BME100 f2017:Group15 W0800 L3
LAB 3 WRITE-UP
Heart Rate (Ox vs Spree)
Pearson's r Coefficient: 0.4772
Temperature (Oral Thermometer vs Spree)
Pearson's r Coefficient: 0.0372
The statistical analysis test that was used for both sets of data (heart rate and temperature) was a two-tailed t test. This test is the most effective test at determining if there is a difference between two sets of data which was the purpose of the experiment. The p-value for the heart rate t test was 0.4271 and the p-value for the temperature t test was 1.09676X10^-21
Summary of Results
There is not statistically significant evidence that the Spree device works any better than the Pulse Ox (they work about the same) because the paired t-test results in a p-value of 0.4271, which is much greater than the α-value (significance level) of 0.05. This is reinforced by the fact that the correlation coefficient is 0.4772, signifying that there is not a strong correlation. In conclusion, for heart rate, the Spree device is not any better than using the Pulse Ox, they give relatively similar results.
Temperature- There is statistically significant evidence that Spree does not work better than the Oral Thermometer because the paired t-test results in a p-value of 1.0967x10-21, which is much lower that the α-value (significance level) of .05. This is confirmed by the correlation coefficient being 0.0372, signifying that there is not a correlation. In conclusion, for temperature, the Spree device does not work better than the Oral Thermometer, in fact, the Spree values were very off.
Experimental Design of High Frequency Ultrasound Probe with Detachable Head
Much like the Spree v. the Gold Standard experiment, we would need to test our device, removable heads for high frequency ultrasounds, against the non removable heads that are already being used. We would set up a paired t-test to make sure the frequency of our prototype matches the frequency given off by the gold standard. The parameters for this test would be testing our prototype, at each frequency between 50 to 200 MHz, against the gold standard 50 times (50 paired machines), repeating this exact set experiment 10 times to make sure the data is reliable and valid. This experiment should also have our prototype tested against multiple ‘gold standards’ because of the devices out there, there are multiple types of high frequency ultrasound machines. We would want to see a correlation coefficient of close to one with a p-value above the significance level, which is the α-value. We would want to see that our prototype works just as well as the gold standard, so that we have evidence that our prototype works just as well but is cheaper and more convenient for use, since having a removable head will make it easier and cheaper to use high frequency ultrasound devices rather than having a non-removable head that can only be used once (or be expensively cleaned). A second experimental design would be testing to make sure the removable head hits the exact same target area as the gold standard.
The final experiment would have to be testing the device for deep brain stimulation. Rather than having a paired t test experiment, there would have to be two separate experimental groups to test for a placebo effect and therefore it would be an independent t test. Four groups of about 50 individuals with Parkinson’s Disease each will be used. Before the tests, a model for Parkinson’s needs to be developed. To accomplish this a test of hand control will be used on each individual. Each person will have a metal hoop around a wire that has curves in it. When both the hoop and wire touch an electrical current that can be detected (and not harmful) can be recorded. Each of the three groups’ scores will be averaged. One group will have deep brain stimulation, one group will have ultrasound stimulation, one group will have no stimulation, and one group will have a turned off ultrasound “give” them stimulation. The last group is to make sure there is no placebo effect. The test mentioned earlier will be performed by the individuals on a regular basis (probably every week) after stimulation. After so many weeks, the results will be graphed. In order for our probe to be “successful” an ANOVA test will determine whether the hypothesis is right or not.