February 5, 2008; Märith Bergström-Isacsson (marith.bergstrom-isacsson@jll.se)
Advantages and Disadvantages of Doing a Research Study Using Quantitative Methods
Introduction
In this short paper I will briefly describe my study and discuss the advantages and disadvantages of doing a research study using quantitative methods. As this is an ongoing Doctoral study I will only share my thoughts about method, and will not be presenting results, at this stage.
The aim for my study is to investigate emotional behaviour in response to music and vibroacoustic therapy in patients with Rett syndrome (RTT) referred to the Swedish National Rett Centre over a period of two years. My sample included a clinical group of 32 participants and a comparison group of ten children aged 1-5 years old with normal development pattern. The purpose is also to investigate if preliminary observations from an earlier study are reproducible in a different sample from the Rett population. The hypotheses of this investigation were formulated on the basis that emotional output would influence the brainstem autonomic activity in the following manner:
- Emotional excitement will increase brainstem sympathetic activity above the resting baseline level.
- Relaxing or calming emotional response will increase brainstem parasympathetic activity above the resting baseline level.
- Arousal response without evoking a relaxed or excitement states will only cause physiological arousal measurable in the brainstem autonomic activity. (Bergström-Isacsson, Julu, & Witt Engerstrom, 2006). Because of the laboratory like the situation, how the study was done and the kind of measurable data collected (baseline and six independent variables) I found that a quantitative approach and a within subject study was the most suitable (Prickett, 2005).
Theory
The body has different types of nerves (brain nerves, spinal cord nerves and nerves that cannot be controlled at will). The autonomic nervous system is among those nerves that cannot be guided at will.
The function of the autonomic nervous system is the basis for this study as this system is part of the motor output of emotion (Best & Taylor, 1966; Guyenet et al., 1996). The autonomic nervous system is made up of the sympathetic and the parasympathetic systems. One can, with girls with Rett Syndrome, observe that the sympathetic and the parasympathetic systems are not in balance. This may be due to immaturity of the brainstem and this disequilibria lies behind considerable concern and anxiety found in the population. They are easily excitable, often appear stressed, find it difficult to relax and calm down and have difficulties in breathing.
These two sub-divisions of the autonomic nervous system influence the most basic function of the body including among other things breathing the heart rate and blood pressure. The sympathetic part affects most of the body’s inner organs and the brainstem sympathetic activity has a very close relationship with the changes in the mean arterial blood pressure (Sun & Guyenet, 1986). When the sympathetic response is stimulated we blush, become excited, alert, and ready to fight or flee. The pupils dilate, the small air bubbles in the lungs inflate and the digestive system slows almost to a standstill since blood from it is re-routed to the brain and muscles to facilitate a reaction to danger. The sympathetic part increases blood pressure and pulse to deal with sudden changes. Therefore, changes in the mean arterial pressure can be used as a non-invasive index of the brainstem sympathetic activity and it is possible to monitor the mean arterial pressure beat-by-beat continuously using non-invasive methods. Moreover, it is also possible to monitor cardiac vagal tone (CVT) continuously in real-time by non-invasive methods (Julu et al., 2001). CVT represents the brainstem parasympathetic activity. The parasympathetic part functions as a natural brake that prevents the sympathetic part from getting out of control. When the parasympathetic part is in action it slows down the pulse, the pump rate of the heart diminishes, blood pressure reduces, pupil size lessens, and salivation and intestinal movement increases. The parasympathetic nervous system creates rest and calm in situations when the body’s reserves are being replenished. Contrary to the sympathetic part that is always active, the parasympathetic is active only when it is called upon to be so. Compare the action of applying the brake of a car, or the pulling of the reins to slow or stop a horse.
As we can monitor cardiac vagal tone, this means that the present advance in medical technology can allow us to monitor both sympathetic and parasympathetic activities in the brainstem using non invasive methods simply by monitoring the changes in the mean arterial pressure and the levels of cardiac vagal tone and these can be monitored simultaneously and continuously in real-time. I have used this new advance in medical technology to monitor brainstem autonomic activity as an index of emotional output in response to music and vibroacoustic therapy in Rett participants and a comparison group of normally developed children.
Design
This experimental study was designed as a within subject study (sometimes referred to as a repeated measures design), where the same subject is used for both a control and experimental condition. The subjects in the study were randomized to the extent that the order of the stimuli (conditions) they received was randomly assigned. Subjects inevitably differ from one to another. It is simply a fact of life that subjects differ greatly. In between-subject designs these differences among subjects are uncontrolled and are treated as error. In within-subject designs, the same subjects are tested in each condition. Therefore, differences among subjects can be measured and separated from error.
All repeated measures’ experiments are factorial. Time is called the within-subject factor because different measurements on the same subject occur at different times. Treatment is often called the between-subject factors where levels change only between treatments; measurements on the same subject represent the same treatment. This study does not have parallel clinical groups and therefore there are no between-subject factors, only within-subject. The statistical analysis of single-factor designs is usually a t test or ANOVA, which can conceptually be thought of as a ratio of the variability between groups to the variability within groups. A repeated measure design increases the sensitivity of the test as subjects serve as their own controls, and thus across-subject variation is not a problem.
To calculate the effect size for this study I have used Power and Precision, ANOVA.
Alpha is 0.05 and within an assumed R² = 0.36 (r =0,6). I aimed to find out the required sample size for 80% power. The result showed that 23 subjects would be needed. This calculation was based on parallel design (23n/group) where as in this study there was a within subject design. The actual power could be higher or lower, dependent on the number of subjects recruited.
The study data was collected during the (approximately) one hour the person was exposed to a registration of basic functions – the brainstems control function of the autonomic nervous system. The control situation was the person’s own baseline which was measured when the person was resting, breathing normally, awake and no epileptic activity was seen on the EEG. We looked at a homogeneous group as far as the neuronal maturity was concerned (Armstrong & Kinney, 2001).
In this study there was allocation concealment. One member of the team, not involved in the study, used a random generator (Excel) and gave the order of the stimuli for each person before every test situation. We have described our stimuli using terms such as: activating, calming or favourite music based on the authors’, parents’ and carers’ descriptions of the subjects’ behaviours towards their favourite tunes. We had a total of six stimuli applied consecutively to each subject in this study.
Brainstem Examination
A brainstem examination entails neurophysiological examination of the dysfunction of the respiratory organs and of the function of the brainstems autonomic nervous systems. This is a painless and non-invasive method. A computerized integrated monitoring machine called the NeuroScope™. This is a machine that quantifies the extent of the communication that exists between the brain and the heart. Neuro ScopeTM from MediFit Instruments Ltd, London, UK is used to record the beat-to-beat heart rate from ECG R-R intervals.
Systolic and diastolic blood pressures are also recorded beat-to-beat by using the Portapres model 2. A photoplethysmographic finger-cuff is used to obtain a continuous digital arterial blood pressure in to the Portapres. The Portapres send the signals to a computer which calculate the recorded signals into beat-to-beat blood pressure; systolic, diastolic and mean blood pressure.
As earlier mentioned the sympathetic activity is very closely related to mean arterial blood pressure (Sun & Guyenet, 1986) and that activity can be monitored from the calculation of mean arterial blood pressure. With the help of this technique it was possible to deduce the sympathetic stimulation to the heart from the heart rate in conjunction with cardiac vagal tone (CVT), which was measured independently. During the measurement period, pre-jelled electrodes on the girl’s chest registered ECG. Cardiac vagal tone was monitored continuously in real-time from ECG signals using a modification of principles described by Julu (Julu, 1992) and now implemented in the Neuro Scope TM. CVT represents the brainstem parasympathetic activity. CSB, cardiac sensitivity to baroreflex, also represent the brainstem parasympathetic activity. It is defined as the increase in pulse intervals per unit change in systolic blood pressure.
The computer calculates everything in real time. The method is described very well by Dr.Julu (Julu, 2001; Julu et al., 2001). Breathing was registered with a pieuzo-electric plethysmographic belt round the lower part of the rib cage. All breathing movements are also integrated in the computer together with pressure of oxygen and carbon dioxide levels registered transcutaneously by using electrodes fastened immediately above the liver and connected to a Radiometer transcutaneous monitor (TINA TCM3®).The equipment for breathing movements, arterial blood pressure, the transcutaneous oxygen and carbon dioxide were all connected to the Medulla, which was connected to the computer. The computer calculated all data and everything was possible to study beat-by-beat. The Electroencephalogram (EEG) was also measured using an electro cap, ECI, with built-in electrodes.
All data were registered in a computer and could be seen as graphs on the screen. This gave one the opportunity to discuss the developments minute by minute during the registration. The whole process was time synchronised with two digital video cameras and filmed, which enabled one to later examine the results in detail. The EEG camera was focused on the subjects face while another camera was used to record all other behaviour and movements during the registration.
During registration the girl sat in a sack named “Music Molly” from the Swedish firm “Kom i Kapp” with built in loudspeakers. During VT as well as VT and music, the low-frequency tones, 40 Hz with a sinus curve with 5 seconds’ duration and a peak of 89.4dB at the point where the person sat, come from the loudspeakers in the sack and were experienced as vibrations. An external CD player were connected to the “Music Molly” and delivered the vibrations. An additional CD player, was connected to loudspeakers to provide the music played during VT and music as well as the patents or carers selected music.
Procedure
As described earlier the subjects were exposed to six stimuli. To secure the situation from carry-over effects there was allocation concealment. One member of the team, not involved in the study, used Excel random generation and gave out the order of musical stimuli before every examination. Even though the subjects were back to baseline or as close to baseline as possible, there could be a suspicion of carry-over effects and the randomisation eliminated that risk.
When there was time to do the assessment, the subject was put into the sack (Musik-Molly) with built in loudspeakers. All the electrodes for ECG and oxygen and carbon dioxide, breathing belt, finger cuff and EEG cap were connected to the person and to the technical equipment. The video cameras, the microcomputer and the EEG machine were time synchronized and the monitoring could start. First, we had to establish a baseline.
The definition of baseline brainstem autonomic function as the activity is recorded during a period when the subject is breathing normally, blood gases within the normal range, awake, no signs or evidence of agitation and no epileptic activity. This definition was formulated by Julu (2001). One minute is sufficient time to record the baseline, but longer periods of up to five minutes can be recorded.
After established baseline the subject was exposed for the first stimuli and it could be any of the six; horn (a, for the participant, unknown piece of goat-horn-music and chosen by the therapist), calming music (chosen by the parents/carers), activating music (chosen by the parents/carers), VT (40Hz with a sinus curve with 5 seconds’ duration and a peak of 89.4dB), VT combined with calm and structured music chosen by the music therapist or that music without VT depending on the randomisation.
The person was exposed to the music parts in two minutes chosen excerpts. The response was measured during the last minute of the music in those two minutes. That was because the initial bit of the response includes the readjustment to the new stimulus. A piece of music of activating or calming character is generally approximately two minutes and should therefore not be longer than that in the measurement situation.
The person was exposed to the other three conditions (VT, VT and music, and the music used in VT alone) for 10 minutes. Again, the response was measured throughout, but analysis was undertaken only on the last minute of the test.
The reason for the time difference between VT and the music parts was because of an attempt to resemble as normal a situation of treatment as possible. A VT session is going on for approximately 25 minutes but it was impossible to expose the participants to 3 x 25 minutes during the monitoring. Therefore, it was necessary to make a choice. The last minute was chosen because 10 minutes is the minimum of time for using VT and to optimise as much as possible the last minute was used.
In between every stimulus the subject settled down and returned to baseline or as close to baseline as possible.
Analysis Methods
The time for the starting and stopping of different stimuli were marked in the computer. This was because of the possibility of afterwards being able to analyse the course of events, find peaks and mean for CSB, CVT and MAP. The exact time was also important for the later video analysis of facial expression. All the different values were put into a database were one also could find the baseline for each person.
Together with the analysis of CVT, CSB and MAP there will be a video analysis of face expressions and vocal sounds from the last minute of each stimulus. The video analysis is to correlate external expressions, and investigate the possibility of interpreting face expressions, with the results from the monitoring.
A second video analysis is going to be done of the total time the subjects were exposed to musical stimuli. This part included facial expressions connected to the music and to the variations in the music. I want to investigate if facial expressions are possible to interpret (emotional responses or brainstem responses) and if changes in facial expressions are caused by or correlated with what is happening in the music. Therefore an analysis of the music will be done with the help of, based on, Hoopers analysis method and his way to characterise music (Hooper, in press).
Reflections so Far
One reason for me to choose a quantitative study from the very beginning was the milieu where I was, and still am, working. Rett Center is a national specialist healthcare unit and I work very close to people from natural science and therefore I need to follow that tradition. To be able to investigate and to explain the phenomenon music for scientist I need to use the same or at least similar vocabulary as they use. It is a matter of understanding and for me it is important that my colleagues can understand and maybe also accept the powerful therapeutic value of music.
The chosen study design is complicated and complex but suitable for this kind of study but as a music therapist I have to draw a clear line and narrow the field of study.
An advantage about using a quantitative method is the clearness of the hypothesis and the necessity to answer the research questions in a very strict way but that could in the same way be limiting or a disadvantage. So many questions and reflections appear during the way but according to the study design I have to leave them out. Another advantage is that because of the quantitative design people from natural science and medicine (my working area) can understand and maybe even accept the study as research and that might open doors for future cross professional research including music and music therapy.
But on the other hand this kind of study might provoke strictly qualitative researchers. I am looking at music as therapy, but without the active music making characteristic which builds a client-therapist relationship involving communicative interaction.
The way I see it the measurable results from this study might give ideas for a new study. My study is a quantitative study but it might lead to a qualitative one – later.
References
Armstrong, D., & Kinney, H. (2001). The neuropathology of the Rett disorder. In A. Kerr & I. Witt Engerstrom (Eds.), Rett disorder and the developing brain (pp. 57-84). New York: Oxford University Press.
Bergström-Isacsson, M., Julu, P. O. O., & Witt Engerstrom, I. (2006). Autonomic responses to Music and Vibroacoustic Therapy in Rett Syndrome. Submitted.
Best, C. H., & Taylor, N. B. (1966). The basal ganglia, the thalamus and hypothalamus. In C. H. Best & N. B. Taylor (Eds.), The Physiological Base of Medical Practice (pp. 144-161). Edinburgh, London: E. and S. Livingstone Limited.
Guyenet, P. G., Koshiya, N., Huangfu, D., Barbaran, S. C., Stornetta, R. L., & Li, Y. W. (1996). Role of medulla oblongata in generation of sympathetic and vagal outflows. In G. Holstege, R. Bandler & C. B. Saper (Eds.), The emotional motor system (pp. 127-144). Amsterdam, New York, Oxford: Elsevier.
Hooper, J. (in press). Record of Predictable Musical Factors (RPMF): Criteria developed to determine the selection of commerically recorded non-contingent music for research.
Julu, P. O. O. (1992). A linear scale for measuring vagal tone in man. J Auton Pharmacol, 12, 109-115.
Julu, P. O. O. (2001). The central autonomic disturbance in Rett syndrome. In A. Kerr & I. Witt Engerström (Eds.), Rett disorder and the developoing brain (pp. 131-182). New York: Oxford University Press.
Julu, P. O. O., Kerr, A., Apartopoulos, F., Al-Rawas, S., Witt Engerström, I., Engerström, L., et al. (2001). Characterisation of the breathing abnormality and associated central autonomic dysfunction in Rett disorder. Archives of Diseases on Childhood, 85, 29-37. Scandinavica 1990;suppl 1396.
Prickett, C. (2005). Principles of Quantitative Research. In B. L. Wheeler (Ed.), Music Therapy Research; second edition (pp. 45-58). Gilsum NH: Barcelona Publishers.
Skille, O. (1991). Music Therapy: Manual and Reports. Levanger: ISVA Publications.
Sun, M. K., & Guyenet, P. G. (1986). Medullospinal sympathoexcitatory neurons in normotensive and spontaneously hypertensive rats. Am J Physiol, 250(5 Pt 2), R910-917.
Wigram, T., & Dileo, C. (1997). Music vibration and health. Cherry Hill, NJ: Jeffrey Books.