Dr. David Anderson on the Brain Circuits of Aggression, Mating, Fear and Social Isolation

High-Level Overview

Andrew Huberman sits down with Dr. David Anderson, a Caltech professor of biology, Howard Hughes Medical Institute investigator and author of The Nature of the Beast: How Emotions Guide Us, for a deep dive into how the brain produces aggression, sexual behavior, fear and arousal. Anderson argues that emotions are best understood as internal states, neurobiological processes with measurable properties like persistence and intensity, rather than as feelings we can only study by asking people. From there, the conversation walks through decades of work in mice and fruit flies: switches for aggression in the hypothalamus, why fear and fighting neurons sit side by side, how estrogen (not only testosterone) drives male aggression, why new mothers turn fierce, what male-male mounting really means, and how a peptide called tachykinin links social isolation to rage and anxiety.

The tone is curious, collegial and technical, with Huberman adding personal stories, a few wry jokes and occasional detours into human questions like fetishes, online communities and school shootings. Anderson is careful to separate what is known, what is unknown, and what he personally doesn't know. The conversation moves from definitions to lab findings to human implications, ending with a plea for drug companies to test a promising compound for isolation-driven stress.

The Discussion Explores:

• Why Anderson treats emotions as internal states with dimensions like persistence, generalization, intensity and valence
• Evidence that arousal isn't one thing, but comes in behavior-specific forms run by separate circuits
• The ventromedial hypothalamus (VMH), where optogenetics revealed an aggression switch next to fear neurons
• Offensive, defensive and predatory aggression, and why offensive aggression can be rewarding to mice
• The surprising role of estrogen and progesterone in male aggression, and the switch that makes nursing mothers aggressive
• How mating and fighting circuits interact, including "make-love-not-war" neurons and dominance mounting
• Tachykinin, social isolation and a shelved drug that calms isolated mice
• The brain-body connection, the vagus nerve and the somatic marker hypothesis

Deep Dive into Main Ideas

Emotions as States, Not Just Feelings

Anderson sees emotion as one class of internal state, alongside arousal, motivation and sleep. Internal states change how the brain turns input into output: when you are asleep you don't hear what you would hear awake. Framing emotion as a state makes it a neurobiological process that can be studied in animals, avoiding the endless debate over definitions that comes from equating emotion with subjective feeling, which only humans can report. He compares it to an iceberg: the feeling is the tip above the surface of consciousness, and the state is the much larger part below.

With his colleague Ralph Adolphs, he has expanded the classic two dimensions of emotion, arousal (how intense) and valence (good or bad), to include features like persistence and generalization. Reflexes stop when the stimulus stops, like a knee jerk. Emotions outlast their triggers: hear a rattlesnake on a Southern California trail and your heart keeps pounding, and you jump at sticks for a while after the snake is gone. Generalization means a state spills into unrelated situations, like reacting to a screaming child very differently after a bad day at work than after a good one. These features separate emotions from motivational states like hunger or thirst, which are specific and end once the need is met.

Arousal Comes in Different Flavors

Huberman asks whether a simple neurochemical signature, such as dopamine, flips arousal from positive to negative. Anderson thinks it's more likely a circuit flip than a chemical one; the same chemical might be involved in both. In a 2009 fruit fly study, his lab found two types of arousal: sleep-wake arousal, and a startle-like arousal where flies puffed with air came back more and more vigorously, like wasps you keep swatting away from a picnic burger. Both required dopamine but ran through completely separate circuits.

The takeaway is that arousal isn't unitary. There appear to be behavior-specific forms, sexual arousal feels different from aggressive arousal, and whether a truly general arousal exists is still an open question. Huberman notes this upends his old picture of arousal as a single continuum running from coma to panic attack.

Finding an Aggression Switch in the Hypothalamus

Anderson stresses that "aggression" describes a behavior, not a state; it can reflect something like anger, fear or even hunger in predatory aggression. Walter Hess's Nobel Prize-winning work showed that stimulating a cat's hypothalamus could trigger defensive rage (ears back, teeth bared, hissing) or predatory attack depending on electrode placement, and Menno Kruk later triggered fighting in rats. But when postdoc Dayu Lin tried electrical stimulation of the ventromedial hypothalamus (VMH) in mice, it failed in over 40 attempts, producing only freezing and crouching. Anderson convinced her to try optogenetics, which was newly usable deep in the brain, and it worked.

The explanation lies in the anatomy. Anderson describes the VMH as shaped like a pear sitting on the ground: aggression neurons sit in the fat bottom and fear neurons in the upper part. In a tiny mouse brain, electric current spreads through the whole pear and the fear circuits dominate, while optogenetics can target just the aggression cells. Later work showed this offensive aggression is rewarding to male mice: they will learn to poke their nose or press a bar for a chance to beat up a subordinate male, and they gravitate to a compartment where those neurons are activated. Huberman and Anderson link to striking videos in which stimulating these neurons turns a mating mouse into an attacker, or drives a lone mouse to attack a glove.

Types of Aggression and Why Fear Sits Next to Fighting

Anderson distinguishes offensive aggression (the rewarding kind seen in mice), defensive aggression (the unpleasant fury of being attacked or cheated, whose brain circuitry is still a mystery) and predatory aggression (used by mice to catch crickets, running through different circuits). In rats, offensive bites target the flank while defensive bites go for the neck and throat. Huberman adds that some people, such as a lawyer relative and a former special operations friend, openly love to fight.

Why would fear and aggression neurons be neighbors? Anderson offers an evolutionary guess: defense against predators probably came first, and aggression cells may have arisen by duplicating and modifying fear cells; developmentally they share a common origin. Functionally, strong fear shuts offensive aggression down, while fear actually enhances defensive aggression. Stimulating the fear neurons during a fight stops it dead, with both animals retreating to freeze. The VMH also contains metabolic neurons (destroying it in rats causes obesity), and Anderson describes a cluster of what he calls the four Fs, "feeding freezing fighting and mating," intermingled perhaps so the brain can prioritize among them.

Hydraulic Pressure: Drives That Build Up

Huberman raises Konrad Lorenz's idea of hydraulic pressure toward behavior, noting that once he loses his temper, it takes a long time to simmer down. Anderson separates need-based homeostatic drives like hunger, thirst and temperature, which work like a thermostat, from aggression, which he doesn't think builds up as an accumulating need to fight (Huberman jokes that Twitter may be the exception). Scott Sternson's work shows that activity in feeding circuits climbs as animals get hungrier and drops when they eat, and that this state feels negative.

In aggression, the harder the VMH is driven, the less it takes to set the animal off, but with nothing to attack, a stimulated mouse just wanders. The same holds for mating circuits: stimulate them and nothing happens until another mouse appears, at which point the male will try to mount it or even a kumquat, "any port in the storm." How the sight or smell of a target releases the valve is a key open question. Anderson describes the VMH, with about 30 inputs and 30 outputs, as both an antenna and a broadcasting center, like a satellite dish that integrates sensory information and broadcasts a pressure to attack across the brain, including the systems that weigh the risk of fighting. Stimulating the neighboring fear neurons produces pupil dilation, stress hormone release and a faster heart rate.

Hormones and the Mother's Switch

Huberman seizes the chance to bust a myth: testosterone doesn't simply make animals aggressive, and estrogen doesn't make them placid. When Anderson's lab found a molecular marker for the VMH aggression neurons, it was the estrogen receptor. Knocking out that receptor in male mice stops fighting, and work from Nirao Shah shows that a castrated male's aggression can be restored with estrogen as well as testosterone, because much of testosterone's effect comes from its conversion to estrogen by the enzyme aromatase. Giving males aromatase inhibitors, which are used against breast cancer in women, stops both fighting and sexual activity in animal studies. The aggression neurons also carry progesterone receptors.

Female mice are different: they fight mainly while nursing pups, becoming hyper-aggressive until the pups are weaned. A virgin female mates with a male, while a new mother attacks him. Anderson's student Mengyu Liu found two subsets of estrogen receptor neurons in the female VMH, one for fighting and one for mating. Stimulating the fighting subset makes even a virgin attack; stimulating the whole population had canceled the effects out. Like a seesaw, mating neurons dominate in virgins and fighting neurons in mothers. A nursing mother doesn't need her pups present to attack, and her aggression neurons fire equally for male and female intruders, whereas a male's fire only for males. Male mice also have male-specific aggression neurons, and fruit flies turn out to have female-specific fighting neurons.

Where Mating and Fighting Meet

Huberman describes watching ferrets mate, with neck biting and squealing, and asks whether mating and aggression circuits blur. Anderson notes that within the male aggression region there is a subset of neurons activated by females during mating, and silencing them makes mating less effective. The main mating region is the medial preoptic area (MPOA). Activating mating neurons there in a male mid-attack makes him stop fighting, start singing and try to mount his opponent. Anderson calls those "make-love-not-war neurons" and the VMH cells "make-war-not-love" neurons. The two areas are densely connected, partly by mutual inhibition, but may also cooperate, which could explain how a tender moment between lions turns into a snarl.

On fetishes, Huberman shares a psychiatrist friend's observation that people tend to develop fetishes for things that are usually disgusting, such as feet, feces or dead bodies, rather than for mailboxes. Anderson suggests appetitive conditioning: something natively aversive, repeatedly paired with reward, flips its valence. He raises a sobering point: "there is no treatment that we have for a violent sexual offender that eliminates the violence but not the sexual desire," since both physical and chemical castration remove both. They also discuss temperature: the MPOA contains temperature-regulating neurons, as does the VMH, and imaging reveals separate neurons for sniffing, mounting, thrusting and ejaculation. Anderson wonders if temperature ties into aggressiveness, noting that violence rises in summer, though social reasons may explain it.

Mounting: Sex or Dominance?

When two male mice mount each other, is it sexual? Machine learning can't easily tell male-male from male-female mounting apart, because the movements look so similar. Two clues solved it. First, context: male-male mounting is most common in inexperienced fighters and gives way to attack as they gain experience. Second, at the suggestion of computational neuroscientist Ann Kennedy, the lab listened for ultrasonic songs, which males sing when mounting females. Males mounting males don't sing at all. Dominance mounting is driven by weak activation of the VMH, while sexual mounting is driven by the MPOA and always comes with song. The lesson is that "the same behavior can mean very different things depending on the context."

Females also show male-type mounting. A female returned to her sisters after mating will start mounting them, and stimulating the MPOA mounting neurons in a female makes her crawl out from under a male and mount him, echoing earlier work from Catherine Dulac showing that male-type mounting circuits exist in females. Huberman connects dominance mounting to jiu-jitsu and to fighters who mock-thrust on a beaten opponent, and to dogs, where dominance hierarchies within a litter cross sex lines.

The PAG and Fear-Induced Analgesia

The periaqueductal gray (PAG) seems involved in almost everything, from pain to fear to the female mating posture. Anderson compares it to an "old-fashioned telephone switchboard," routing signals for many innate behaviors. The dorsal PAG handles panic-like running and the ventral PAG freezing, and hypothalamic regions project to different sectors, arranged like numbers on a clock face (in cross-section, he admits, it looks like water in a toilet bowl). Huberman recalls that punches during martial arts barely hurt until afterward; Anderson explains fear-induced analgesia and mentions an adrenal peptide, BAM22, which acts on a receptor his lab found and turns out to suppress pain. Whether such analgesia happens during offensive aggression or mating, and whether in the PAG or the spinal cord, he says he doesn't know.

Tachykinin, Social Isolation and a Shelved Drug

The most practical part of the episode concerns tachykinins, a family of neuropeptides (tachykinin 1 is substance P). Searching for peptides that control aggression in flies, Anderson's lab found that tachykinin neurons strongly promote fighting, and that social isolation, which increases aggression in flies as in people, raises tachykinin levels. "Putting a violent prisoner in solitary confinement is absolutely the worst most counterproductive thing you could do to them," he says. Former postdoc Moriel Zelikowsky found that two weeks of isolation causes a massive rise in tachykinin 2 in mouse brains, enough that a fluorescent tag makes the brain glow green, and that this rise drives the increased aggression, fear and anxiety.

A drug called osanetant, which blocks the tachykinin 2 receptor and has a good human safety record, reverses all of it. Isolated mice given it "just look chill," without sedation, and an isolated male that would otherwise kill his littermates overnight can safely live with them again. In humans, psychiatrist Emil Coccaro found that aggression in people with borderline personality disorder correlates with blood levels of tachykinin 1. Yet drug companies won't test these compounds for stress or bereavement: after they failed in schizophrenia trials, where a failed phase three trial costs about $100 million, the industry concluded animal data don't translate. Anderson even proposed that a company running a hot flash trial during COVID ask isolated participants about stress, but it declined because any bad finding would have to be reported to the FDA: "it's better not to ask and not to know." He also suggests testing it for separation anxiety in pets bought during the pandemic. Huberman puts out a call for researchers or funders to pursue it, noting that isolation is a common thread among perpetrators of mass shootings.

The Body in Emotion

Finally, Huberman asks how to think about the body in emotion, citing heat maps in Anderson's book showing where people say they feel anger, sadness or calm. Anderson connects this to Antonio Damasio's somatic marker hypothesis, the idea that feelings are tied partly to sensations in particular body parts, which may reflect changes in blood flow. Brain and body talk in both directions through the sympathetic and parasympathetic systems and the vagus nerve, whose fibers both sense the body (like a knotted stomach) and send commands to it. Recent work is decoding specific vagal fibers to the lungs, gut and other organs as labeled lines, and new tools may soon let scientists switch particular fibers on and off to see how they shape emotional behavior. He calls this brain-body loop a central feature of emotion states, something Darwin recognized.

Conclusion

Anderson's big idea is that emotions are internal states built from specific, intermingled circuits, and that the same behavior (a bite, a mount, a chase) can come from very different states depending on context. His work shows aggression switches sitting next to fear cells, estrogen driving male fighting, a seesaw that makes nursing mothers fierce, and neurons that can turn war into love and back. The most urgent takeaway is about social isolation: it raises a peptide that fuels aggression, fear and anxiety, and an existing, well-tolerated drug reverses this in mice, yet economics keep it from being tested in people. Anderson is candid about how much is still unknown and hopes the next generation of neuroscientists will fill those gaps, which he calls essential for better psychiatric treatment.

This is a summary of "The Biology of Aggression, Mating, & Arousal | Dr. David Anderson" from Huberman Lab, giving you the GIST so you can decide if you want the REST. Listen to the full episode.