Appendix A (Online supplement)
Psychophysiological Data Collection
Heart rate was collected with a three lead electrocardiogram (ECG) using a Biopac data acquisition unit (MP36: Biopac Systems Inc., US). The disposable Ag/AgCl electrodes filled with electroconductive gel were placed on the participant’s right clavicle, the lower left rib cage, and on the palmar surface of the left foot. The electrode on the foot was part of the EDA lead system, but was shared by both systems as the ground lead (more than one ground lead can lead to signal interference across systems). Trained research assistants placed the electrodes on participants and then attached the leads. The data were sampled at rate of 500Hz, and were collected and analyzed using AcqKnowledge 4.1 software (Biopac Systems Inc., US).
Electrodermal activation (EDA) was obtained using a galvanic skin response amplifier, which was connected to the BiopacMP36 data acquisition unit. Participants were required to remove both shoes and socks before two disposable Ag/AgCl electrodes filled with electroconductive gel were attached to the plantar surface of their left foot. The electrodes were placed approximately one inch apart on the midfoot. The participants were then required to place both of their legs on a leg rest so that the electrodes on their feet would not touch the ground. The foot was chosen as the acquisition site instead of the traditional phalange placement because this study required participants to type, write, and engage in other tasks with their hands, which would degrade the transmission of the signal. Like the ECG acquisition, the EDA data were received and analyzed using AcqKnowledge 4.1 software.
Psychophysiological Data Scoring and Reduction
The raw ECG data underwent a bandpass filter of 0.5Hz to 35Hz to eliminate noise from the signal. Similarly, the raw EDA data were smoothed with a smoothing factor of 500. Any data that were more than approximately 40% noise or movement artifact were removed from the analysis. This resulted in the removal of 4 ECG and 3 EDA data points.
The data for each time of measurement (three training periods and three stressor tasks) was cut down into epochs of equal time length. For the training periods, the data were parsed into 120 second epochs, with the maximum number of epochs for the first training period being four, followed by three epochs for each of the following two training periods. This resulted in a total of 10 possible epochs for both ECG and EDA over the course of training. If participants took more time to complete the training than the number of epochs being collected for that training period (e.g., more than 360 seconds for the second training period), then the remaining data were disregarded. Similarly, if participants completed the training period before the last epoch was complete within a given time period (e.g., less than 360 seconds first for the second training period), data for the incomplete epoch were not recorded. Only data from completed epochs were used. For the stressor tasks, deviance statistics were calculated across participants and used to create an epoch time that included the most people. The epoch length for the anagram task was 120 seconds, while the length for the thought task was 90 seconds, and the trash task was 60 seconds. Raw ECG data were converted into heart rate (HR), and then into mean heart beats per minute (BPM), and EDA were calculated within each epoch and used as the indices of physiological arousal. See Cacioppo, Tassinary, and Bernston (2007) for further information about guidelines for psychophysiological data acquisition.