How should neuroscience study emotions? by distinguishing emotion states, concepts, and experiences

Abstract

In this debate with Lisa Feldman Barrett, I defend a view of emotions as biological functional states. Affective neuroscience studies emotions in this sense, but it also studies the conscious experience of emotion ('feelings'), our ability to attribute emotions to others and to animals ('attribution', 'anthropomorphizing'), our ability to think and talk about emotion ('concepts of emotion', 'semantic knowledge of emotion') and the behaviors caused by an emotion ('expression of emotions', 'emotional reactions'). I think that the most pressing challenge facing affective neuroscience is the need to carefully distinguish between these distinct aspects of 'emotion'. I view emotion states as evolved functional states that regulate complex behavior, in both people and animals, in response to challenges that instantiate recurrent environmental themes. These functional states, in turn, can also cause conscious experiences (feelings), and their effects and our memories for those effects also contribute to our semantic knowledge of emotions (concepts). Cross-species studies, dissociations in neurological and psychiatric patients, and more ecologically valid neuroimaging designs should be used to partly separate these different phenomena.

Keywords: affective neuroscience; amygdala; concepts; emotion; feelings.

Fig. 1.

Rate your position. Which applies to emotions? Indicated are my own (blue), my take on Lisa Barrett’s (Barrett, 2006; Wilson-Mendenhall et al., 2011) (pink) and my take on Jaak Panksepp’s (1998) (gray), to provide three different views (any errors are of course mine). Many of the terms have unclear meanings, and the figure is intended only to give a rough starting point for discussions, not to quantify theoretical frameworks. Lisa saw a prior version of this figure and sent some corrections to my original take on her view. My original depictions of her positions are indicated by circles; the corrected positions from Lisa are denoted by triangles.

Fig. 2.

Features of an emotion. These features underpin the functional properties of emotions, and may help define dimensions or categories of emotions in a mature theory of emotion; they are not necessary and sufficient conditions, and the list is no doubt incomplete. Several of these are modified from (Anderson and Adolphs, 2014).