Four criteria for localizing emotions to specific brain regions

Chapter 1 endnote 44, from How Emotions are Made: The Secret Life of the Brain by Lisa Feldman Barrett.
Some context is:

Overall, we found that no brain region contained the fingerprint for any single emotion.

To confidently claim emotions to be localized to different regions of the brain, our meta-analysis would have to meet four criteria:

Statistical significance

If a group of voxels increased in activation, this must occur with a probability greater than what would be expected by chance.

Reliability (also called consistency)

Experiences or perceptions of an emotion (say, fear) must be consistently associated with increased activity in the voxels within the claimed brain region (say, the amygdala). When analyzing studies of emotional experiences, for example, this would mean that regardless of the way that the emotion is evoked (e.g., using memory, movies, images, sounds, etc.), regardless of who the test subjects are (e.g., men, women), regardless of the whether test subjects reported their experiences or not, and so on, we must see a consistent link between the activity in a particular brain region and the experience of a particular emotion.

Specificity

The voxels in a given brain region must increase in activity only during the experience or perception of that emotion (e.g., fear), but not during the experience or perception of other emotions (anger, disgust, etc.) nor during non-emotional events (thoughts, perceptions, etc.).

Alternative explanations do not apply

The voxels of a given brain region must not increase in activity in relation to other factors such as the type of stimulus used, the type of experimental method used, and so on.

Of these criteria, only reliability was met in our meta-analysis, but not at a high enough level to confirm that the neurons in these voxels were an emotion essence. For example, the amygdala showed a consistent increase in activity for studies of fear, more than what you’d expect by chance, but only for 27% of the experience studies and 42% of the perception studies. Specificity was not supported because the amygdala showed this consistent increase for every other emotion we studied as well (anger, disgust, sadness, and happiness). Amygdala activity also routinely increases during events that are considered cognitive and perceptual, not just emotional. This includes when you meet new people, when you are trying to remember something, when you are feeling pain, when you are making a decision, and when you are viewing a beautiful sunset. Because this information is new to you, the neurons in your amygdala are probably more active right now as you read these words.

This pattern of results was basically the same for every emotion we tested. For example, scientists originally thought that the front part of the insula (the anterior insula) was the neural essence of disgust. In our meta-analysis, that brain region showed an increase in activity in only 10% of the experience studies and 8% of the perception studies. This was more than what you’d expect by chance, but much less than what you’d expect for the essence of disgust. In addition, activity in the anterior insula increased for every other emotion we studied, especially sadness (27% of studies). The anterior insula is also routinely activated when you are in pain, when you are smelling and tasting, when you are speaking or listening to someone else speak, when you’re shifting your attention from one thing to another, and frankly in almost every brain imaging study that is ever conducted! It is a “hot spot” the human brain.