Symposia
Suicide and Self-Injury
Elizabeth A. Velkoff, Ph.D.
University of California San Diego
Portland, Oregon
April Smith, Ph.D.
Assistant Professor
Auburn University
Auburn, Alabama
The purpose of this study was to test a putative relationship between decreasing interoceptive attention (IA) and self-injurious behavior (SIB; defined here as disordered eating and non-suicidal self-injury). Previous research finds that greater interoception is associated with lower risk for SIB. However, tests of the temporal relationship between IA and SIB, especially in the hours or minutes before an episode, are lacking. We predicted that people with lower average IA, and times when people had IA lower than their own average, would be associated with greater risk for SIB. Additionally, we predicted that IA would decrease in the hours and minutes leading up to SIB, and that this effect would be stronger before episodes of SIB compared to urges for SIB. Finally, we tested an alternative hypothesis that greater change in IA, regardless of the direction, would be associated with the highest risk for SIB. We used ecological momentary assessment to collect data from adults with recent SIB (N = 128; 70.2% reporting at least one urge or episode of disordered eating during the study period) on IA and SIB for 14 days. Participants reported on urges for and episodes of SIB, and rated IA on a 0-100 visual analog scale for the question “How connected do you feel to your body right now?” We used multilevel models with occasions at level 1 nested within days at level 2, and days nested within persons at level 3. Using binary logistic mixed modeling, we tested our first hypothesis and found that, as predicted, people with higher average IA had lower risk for SIB, but that conversely, risk for SIB was higher at moments when people were above their own average IA. Furthermore, we found that there was a statistically significant but small linear effect such that IA increased prior to urges for or episodes of SIB, but that there were not differences in this effect between urges and episodes. Finally, using a dynamical systems approach, we found that greater changes in IA were strongly associated with risk for urges for or episodes of SIB, regardless of the direction of the change. Findings indicate that people with better average IA are at less risk for SIB, but that higher momentary IA is not always associated with lower risk. We also found that IA seems to have a baseline and fluctuate around that baseline, while larger fluctuations are associated with the highest risk for SIB. Results are consistent with active inference models of interoception and clarify the short-term relationship between IA and SIB. Our findings suggest that IA may be a fruitful target for interventions aimed at reducing SIB.