Evaluating an internet-delivered fear conditioning and extinction protocol using response times and affective ratings

Abstract

Pavlovian fear conditioning is widely used to study mechanisms of fear learning, but high-throughput studies are hampered by the labor-intensive nature of examining participants in the lab. To circumvent this bottle-neck, fear conditioning tasks have been developed for remote delivery. Previous studies have examined remotely delivered fear conditioning protocols using expectancy and affective ratings. Here we replicate and extend these findings using an internet-delivered version of the Screaming Lady paradigm, evaluating the effects on negative affective ratings and response time to an auditory probe during stimulus presentation. In a sample of 80 adults, we observed clear evidence of both fear acquisition and extinction using affective ratings. Response times were faster when probed early, but not later, during presentation of stimuli paired with an aversive scream. The response time findings are at odds with previous lab-based studies showing slower as opposed to faster responses to threat-predicting cues. The findings underscore the feasibility of employing remotely delivered fear conditioning paradigms with affective ratings as outcome. Findings further highlight the need for research examining optimal parameters for concurrent response time measures or alternate non-verbal indicators of conditioned responses in Pavlovian conditioning protocols.

Figure 1

Experimental design. The top panel shows the outline of each experimental trial during the acquisition phase. A brief fixation cross was shown on the screen followed by an image of a neutral female face. During CS+ trials the presentation was followed by an aversive sound (scream) accompanied by the same face with a fearful expression. During CS− trials only the neutral expression was shown and no sound was played. During CS presentation, response times were probed after either 500 ms or 2500 ms by playing a brief 440 Hz tone. Participants were instructed to press the space-bar as quickly as possible when the probe was played. Trials during extinction and reinstatement had the same structure but the US (sound and fearful expression) was always omitted. The face-images shown in the figure are sample images from the FACES database, but not the same as the ones used in the experiment. The bottom panel illustrates the outline of the entire experiment, including number of trials of each type for each experimental phase.

Figure 2

Affective ratings to conditioned stimuli (CS). Negative affective ratings were similar for stimuli paired with aversive outcome (CS +) and stimuli never paired with aversive outcome (CS−) pre acquisition, but differed after acquisition (acq), with increased negative ratings of CS+ and reduced negative ratings of CS−. Post extinction (ext), ratings of CS+ and CS− still differed, but to a lesser degree, whereas reinstatement (reinst) increased negative ratings of both CS. Affective ratings is a composite score consisting of the mean rating (1[not at all] to 7[totally]) of how much participants agree with feeling fear, discomfort, irritation and calm (reversed score). Points and error-bars indicate means and SEMs.

Figure 3

Response times to early and late auditory probes during presentation of conditioned stimuli (CS) in the acquisition phase. Response time to the neutral stimulus paired with an aversive outcome (CS +) was faster than to the stimulus never paired with an aversive outcome (CS−) during early (500 ms after onset of CS) but not late (2500 ms) probes. Points and error-bars indicate means and SEMs.

Figure 4

Response times to early auditory probes during presentation concomitant with conditioned stimuli (CS) in the acquisition (acq) and extinction (ext) phases. Response time to stimuli paired with an aversive outcome (CS +) was faster than to stimuli never paired with an aversive outcome (CS−) during acquisition but not extinction. Response times for both stimuli was higher during acquisition than during extinction Points and error-bars indicate means and SEMs.

Figure 5

Associations between trait anxiety and learning variables. Scatter plots and linear fit lines showing correlations between trait anxiety, measured with Spielberger State-Trait Anxiety Inventory-trait version (STAI-T) and (left) negative affective ratings to conditioned stimulus never paired with an aversive outcome (CS−), (middle) response time to early probes presented during CS−, and (right) difference in response time between probes presented during stimuli paired with and aversive outcome (CS +) and CS−. A negative difference score indicates that CS+ responses were faster than CS− responses, and thus reflects stronger learning.