Back

Test-retest reliability of a 2-dimensional and 3-dimensional visual assessment of body image disturbance in anorexia nervosa

Ralph-Nearman, C.; Arevian, A. C.; Karem, A.; Llano-Rios, T. F.; Sinik, M.; Moseman, S.; Feusner, J. D.; Khalsa, S. S.

2023-06-20 psychiatry and clinical psychology
10.1101/2023.06.14.23291397 medRxiv
Show abstract

Body image disturbance (BID) is a diagnostic feature of anorexia nervosa (AN), with few reliable visual perceptual or attitudinal markers. Somatomap is a 2-dimensional (2D) and 3-dimensional (3D) digital assessment of BID, which has demonstrated utility. Test-retest reliability of Somatomap 2D and 3D digital assessment of BID in AN was examined. Fifty-nine inpatient participants with AN performed test-retest by a) outlining body concern areas on a 2D avatar for each independent area of concern; and b) sculpting 23 independent body parts on a randomized 3D avatar to reflect their perceived body size in length and girth. Participants corresponding body parts were physically measured to calculate discrepancy scores (i.e., 3D perceived minus measured values). Regional 2D BID test-retest differences were evaluated using z-scores to generate statistical visual body maps. Test-retest of 3D assessment reliability was evaluated by Intraclass Correlation Coefficient for individual and aggregated body parts. Somatomap 2D demonstrated excellent test-retest reliability with no statistical differences between test and retest z-scores. All 23 body parts on Somatomap 3D demonstrated statistically significant fair-to-excellent test-retest reliability in AN. Regions that are commonly of concern in AN, and combined measures, showed the highest reliability. Results suggest that Somatomap 2D and 3D may provide a reliable perceptual marker of visual BID in AN.

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.