Abstract
Our ability to shift from one emotion to the next allows us to adapt our behaviors to a constantly changing and often uncertain environment. Although previous studies have identified cortical and subcortical regions involved in affective responding, none have shown how these regions track and represent transitions between different emotional states, nor how such responses are modulated based on the recent emotional context. To study this, we commissioned new musical pieces designed to systematically move participants (N = 39, 20 males and 19 females) through different emotional states during fMRI and to manipulate the emotional context in which different participants heard a musical motif. Using a combination of data-driven (hidden Markov modeling) and hypothesis-driven methods, we confirmed that spatiotemporal patterns of activation along the temporoparietal axis reflect transitions between music-evoked emotions. We found that the spatial and temporal signatures of these neural response patterns, as well as self-reported emotion ratings, were sensitive to the emotional context in which the music was heard. In particular, brain-state transitions associated with emotional changes occurred earlier in time when the preceding affective state was of a similar valence to the current affective state. The findings argue that emotional changes are an essential signal by which the temporoparietal lobe segments our continuous experiences, and further clarify its role in linking changes in external auditory signals with our dynamic and contextually dependent emotional responses.