The Biology of Social Interactions & Emotions | Dr. Kay Tye

KEYPOINTS

• 01

  The amygdala does far more than fear

  Tye, a Salk Institute neuroscientist and HHMI investigator, notes that the early Klüver-Bucy lesion studies left monkeys emotionally flat toward everything: food, feces, objects. Fear dominated research because it was easy to study. As a graduate student working on the basolateral amygdala, she showed its neurons strengthen their synapses and fire to cues that predict reward, which surprised the field when she presented it at the first Gordon Research Conference on the amygdala.

• 02

  Valence is assigned at a fork in the road

  Valence means net good versus net bad, which Tye distinguishes from value, a scale of magnitude in one direction. Her lab traced amygdala projections and found separate populations sending mostly positive or mostly negative signals to different downstream targets, which drive approach or avoidance. At the time it was hard to publish because few thought neurons in one region could have completely different functions depending on where they project.

• 03

  Novelty gets a single chance to matter

  The amygdala responds to anything new, like an unfamiliar tone or a fire alarm, then quickly stops responding if it predicts nothing. Tye calls this giving everything novel a chance to prove its importance in one trial. It's why someone else's ringtone startles her when her own phone is on do-not-disturb, while its owner barely notices it.

• 04

  The amygdala evaluates, but doesn't generate panic

  Patient SM, who has damage to both amygdalae, shows no response to fearful faces or threats yet can still panic in response to suffocation. Tye's working model is that the amygdala's roughly 13 subnuclei assign importance and meaning rather than producing autonomic arousal themselves.

• 05

  Hunger can flip fear and reward priorities

  Normally the negative-valence amygdala pathway can suppress the reward pathway, because escaping a predator matters more than eating. After one day without food, a big deal for mice given their high metabolism, the balance inverts and reward signaling can silence the fear pathway. The shift appears reversible, and the amygdala may sense hunger through receptors such as those for ghrelin. Tye cites the controversial 2011 Danziger study, in which judges' parole approvals were highest after meal breaks and plunged toward single digits before them. She stresses it was correlational.

• 06

  Loneliness neurons were found by accident

  Former postdoc Gillian Matthews was testing whether dorsal raphe dopamine neurons respond to cocaine. The saline controls showed synaptic strengthening too, because injected mice were housed alone. These neurons, unlike the familiar VTA dopamine neurons, are unpleasant to stimulate: mice avoid places where it happens. Yet stimulating them makes mice more social. Tye compares this to eating out of uncomfortable hunger rather than pleasure, and concludes these may be loneliness neurons, encoding an aversive need for social contact.

• 07

  Social homeostasis: your set point shifts

  Acute isolation produces a rebound of affiliative behavior when reunited. Chronic isolation produces aggression, territoriality or avoidance across humans, monkeys, mice and even flies. Tye's model is that you detect a deficit and try to reconnect, and if those efforts fail long enough, behavior drops off sharply and you adapt to a new baseline. Then your old normal level of contact feels like overload. She thinks many people experienced this shift going into and out of the pandemic.

• 08

  Empathy tracks whether someone is an ally or an adversary

  Tye defines empathy as understanding another's emotion and taking it on, distinct from emotional contagion like group panic. She speculates that empathy is inversely correlated with seeing someone as a competitor standing in your way. Once you realize someone is out to get you, empathy disappears. Long isolation may make everyone feel like a competitor. Whether empathy for positive and negative events is truly asymmetric remains an open question.

• 09

  Prefrontal neurons predict who wins

  In stable groups of four cage-mate mice, food-deprived and competing for a reward only one could collect, prefrontal activity stably encoded who was dominant. It also predicted above chance, up to about 30 seconds ahead, which mouse would win the next trial. For subordinates, prediction accuracy jumped near the cue, suggesting they watch the dominant and go when it seems disengaged. Dominants don't watch the subordinates at all. Huberman likens this to Don Draper in Mad Men saying, "I don't think about you at all."