Essentials: The Biology of Aggression, Mating & Arousal

Why does it matter? Because the hormones and circuits driving human aggression are almost nothing like what we assume

Dr. David Anderson's research dismantles some of the most stubborn myths in behavioral neuroscience — testosterone isn't the aggression hormone, solitary confinement makes violent people more violent, and the line between mating and fighting is thinner than anyone wants to admit. The biology is weirder, more specific, and more actionable than the folk psychology version.

• Estrogen receptors, not testosterone receptors, are the molecular signature of aggression neurons in the brain
• Two weeks of social isolation floods the brain with tachykinin-2, and a drug already tested in humans can fully reverse the resulting aggression and fear
• Stimulating fear neurons in the hypothalamus stops an active fight dead — fear is neurologically dominant over offensive aggression
• Activating mating neurons mid-fight causes a male mouse to stop attacking and attempt to court its opponent

The aggression hormone isn't testosterone — it's estrogen

The neurons that drive male aggression don't carry testosterone receptors. When Anderson's lab identified the VMH cells responsible for fighting using a molecular marker, that marker turned out to be the estrogen receptor. Knock out that gene in the VMH and the mice stop fighting entirely.

The mechanism is aromatization: testosterone gets converted to estrogen by the enzyme aromatase, and it's the estrogen that actually activates the aggression circuit. The proof is clean. Castrate a mouse, it loses the drive to fight. Restore it with a testosterone implant — aggression returns. Restore it with an estrogen implant instead — same result. You can bypass testosterone completely.

Anderson notes the irony that most listeners will recognize aromatase from a very different context: aromatase inhibitors are standard adjuvant chemotherapy for breast cancer in women. The same enzymatic pathway is doing heavy lifting in male territorial violence.

The practical reframe is significant. Blaming testosterone for aggression misidentifies the active molecule at the neural level. What's actually driving the fight is estrogen, acting on estrogen receptors in the ventromedial hypothalamus — a fact that reshapes how we should think about pharmacological interventions for aggression in humans.

Two weeks alone spikes a brain chemical that makes mice lethally violent — and a shelved human drug reverses it completely

Isolate a mouse for two weeks and its brain lights up with tachykinin-2. Anderson describes the visual: tag the peptide with green fluorescent protein and the entire brain glows because there's so much of it. That surge isn't cosmetic — it's causal. Postdoc Moriel Zelikowsky showed the tachykinin-2 increase is directly responsible for the spike in aggression, fear, and anxiety that follows social isolation. Block the gene in flies, isolation stops increasing aggression. The same logic holds in mice.

The pharmacological angle is striking. A drug called osanetant, which blocks the tachykinin-2 receptor, was tested in humans and abandoned — not because it was unsafe, but because it failed efficacy trials for the conditions it was tested against. Give it to a socially isolated mouse and something remarkable happens: the aggression disappears, the fear disappears, and the mouse, in Zelikowsky's description, just looks chill. Not sedated — chill. Most remarkably, a mouse so violence-primed it would kill its littermates overnight can, after osanetant, be returned to the same cage and coexist peacefully.

Anderson is blunt about the policy implication: putting a violent prisoner in solitary confinement is absolutely the worst, most counterproductive thing you could do. Isolation doesn't calm dangerous people down. Neurochemically, it makes them more dangerous. A drug with an existing human safety profile could potentially reverse that — if anyone could be persuaded to fund the trial.

Fear beats aggression every time — the hierarchy is wired directly into the hypothalamus

In the ventromedial hypothalamus, aggression neurons sit at the base and fear neurons cluster at the top — neighboring populations with antagonistic jobs. Anderson's pear analogy is precise: fat end at the bottom for fighting, narrower tip for fear, pressed cheek-to-jowl in the same structure.

The functional relationship is hierarchical, not symmetrical. Stimulate the fear neurons while two animals are mid-fight and the fight stops dead in its tracks. Both animals retreat and freeze. Strong fear overrides offensive aggression completely.

The reverse isn't true. Defensive aggression — the cornered-animal response — is actually enhanced by fear, not suppressed by it. That's one of the sharpest distinctions between offensive and defensive aggression at the circuit level: the same fear signal that shuts down an attack launched from a position of dominance will intensify a fight launched from a position of threat.

Anderson speculates the proximity of these populations is evolutionarily meaningful — defensive circuits probably came first, since survival from predation is more urgent than establishing social hierarchy. Offensive aggression may have been built by duplicating and modifying the older defensive architecture. The functional result is an inhibitory relationship: fear neurons have veto power over the fight.

Activate the 'make love' circuit mid-fight and the animal stops attacking and tries to mate with its opponent

The medial preoptic area houses what Anderson calls the make-love-not-war neurons. Activate them in a male mouse while it's actively attacking another male, and the fight stops — the attacker starts singing courtship vocalizations to its former opponent and attempts to mount him. Shut off the stimulation and the fighting resumes.

The VMH aggression neurons and the medial preoptic mating neurons are anatomically close and densely interconnected. Anderson describes the relationship as normally antagonistic — each tends to suppress the other — but the connections allow for cooperation too. The balance between those modes may be what determines whether a mating encounter tips into aggression or stays sexual.

Anderson says the discovery immediately raised a clinical question: whether serial rapists might have a version of this crosstalk where circuits that should be mutually suppressive have become cooperative and mutually reinforcing. The behavior that looks like a monstrous psychological aberration may have a concrete neural substrate — not an excuse, but a potential target. The wiring for that pathology exists in every mammalian brain. What differs is whether those connections are inhibitory or excitatory at any given moment.

Emotions are brain states that persist and generalize — being snappy at home after a bad day is neurobiology, not character

Anderson draws a firm line between feelings and emotional states. Feelings are the tip of the iceberg — the subjective, reportable surface. The state is everything below: the level of neural activity, how long it persists, how broadly it generalizes to new situations.

Persistence is the key diagnostic. A reflex starts and stops with its stimulus — tap the knee, the leg kicks. An emotion outlasts its trigger. Hear a rattlesnake, jump clear, watch it disappear into the brush — your heart is still pounding, your palms still sweating, every stick on the trail still looks like a snake. Anger is the same: long after the argument ends, the arousal lingers and colors the next interaction you have.

Generalization is the other feature. The same emotional state that was triggered at work gets applied at home. Anderson's example: a screaming infant gets soothed if you had a good day and gets a very different reaction if you didn't. That's not a parenting failure — it's a state transferring across contexts, which is exactly what emotional states are built to do.

Framing emotion as a neurobiological state rather than a psychological label makes it scientifically tractable. It also offers a quieter practical insight: when your reactions feel disproportionate, the mismatch is probably real. You're running a state that was calibrated somewhere else.

The VMH is simultaneously a receiver and a transmitter — aggression threshold is continuously computed, not fixed

Anderson's architecture of VMH aggression is hub-and-spoke. The region receives input from roughly 30 brain areas and projects back out to roughly 30 areas — integrating smell, vision, mechanosensation, and other signals into what he describes as a low-dimensional representation of pressure to attack, then broadcasting that signal system-wide.

The pressure model has a direct behavioral correlate. Drive VMH activity higher optogenetically, and the animal's threshold to fight drops: less and less provocation is needed to trigger a full attack. The more the antenna has accumulated, the shorter the fuse.

This means aggression isn't a switch — it's a running calculation. The brain is continuously asking whether the cost-benefit math justifies a fight, because losing a fight can mean death. Prior state, sensory context, hormonal background, social history — all of it feeds into VMH and shifts what counts as sufficient provocation. That's why the same mild irritant produces completely different responses depending on what came before it.

The vagus nerve runs color-coded lines to specific organs — gut knots and racing hearts are real feedback signals, not metaphors

Anderson describes the vagus nerve not as a single calm-down cable but as a labeled bundle of distinct fiber populations: specific lines to the lung governing breathing responses, separate lines to the gut, separate lines to the heart. The gut-in-knots sensation during anxiety isn't poetic — those vagal fibers are sensing actual contraction of gut smooth muscle and reporting it back to the brain.

The communication is bidirectional. The brain sets the state, activates sympathetic and parasympathetic outputs, which change what's happening in the body, which the vagus then senses and feeds back upstream. Emotion isn't just top-down; the periphery participates in maintaining and modulating the state.

What's coming, Anderson says, is the ability to turn specific vagal fiber subsets on or off with precision and ask what each one does to emotional behavior. Breathing techniques and gut-targeted interventions likely work through distinct labeled pathways to the brain — not a diffuse relaxation signal but organ-specific feedback loops with specific central targets.

The next frontier is psychiatric treatment built on causal circuit knowledge — and most of it hasn't been mapped yet

Anderson closes with a candid acknowledgment: the field knows the outlines of these circuits but very little of the detail. Most of the VMH's 30 input and output regions remain uncharacterized in terms of their specific contributions to aggression, fear, or mating. The vagal fiber map is weeks old. The tachykinin-2 drug sits unused in a drawer.

What this conversation actually reveals is that the gap between mechanism and medicine is not conceptual — it's economic and generational. The biology is specific enough to point at real targets. What's missing is the will to run the trials and the researchers to do the mapping.

The next generation of psychiatry won't adjust neurotransmitter levels with a blunt instrument. It will route specific signals through specific labeled lines. That future is closer than it looks — and most of the circuits that will drive it haven't been drawn yet.

---

Topics: aggression, neuroscience, hormones, estrogen, testosterone, VMH, tachykinin, social isolation, mating behavior, periaqueductal gray, vagus nerve, emotion, fear, hypothalamus, optogenetics