We recruited n=22 medical students per an approved IRB protocol to participate in a longitudinal study regarding the relationship of sympathetic arousal and skill in learning micro-surgical tasks. The subjects had to pay five visits, lasting one hour each, in order to practice micro-surgical cutting and suturing in an inanimate simulator. In their first visit, and after signing an informed consent, the subjects completed a biographic questionnaire, and a trait anxiety inventory. At the end of their last visit they completed a post-study questionnaire.
During the main part of each session, the subjects underwent the following treatments:
Baseline: The subjects were relaxing for 5 min, listening to spa music. They were facially recorded by a thermal and visual camera. Cutting: The subjects had to perform precision cutting in the inanimate simulator. They were facially recorded by a thermal and visual camera. Sutiring: The subjects had to perform suturing in the inanimate simulator. They were facially recorded by a thermal and visual camera. After the cutting treatment the subjects had to fill out a NASA-TLX questionnaire. The subjects also filled out a NASA-TLX questionnaire after the suturing session. The NASA-TLX instrument features five subscales measuring different aspects of the subjects' perceptions regarding task difficulty.
Data Organization
The data is organized per subject in a way that largely reflects the experimental design. At the root level, there is a master index file that contains information regarding the state of the dataset. In this file, "1" stands for a data item that is present, while "0" stands for a data item that is missing, due to technical or other reasons. At the root level, there is also a file reporting the time and accuracy performance of the subjects, as recorded by the evaluating surgeon.
Within each subject folder there are files containing biographic, trait psychometric, and post-study views of the subject. There are also the subfolders for each of the visits the subject paid. Each of these subfolders holds the measurements that were taken in each task. These are stress measurements expressed as perinasal signals for the duration of the task, and NASA-TLX scores taken at the end of the respective treatment.
Quality Control
Biographic Data: Draw the barplot for gender distribution; histogram for age distribution Trait Psychometric Data: Draw the histogram for TAI scores. Please consider only the total TAI score, which can be found at the *tp.csv file residing at the root of each subject file system. TAI takes values in the range 20-80, with scores up to low 40s considered normal, while higher scores considered indicative of overanxious individuals. State Psychometric Data: For each subject draw the barplots for all the NASA-TLX subscales per task. This will give two figures per subject per subscale, one for suturing and one for cutting, where the evolution of the scores from the initial to the final session will be evident. There should be a downward trend, reflecting increased facility with the tasks. Keep in mind that NASA-TLX subscales are scored in the range 1-20. Perinasal Perspiration (Stress) Signal Data: For each session of each subject draw the stress signals, using black for baseline, green for cutting, and red for suturing. Generally speaking, the baseline signal should be at a lower level than the other two. In total, you will draw five figures for each subject or whatever the number of his/her sessions is. It is worth noting that the sampling rate fluctuates around 7 frames per second. We suggest that you downsample to 1 frame per second and use these downsampled signals to perform statistics. This averaging will give you a smoother signal without affecting the validity of the data, as arousal responses manifest in 2-4 seconds, that is, at a resolution higher than the averaging that we suggest. Performance Data: Draw the time barplots for each subject, using different colors for each task (cutting vs. suturing). You should observe a downward trend for the cutting task. Draw also the accuracy barplots for each subject, using different colors for each task (cutting vs. suturing). You should observe an ascending trend. Please note that there are two accuracy variables: one provides an overall accuracy score, while the other indicates the number of sutures properly completed in the allotted time. The allotted time was 20 minutes for the suturing task. IMPORTANT NOTE wrt Performance Data We have uploaded new performance data, with excplicit accuracy scores per task. Hence, the cutting task has its own accuracy scores and so is the case with the sutruting task. Please use these accuracy scores instead of the legacy generic accuracy score. Obviously, these more specific scores will enable more specific and sound inferences.
HYPOTHESES - TESTING - INTERPRETATION You need to formulate a set of pre-planned hypotheses that you would like to test. Such a set should be driven by the experimental design. This is a micro-surgical education study, with two distinct tasks that present different challenge levels. Some key questions are if the levels of stress and the type of task played any role in the accuracy scores, and if the accuracy scores changed over time. You can formulate a linear model to test this set of hypotheses. The results of these tests can provide revealing insight. For example, if you find that there was no accuracy improvement over time in the suturing task, this means that the educational process was a failure in this respect.