al., 2001). Thus to ensure comparability between the population used in this study with the coefficient of variation reported in earlier study of chronic LBP participants, intra-subject variability was calculated using Winter’s CV. However, the use of this method has recently been criticised due to the effect of waveform mean offsets altering relative variability away from the true variability in the system
(O’Dwyer et al., 2009). O’Dwyer et al. (2009) noted that variability of mean offsets (CVo)
and waveform pattern variability (CVp) should be calculated separately to account for the
different information they provide; CVo being determined by the reference frame used,
identification of anatomical landmarks, markers and their configuration, whereas CVp is
more representative of repeatability of motor performance. Adding to this, the model used in this study has been examined for within-day repeatability previously and it was reported that marker reapplication errors and their effect upon daily mean offsets were the main source of concern (Schache et al., 2002b). Thus Winter’s CV (Eq. (1)), CV
p (Eq. (2))and
CVo (Eq. (3)) were all calculated using the following equations:
157 | P a g e Where N is the number of intervals over the stride period, Xi is the mean waveform at the ith interval and ithe standard deviations about Xi.
To counteract the effects of mean offset variability in examining the variability in the waveform pattern the raw waveforms for each subject may first be transformed to zero mean before averaging. The same approach as in Eq. (1) may then be employed to compute a CV for pattern variability (CVp):
Eq. (2)
Where Xi(zero)is the average of the zero mean waveforms at the ith interval and i(zero)
the standard deviation about Xi(zero). Xi(zero) is calculated:
Where S is the number of subjects, Xji the raw waveform value for subject j at the ith
interval and the offset for subject j over all intervals in the stride period. i(zero) is calculated as:
In line with the computation in Eq. (2), the standard deviation of the mean offsets of the raw data is also compared with the grand mean of the absolute value of the zero mean waveforms over all intervals:
158 | P a g e
Where is the standard deviation of the
pooled subjects’ offsets and the mean offset over all subjects.
This allowed differentiation of offset variability (CVo) from pattern variability (CVp), the
latter being better representative of motor performance repeatability (for further discussion regarding the calculation please refer to O’Dwyer et al., 2009).
3.4.2 Participant Testing: Study 3
Participants underwent testing for ILEX strength and seated stadiometry at three points throughout the study (T1, T2, and T3). The ILEX test days were separated by at least 72 hours in order to avoid the effects of residual fatigue or soreness. At each time point participants were also required to complete the VAS and the ODI.
All stadiometry measurements were completed at the same time of day and participants were instructed to avoid heavy lifting for at least two days prior to testing (McGill et al., 1996). Measurements were conducted at the same time of day in order to control for diurnal variation. In order to normalise spine height prior to measurement the participant was instructed to lie in the supine position for 10 minutes with his or her hands resting on the stomach, head in a neutral position and supported by a pillow, and legs uncrossed with a pillow under the knees for support, as per the standard procedures used in the extant literature (Magnusson et al., 1996; Stothart & McGill, 2000; Rodacki et al., 2001; Kanlayanaphotporn et al., 2002; Rodacki et al., 2003). A custom set-up (See Figures 9 and 10) was used in combination with the same wall mounted stadiometer used for standing measurements. The wall mounted stadiometer calibration was checked before every test. Once 10 minutes elapsed participants were seated in the stadiometer setup
159 | P a g e with their sacral crest against the rear board of the seat, hip, knee and ankle angles at 90o, and arms rested comfortably on a pillow across their lap. A line traced along the
centre of the wooden seat was used to guide the participants in sitting centred when
moving into the seat. The participants’ feet were supported by mats to ensure hip, knee
and ankle angles were at 90o with the number of mats used recorded and used during
each test. Five anatomical points were identified and custom built adjustable rods were used to note the position of these for repeated testing (Healey et al., 2011). The points identified were: 1) the most posterior distension of the head; 2) the deepest point of the cervical lordosis; 3) the most prominent point of the thoracic kyphosis; 4) the deepest point of the lumbar lordosis; 5) the buttocks at the sacral crest (against the seat backboard). Control of these points (by noting during initial testing and replicating throughout further testing the vertical, horizontal and coronal position of the postural rods) ensured that participants adopted the same posture during all testing. After participants were seated in the stadiometer their heads were aligned in the Frankfurt plane to control their position and they were instructed to breathe in deeply maintaining their posture. They were instructed to hold their breath for 2-3 seconds whilst the head platform of the stadiometer was lowered until it made contact with the top of the head and measurement was taken. The testing was conducted by the present author. However, measurements were recorded by a research assistant and the results not disclosed to the primary investigator until both pre- and post-test data were collected in order to avoid investigator
bias. The measurement dial on the stadiometer was obscured from the researchers’ view
during testing. Ten repeated measurements were taken over a period of ~3 - 3.5 minutes with the participant remaining in the stadiometer between measurements (Stothart & McGill, 2000). The reliability of this custom set-up was also examined and details of this investigation are provided in the appendices (Appendix 7.6).
160 | P a g e