Table 1.
Participant Demographics.
Fig 1.
Example trial of the matrix task.
Note: In our actual trials, we had the following constraints, consistent with Lazarov, Abend, and Bar-Haim (2016): (a) each actor could appear only once on the matrix, (b) there was an even split of genders in each matrix (8 male and 8 female), (c) there was an even split of valences in each matrix (e.g. 8 neutral and 8 sad), and (d) the four inner faces always contained two emotional and two neutral faces. This sample trial uses images approved for publication, so while criteria (b) - (d) are met, criterion (a) is not fulfilled in this example [41].
Fig 2.
Example trial of matrix task with Mouseview.js when stimuli are occluded via Gaussian Blur, and the mouse-aperture is hovering over an image.
Table 2.
Internal Consistency Indices.
Table 3.
Experiment 1: Mixed effects model predicting Dwell Time (ms) from depression severity (BDI-II), stimuli valence, and trial.
Fig 3.
Significant Interaction of depression severity (BDI-II) and stimuli valence on dwell time.
Table 4.
Experiment 2: Mixed effects model predicting Dwell Time (ms) from depression severity (PHQ-8), neutral stimuli valence, and trial.
Fig 4.
Interaction of depression severity and stimuli valence when using PHQ-8 (4A, marginally significant) and IDAS Dysphoria subscale (4B, significant).
Table 5.
Experiment 2: Mixed effects model predicting Dwell Time (ms) from depression severity (IDAS dysphoria subscale), neutral stimuli valence, and trial.
Table 6.
Experiment 3: Mixed effects model predicting Dwell Time (ms) from depression severity (PHQ-8), neutral stimuli valence, and trial.
Fig 5.
No interaction of depression severity and stimuli valence when using PHQ-8 (5A) or IDAS Dysphoria subscale (5B).
Table 7.
Experiment 3: Mixed effects model predicting Dwell Time (ms) from depression severity (IDAS dysphoria subscale), neutral stimuli valence, and trial.