Neuroscience of Emotions & Tools for Improving Emotion Regulation | Dr. Ralph Adolphs

I’m Andrew Huberman, and today I’m talking with Dr. Ralph Adolfs, a Caltech professor working across psychology, neuroscience, biology, animal research, and human emotion. Ralph pushes back on the easy story that one brain area equals one feeling. I wanted to understand what emotions are, why they commandeer decisions, and how we can regulate them without becoming robotic.

I’m trying to connect fields that barely speak to each other: animal work, social psychology, neuroscience. I think emotions are functional states, defined by what they accomplish rather than particular hardware in a human, mouse, octopus, or perhaps AI. They guide behavior when a challenge is too variable for reflex but too urgent for leisurely planning.

Fear, anger, disgust, happiness: clear examples, but what makes them one category? Emotions have priority; they interrupt what I’m doing. They have valence, broadly approach or avoidance, pleasant or unpleasant. They also scale: a distant threat can produce watchfulness, then fear, then panic. And they persist, because the threat may still be out there.

That persistence is not merely memory. Patients with profound amnesia watched sad films, became tearful, and still felt sad minutes later, though they could not recall the film. Sadness remained while the event disappeared. That differs from a reflex: pull your hand from heat and it’s over; if there’s a bear in the forest, the internal state must remain active long enough to keep you safe.

Many animals meet these functional criteria, and perhaps some AI systems may as well. Conscious feeling is harder. We naturally treat how emotion feels as ground truth, but we cannot directly test conscious experience in animals. I don’t deny feelings in people or animals. I’m saying we can study emotional processes across species without first solving consciousness, which is a tall order.

I asked about noticing irritation before it takes over: getting fed up in a checkout line, or watching people melt down online and on airplanes. I also floated deliberate cold exposure. It sounds comical, but ice baths may be practice in not letting the first autonomic surge run the whole show.

Humans can monitor our emotions, and that gives us leverage. Most difficulties are not that emotion exists; they’re about regulation. Emotional granularity matters: the finer my sense of what’s happening, the more responses I have. It needn’t be verbal, but I need enough of a concept to differentiate what is happening and evaluate it.

Regulation is trainable, but reappraisal can cut both ways. If I keep telling myself, ‘I don’t want to feel sad,’ I may intensify sadness. Earlier intervention is generally better: structure my environment to avoid situations that reliably make me anxious, guilty, or angry. Once I’m in one, strategies are individual. The goal is control that becomes quick, strong, and automatic.

That’s why I got interested in ice baths. At first they were awful: pain, rapid breath, rapid heart rate. With practice, my breathing and heart rate now settle almost immediately. The people at the ice shop joke about my parties. It also seemed to generalize: a driver honking while I parked once triggered instant anger; afterward, the response was flat. It’s an experiment of one, not a published claim, but it felt like autonomic training.

I’ve also come to value gratitude and awe. After adversity with an endpoint, belief in my ability to endure matters, but so does waking with appreciation that I’m here. Awe may depend on a human capacity to step outside the immediate moment: imagine your perspective, revisit the past, picture futures, and zoom out from present sensation. That flexibility gives us planning, fiction, and the distinctive vastness of awe.

We are far less accurate at reading emotion from a face than we think. Basic-expression pictures are posed, extreme actor displays paired with multiple-choice labels. Ordinary faces are dynamic and ambiguous, so I need context, tone, what happened, and someone’s usual pattern. Dog owners are certain they recognize guilt in dogs, but under controlled conditions they do about as well as chance.

The amygdala case of patient SM made this vivid. She showed little fear toward external threats—haunted houses, snakes, spiders—but inhaling carbon dioxide and experiencing suffocation produced a full panic attack. Fear of something outside and panic arising from inside the body recruit different systems. Saying the amygdala is simply a fear center is much too crude.

Social intelligence is not one perfect skill. I want many ways to regulate emotion and choose according to the situation; we regulate one another constantly. With autism, I view social traits as dimensions of individual difference, not just a category. In webcam studies, people higher on autistic traits tend, on average, to look less at faces and social cues and more at distractors. But individuals vary enormously, and these gaze patterns are fairly stable over time.

Being alone with your mind can be training. Running, meditation, or an ice bath removes external demands and lets me practice control of thought and emotion. My lab begins meetings with silence. It started as remembrance in 2020 and became a way to settle before work. Three quiet minutes can stress new students, but it becomes easier. Switching tasks also leaves residue, so transition rituals may matter. During my illness, I regulated fear, worry, and sadness with my wife, then found more acceptance, gratitude, and awe. Death is not only the world losing me; eventually, I lose the world. So: live it up.

That is a beautiful place to leave it. Ralph, thank you for bringing careful science to emotions while showing how these ideas appear in your own life: ice baths, silence, illness, gratitude, and awe. Thank you all for joining us, and thank you for your interest in science.