Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Fig 1.

Example trials for face similarity task.

Each trial presents a pair of emotional faces. Participants rate the similarity of the expressed emotions. The images pictured here are similar, but not identical, to the original images used in the study, and are for illustrative purposes only.

More »

Fig 1 Expand

Fig 2.

Ground truth calculations.

A) The ground truth affective similarity is based on valence and arousal ratings. These ratings served as x-y coordinates for each face. The affective similarity score was calculated as the Euclidean distance between these values. B) The ground truth visual similarity is based on the Euclidean distances between all 68 facial landmarks. These landmarks were obtained from automated software (ResNext) for each face. Distances between each landmark were represented in a 68x68 matrix for each image. The global visual similarity score was calculated as the correlation between these two vectorized matrices.

More »

Fig 2 Expand

Fig 3.

Cue use calculations.

The participants’ similarity ratings were predicted by A) the affective similarity score, to calculate the affective cue use score, and B) the visual similarity score to calculate the global visual cue use score.

More »

Fig 3 Expand

Fig 4.

Effect of mask exposure on eye cue use in Study 1.

Use of eye cues in emotion judgements increased as exposure to masks increased. Transparent band reflects 95% confidence interval. Created using ggplot2 (v3.3.0) in RStudio.

More »

Fig 4 Expand

Fig 5.

Change in facial cue use between time 1 and time 2 in Study 2.

A) Eye and mouth cues use increased from time 1 to time 2. B) Arousal cue use decreased from time 1 to time 2. * p < 0.001. Error bars represent 95% confidence intervals. Created using the seaborn package (v0.10.0) for Python (v3.7.6).

More »

Fig 5 Expand

Fig 6.

Effect of mask exposure on global visual cue use as a function of social interaction in Study 2.

Social interaction mediated the relationship between mask exposure and visual cue use between time 1 and time 2. People with the least amount of social interaction had the greatest decrease in global visual cue use, and people with the most amount of social interaction had the greatest increase in use of global visual cues. Transparent bands reflect 95% confidence intervals.

More »

Fig 6 Expand