written: 2019-06-09
last run: 2019-06-09

Description of task

The EA task presents videos (visual and auditory) of actors recounting a positive or negative experience. While undergoing fMRI, participants are instructed to dynamically rate how they think the actor is feeling on a 9-point Likert scale (“How negative or positive does this person feel?”).

We presented a subset of 9 EA videos across 3 blocks, for a total run-time of ~20 minutes. The videos were presented in the same order for all participants, as follows:

video order actor sex valence emotion description timepoints
2 second AR male negative sad soccer 66
3 sixth ME male positive amused movie 70
4 seventh NW female positive delighted trip 91
5 eighth TA female positive amused comedian 74
6 third CT male negative anger paycheck 74
8 fifth HR female negative sad death 83
10 fourth DH female negative anger truck 56
12 first AR male positive delighted wedding 73
13 ninth TA female negative anger room-mate 61

Gold standard ratings

The EA task also provides ratings of emotional valence from the actors depicted in the stimuli, i.e., after recounting their story, the actors watched their footage and indicated dynamic changes in their own emotion. This rating is taken as the “ground truth” or “gold standard” rating. The ratings provided by the actors are as follows:

HC ratings (n=152)

The plot below shows the average rating of HCs in the thick blue line. Individual ratings from HC participants are represented as blue dots, with the increasing size of the dot indicating increasing agreement. For comparison, the gold standard rating is shown in the thick black line.

Summary. We see that there’s a wide range of HC values. At some timepoints, the actor has rated themselves in a way that most HCs did not!

SSD ratings (n=230)

This plot below shows the average rating of SSD participants in the thick orange line. Individual ratings from SSD participants are represented as orange dots, with the increasing size of the dot indicating increasing agreement. As above, the gold standard rating is shown in the thick black line.

Summary: We see that the SSD participants have very similar ratings to the HCs. In fact, SSD and HC participants are more similar to each other than they are to the gold standard ratings. As an overlay of mean SSD and HC scores is near-identical, it is not displayed here.