Fig 1.
Schematic depiction of the CPM.
Following a relevant event, components of Appraisal, Motivation, Physiology, and Expression dynamically interact through feedforward (thick arrows) and feedback (thin arrows) processes, producing globally synchronized states that correspond to an emotional episode. CPM, Component Process Model.
Fig 2.
Illustration of the interactive video game.
(A) Experimental conditions defined by a 2 × 3 design manipulating Appraisals of (i) Goal Conduciveness (bad, neutral, good monsters) and (ii) Coping Potential (no-power, power). The colors of bad and good monsters were counterbalanced across participants. (B) Illustration of the maze interface in one trial. The maze configuration and colors varied across trials within each condition.
Fig 3.
Response frequencies across participants for subjective ratings (post scan), using 5-point Likert scales (see Methods).
Response labels used for the different questions are displayed on the right lower panel. (A) Ratings for Appraisals of (i) Coping Potential and (ii) Goal Conduciveness. (B) Discrete emotion ratings for (i) Boredom, (ii) Satisfaction, (iii) Frustration, and (iv) Anxiety. (C) Dimensional emotion ratings for (i) Valence, (ii) Arousal, and (iii) Dominance. Data used in this figure can be found in S1 Data.
Fig 4.
Component brain networks engaged by Appraisal processes.
(A) Effects of Coping Potential showing effects of high and low power. (B) Effects of Goal Conduciveness showing the differential effects between good and bad monsters. (C) Interaction effects between Coping Potential and Goal Conduciveness. Effects are presented on axial slices of a mean brain image created by averaging the participants’ normalized structural images, with a statistical threshold of pFWE < 0.05. Individual beta maps used for this figure are available on Neurovault at https://identifiers.org/neurovault.collection: 8740.
Table 1.
Behavioral measures indexing the Motivation component, averaged across participants (±SD).
Data used in this table can be found in S2 Data.
Fig 5.
Component brain networks whose activity is modulated by Motivation (A), Expression (B), and Physiology (C). Red/orange colors reflect positive and blue colors reflect negative correlations with the respective indices. Effects are presented on axial slices of a mean image created by averaging the participants’ normalized structural images and displayed at a voxel height threshold p < 0.001 with a cluster-level threshold of pFWE < 0.05. Individual beta maps used for this figure are available on Neurovault at https://identifiers.org/neurovault.collection: 8740.
Fig 6.
Overview of the analysis pipeline for the brain-based calculation of synchronization between emotion component networks.
Representative time courses of emotion component networks were computed from the time point by time point scalar product between the emotion component maps learned for the training set and the BOLD time course of the test set. Synchronization between all 4 representative network time courses was estimated using a multivariate version of the instantaneous phase coherence as a similarity metric. Brain maps of regions selectively activated during high synchronization between all components were obtained by computing z-score across all BOLD volumes that were associated with 5% of the highest synchronization values. a.u., arbitrary units; BOLD, blood oxygenation level dependent; FDR, false discovery rate; iPCMulti, multivariate instantaneous phase coherence; TR, repetition time.
Fig 7.
Overlap between brain activation patterns associated with synchronization between emotion components estimated by the brain-based synchronization and the model-based synchronization indices.
Effects are presented on sagittal, coronal, and axial slices of a mean image created by averaging the participants’ normalized structural images. Individual beta maps used for this figure are available on Neurovault at https://identifiers.org/neurovault.collection: 8740.