In 1939 the American psychologists Heinrich Klüver and Paul Bucy removed both temporal lobes (including the amygdalae) from rhesus monkeys in search of the seat of complex visual discrimination, and got something they did not expect: the animals became strikingly placid, unable to recognize threats that had previously terrified them, tried to eat every object (food, feces, stones), tried to mate with every conspecific regardless of sex or appropriateness, and picked up snakes without flinching. The Klüver-Bucy syndrome — psychic blindness to emotional and biological significance — was one of the most theoretically loaded findings of twentieth-century neuroscience. Some small almond-shaped subcortical structure (Latin amygdala, "almond") was responsible for attaching emotional salience to perceptual content.
The amygdala is a cluster of nuclei in the anterior medial temporal lobe, bilaterally symmetric, about 1.5 cm³ in humans. The basolateral complex is the primary input region — sensory information from thalamus, sensory cortices, prefrontal cortex — and the central nucleus is the primary output, projecting to brainstem nuclei for freezing and sympathetic activation and to the bed nucleus of the stria terminalis for sustained anxiety as opposed to acute fear. Fear conditioning is the canonical paradigm: a conditioned stimulus paired with an unconditioned stimulus until the CS alone elicits freezing, autonomic activation, and startle, with the rodent circuitry (LeDoux, Davis, Maren, since the 1980s) mapping CS information arriving at the lateral amygdala, converging with US information, and strengthening synaptic connections through NMDA-receptor-dependent LTP. Two pathways carry the CS: a fast low road (thalamus → amygdala, ~12 ms) delivers a coarse signal sufficient for immediate startle; a slow high road (thalamus → cortex → amygdala) delivers the contextualized signal supporting nuanced responses. Extinction is new learning — a competing safety memory in infralimbic prefrontal cortex that inhibits the fear memory in appropriate contexts — and failure of extinction is the core of anxiety disorders and PTSD. Patient S.M., a woman with bilateral amygdala calcification from Urbach-Wiethe disease, cannot feel fear: she handles snakes, walks through haunted houses laughing, and has been threatened with weapons without responding appropriately. Her case is the canonical demonstration that the amygdala is necessary for the conscious experience of fear in humans, though it is not simply the fear center — it also processes faces, social threat, novelty, and reward.
PTSD is increasingly understood as amygdala-driven hyperconsolidation of traumatic memories combined with prefrontal-cortex failure to extinguish the resulting fear responses. Treatments include prolonged exposure therapy, EMDR, propranolol (a beta-blocker that during memory reactivation dampens reconsolidation of the emotional component), and MDMA-assisted therapy (in Phase 3 trials, ~67% no longer met PTSD diagnostic criteria, vs ~32% on placebo). Anxiety disorders all involve amygdala-circuit dysregulation and respond to similar logic. fMRI of the amygdala is now standard in psychiatric research, with reactivity to threat predicting vulnerability to anxiety and PTSD. AI safety researchers occasionally invoke the amygdala as a biological proof of concept for a small specialized subsystem that vetoes a much larger general-purpose system on the basis of learned aversive associations.