Participants
Thirty-two university students (17 males and 15 females aged 20–28 years, M = 21.7 years) participated in this experiment after providing written consent forms. The participants received an honorarium of 1,120 yen after completion of all the experimental sessions. All participants reported normal or corrected normal vision. The requisite sample size of 32 participants was determined using the G*Power program (Faull et al., 2007) for a mixed-design repeated-measures ANOVA (two within-participant factors consisting of four tests and one between-participant factor consisting of two groups) based on α = 0.05, power (1–β) = 0.80, and an effect size of f = 0.25. Note that the estimated sample size is greater than the calculated value of 24, accounting for some exclusions in the data.
The present experiment was a between-participants design, with two response modes: hitting or avoiding a target probe. Each participant performed either of the two modes in the visual-probe task. The participants completed the trait scale of the Spielberger State-Trait Anxiety Inventory Japanese version (STAI-JYZ) prior to beginning the experiment. There was no significant difference in the trait anxiety score between the two groups (t(30) = 0.69, p = 0.485, Cohen’s d = 0.16), with M = 50.4 and SD = 8.6 for the hit group and M = 48.2 and SD = 9.8 for the avoidant group. The group means did not significantly differ from those for the Japanese population, with µ = 48.2, σ = 10.0 (Hinoda et al., 2000) for the hit group (Z = 0.89, p = 0.375) and the avoidant group (Z = -0.01, p = 0.990).
Apparatus and materials
We followed the threat-of-shock procedure delineated in previous studies (e.g. Bublatzky et al., 2017; Clark et al., 2012; Schmitz & Grillon, 2012). A portable constant-current stimulator (USE-100, Unique Medical, Japan) was set to deliver bipolar electric pulses with a duration of 5.0 ms and a frequency of 5.0 Hz. The stimulation intensity was varied between 4 mA and 10 mA. A watch-type heart rate (HR) monitor (Polar Vantage M, Polar, Finland) and three visual analogue scales (VAS) with a line segment of 100 mm to indicate their levels of anxiety, fear, and happiness were used to assess the degree of threat manipulation.
The participants executed the visual-probe task using three buttons on a handmade response box (233 mm long, 187 mm wide, and 50 mm high). The three buttons with a diameter of 20 mm were arranged in a horizontal row with a gap of 25 mm. During the task trials, the participants pressed each button with the index finger while resting their palm on this device.
A fixation point (black crosslines with a line length of 110 pixels and a line width of 5 pixels), face images, and a probe image were presented in a light grey target frame (1840 × 780 pixels) with a black background. The face images were collected from the FACES database developed by the Max Planck Institute for Human Development, Centre for Lifespan Psychology (Ebner et al., 2010). Three face images representing Angry, Happy, and Neutral emotions were selected from two male models (ID: 008 and 114) and two female models (ID: 048 and 140). In total, 12 images were prepared. Irrelevant backgrounds surrounding the faces in the photographs were removed, and the image sizes were adjusted so that the faces fit within a transparent frame of 390 × 520 pixels. To minimise the effect of colour on the participants’ responses, all pictures were transformed into grayscale images. A black shadow illustration of a human-like figure was used as a probe stimulus to indicate the start signal and the direction of the response actions. The image size was adjusted to fit within a transparent frame of 520 × 520 pixels. All the visual stimuli were presented on a 21-inch LCD with a 1920 × 1080 pixel resolution and a refresh rate of 60 Hz. During the experiment, the participants faced the display at a distance of 1.2 m. The visual angle of each face image was then approximately 5.8° × 7.5° with a 13.5° disparity between the centres of the two face images.
The presentation of the visual stimuli, collection of participants’ responses, and trigger of electric shocks were controlled by a personal computer running an application of the visual-probe task that was originally developed using commercial software (LabVIEW ver. 2019 for Windows, National Instruments, USA). The participants’ responses to manipulating the buttons were acquired at a sampling rate of 1 kHz.
Experimental design and procedure
Half of the participants performed the hit task, and the other half performed the avoidant task. Each action task was block-designed with threat-of-shock conditions (i.e. safe and threat conditions), each of which consisted of three blocks. There were no shocks in the three safe blocks. Electric shocks were delivered after the 10th trial in the 1st threat block and the 30th trial in the 2nd threat block. No shocks were given in the 3rd threat block. Since two trials after the electric shocks needed to be excluded from the data, two additional trials were added at the end of those blocks. Each of the 6 blocks contained 40 trials, resulting in 240 trials in total. The safe and threat blocks alternated, and the order of the sequence was counterbalanced across the participants within each task group.
Figure 1 illustrates the sequence of the stimulus presentations and motor responses in each trial. The illustration exemplifies a trial in which the locus of a threat cue (an angry face) is spatially congruent with the locus of the probe. Each trial began with a PC beep sound that requested each participant to keep pressing the response box’s centre button (i.e. the home button). Then, the fixation point was displayed at the centre of the target frame. It remained visible until each participant pressed either the right or left button (i.e. response buttons). After the presentation of the fixation point, a standby interval of 500 ms followed. Next, a pair of emotional cues (a pair of happy and neutral faces or a pair of angry and neutral faces from the same model) were simultaneously presented on the left and right sides of the fixation point and remained visible for 200 ms. Next, the probe stimulus appeared on either the left or right side of the fixation point and remained visible for 150 ms. The interstimulus interval was fixed at 350 ms. The hit task was to press the response button that spatially corresponded to the side on which the probe was presented. The avoidant task was to press the button on the opposite side.
Before beginning the hit or avoidant task, each participant decided on the preferred hand to manipulate the response device. The HR monitor was attached to the contralateral wrist. Two disposable electrodes were placed on the surface of the dorsal skin above the third proximal and third intermediate phalanx bones of the contralateral hand. The shock intensity was calibrated for each participant to a level that was ‘quite unpleasant/uncomfortable, but not painful’, following the established protocol (Schmitz & Grillon, 2012). The participants were informed that electric shocks might arrive during some of the trials within the threat blocks but never within the safe blocks.
While executing the task, the participants were restricted to using their index fingers alone to press the buttons. They were asked not to release the finger from the home button earlier than the appearance of the probe and to press the correct response button as quickly and accurately as possible when the probe stimulus was presented. The participants performed 40 practice trials without electric shocks, and then the test trials began. At the beginning of each block, a warning appeared on the display: ‘You are now safe from shock this next set of trials’ in black letters on a blue background for the safe condition blocks and ‘You may receive an electric shock at some points during the next set of trials’ in black letters on a red background for the threat condition blocks. A 3 min break was given between blocks. Immediately after completing each block, the participants answered the VAS questionnaires about their moods during the block. The experiment took 60 min. Afterwards, the experimenter thanked each participant for their participation in the experiment.
Measurements and statistical analysis
HR and VAS scores
The raw time-sequence data of the HR records were averaged during each participant’s test block. Next, the average HR and raw VAS scores were averaged among the three test blocks in each of the safe and threat conditions. The participants’ mean measurements were used to statistically test the effect of anxiety manipulation and the difference in induced anxiety levels between the two action goal groups. Anderson–Darling tests revealed that deviations from the normal distribution were significant for some of these variables (p < 0.05). For variables with a skewed distribution, Wilcoxon signed-rank tests and Wilcoxon rank-sum tests were used to evaluate the effect of anxiety manipulation and the difference between the action goals, respectively. We calculated r (Z statistics divided by the square root of the sample size) as a measure of effect size. For variables with the normal distribution, paired t-tests and t-tests were performed to evaluate those effects, respectively. Cohen’s d was calculated as a measure of effect size.
Reaction times
Task performance was evaluated by measuring the release reaction time (RT), press RT, and total RT. The release RT was defined as the time between the presentation of the probe stimulus and the release of the home button. The press RT was defined as the time between releasing the home button and pressing the right or left response button. The total RT was defined as the sum of the release RT and press RT per trial. Errors, such as pressing the wrong response button, were negligible (8 errors, 0.21% of all hit trials, and 16 errors, 0.42% of all avoidant trials). Thus, these error responses were excluded from the full dataset, and the error rates were not statistically analysed. Trials with release RTs < 100 ms or total RTs > 1000 ms were eliminated from the analysis as outliers: 2.6% of all hit trials and 6.7% of all avoidant trials in the full dataset.
Among each participant dataset, the three RT measurements were averaged within the emotional cue condition (i.e. angry faces and happy faces), threat-of-shock condition (i.e. safe and threat), and spatial congruency condition (i.e. the locus of the angry/happy faces was spatially congruent or incongruent with the locus of the probe). Then, the degree of attentional bias was assessed using the following equation:
RT difference [ms] = RT congruent – RT incongruent.
The individuals’ mean RT differences were used for statistical testing. Anderson–Darling tests revealed that the deviations from the normal distribution were not significant (p > 0.05). A three-way, mixed-design ANOVA with the emotional cue condition and threat-of-shock condition as the within-participant factors and the task group (i.e. hit or avoidance) as the between-participant factor was performed for all the RT variables. If the three-way interaction term reached a significant level (p < 0.05), two-way repeated measures ANOVA was performed separately for each of the task groups to estimate the effects of the emotional cue and threat-of-shock condition factors. If the two-way interactions of this ANOVA were significant, multiple comparison tests using paired t-tests with the Bonferroni correction were performed separately for each combination of interests between the factor’s levels. A Greenhouse–Geisser correction was applied when sphericity was violated for the repeated measures. For all ANOVA tests, the generalised eta squared (η2G) was calculated as a measure of effect size (Bakeman, 2005). For all multiple comparison tests, Cohen’s d was calculated as a measure of effect size.