Supplemental Materials

Interoception in Pathological Health Anxiety

by S. Krautwurst et al., 2016, Journal of Abnormal Psychology

Supplement 1

Description of interoception tasks

Signal-detection task for non-specific skin conductance fluctuations (NSCF task)

For measuring skin conductance level (SCL), two electrodes with a contact surface area of 2 cm2 filled with isotonic paste (Fowles et al., 1981) were placed on the palm of the non-dominant hand. With the help of the Variotest System (Gerhard Mutz, Köln, Germany), skin conductance was continuously analyzed online for the duration of the experiment. Initially, participants were shown their skin conductance levels on a screen and were introduced to the concept of minor changes of physiological arousal, in terms of skin conductance fluctuations. Participants were then instructed to focus internally and to try to detect even minor signs of physiological arousal, and the screen was turned off. Then 20 acoustic signals were presented that were either triggered by NSCFs (arousal signals, 10 times) or by a stable phase of skin conductance for at least 20 seconds (non-arousal signals, 10 times). An arousal signal was triggered when a slope value of 0.07 µ-Siemens per second was detected and a minimal amplitude value of 0.25 µ-Siemens was reached within 1 second (Andor et al., 2008). Participants had to decide whether they felt physiological arousal just prior to hearing the tone. The order of searching for the two different types of triggers was determined randomly. The minimum time interval between the two events was set to 35 s. If the target event did not occur within 150 s, the trial was stopped, but an acoustic signal was nonetheless presented (which then counted as a non-arousal signal). Then, an arousal signal was searched for again. Participants could only be included in data analyses if they had exhibited a minimum of five arousal signals. Since the number of NSCFs decreases with lower levels of arousal, two breaks were implemented, during which participants were asked to talk about a book/movie/vacation. At the end of the task, rest values as a proxy for a true baseline were recorded.

Performance in the NSCF task was parameterized by computing the sensitivity index d′, and the response bias index C. D′ and C are parameters derived from signal detection theory (MacMillan, 2002). D′ is an indicator of the ability of the participants to distinguish between the occurrence of a physiological arousal signal (i.e., an NSCF) and a stable phase without an NSCF. The higher the d′ score, the better participants are able to perceive NSCFs as indicators of phasic bodily arousal. C is the response criterion. C=0 represents a neutral response criterion – participants said yes as often as no. C0 signals a liberal response criterion – participants more often said yes. C0 is a conservative response criterion – participants more often said no.

Heartbeat mental tracking task

Three Ag-AgCl electrodes were placed at the lower left rib cage, at the right mid-clavicle, and at the left mid-clavicle (serving as a ground electrode). The electrocardiogram was sampled with a rate of 512 Hz and analyzed online using the Variotest System (Gerhard Mutz, Köln, Germany). First, participants were asked to relax for 3 minutes to assess their heart rate at rest (baseline measure). Then, participants were instructed to silently count all the heartbeats they could feel within three randomly presented fixed time intervals of 25, 35, and 45 seconds. Participants were explicitly asked not to take their pulse or note the time during the heartbeat perception task. The onset and offset of these time periods were signaled by two acoustic tones. Between intervals there was always a pause of 30 s. Interval and rest durations were all unknown to the participants. At the end of each interval, participants were requested to report all counted heartbeats.

Performance in the heartbeat perception task was quantified by computing a heartbeat perception score (HBP) as follows (cf. Knoll & Hodapp, 1992).

HBP = 1 − 1/3 ∑(│recorded heartbeats – counted heartbeats│/ recorded heartbeats)

The heartbeat perception score is a measure of how well an individual is able to perceive his/her own heartbeat. The higher the score, the better the individual is able to track his or her heartbeat. The possible range is 0 to 1, with 1 indicating a perfect HBP (Schandry, 1981).

Supplement 2

Table 1

Correlations between possible confounding variables and the heartbeat perception score (HBP) within the total sample as well as the healthy comparison group and the pathological health anxiety group separately

Heartbeat perception score
Overall / HC / PHA
Baseline heart rate / -.11 / -.13 / -.11
BMI / -.11 / -.13 / -.08
Sports activity / -.06 / .15 / -.14
Blood pressure / .10 / .13 / .03
Age / -.03 / -.04 / -.01

Note: BMI=Body Mass Index, healthy comparison group (HC; n=56), PHA = patients with pathological health anxiety (n=49).