Method at a glance
- Modality: multimodal (auditory/visual/olfactory CS; shock or nausea US); primarily rodent, with human and nonhuman-primate variants
- Measures: Acquisition and expression of learned threat associations — the capacity of an initially meaningless stimulus, paired with an aversive one, to elicit innate defensive responses (typically freezing) and to modulate instrumental behaviour.
- Procedure: A neutral conditioned stimulus (tone, light, odor) is paired with an aversive unconditioned stimulus (usually footshock). After conditioning, the CS alone elicits species-typical defensive responses. Variants: conditioned taste aversion pairs a taste CS with a nausea-inducing US; conditioned suppression indexes learning by the CS’s suppression of food-motivated bar-pressing.
- Also called: fear conditioning, Pavlovian fear conditioning, threat conditioning, conditioned taste aversion
Strengths
- The best-characterized circuit in behavioural neuroscience — resolved to specific amygdala nuclei and their projections, with cellular and molecular mechanism attached (Johansen et al. 2011).
- Cross-species portability: homologous amygdala circuitry reported in rats, rabbits, nonhuman primates, reptiles, and birds; human imaging and lesion work supports an amygdala role ‘at least to a first approximation.’
- Molecular and genetic continuity extends to invertebrates — many of the transmitters, neuromodulators, intracellular signals, and transcription factors implicated in Aplysia and Drosophila defense learning are implicated in the mammalian amygdala.
- Cleanly separates innate from learned triggers within one circuit, which is what makes it the worked example for Survival circuits.
Limitations
- Freezing dominates the literature and is not representative. Freezing mainly occurs in confined spaces where escape is impossible (Fanselow 1994). Little work exists on other CS-elicited defensive responses, so the paradigm’s canonical measure is partly an artifact of the apparatus.
- Electric shock is analytically awkward. It simulates predator-inflicted tissue damage but its unconditioned stimulus pathways are difficult to trace — one reason the innate/learned circuit relationship remains unresolved. LeDoux flags olfactory US work (Pavesi et al. 2011) as the promising alternative.
- Associability is not uniform. Taste associates readily with gastric discomfort but poorly with shock (Garcia et al. 1968), so CS-US pairings are constrained by evolved preparedness rather than freely substitutable.
- For conditioned taste aversion specifically, the contribution of amygdala areas to learning vs. performance of the avoidance response ‘is less clear than for the standard defense conditioning paradigms.’
- Human resolution is limited to methods whose ‘levels of anatomical resolution obscure circuit details’ — the human literature cannot adjudicate the nucleus-level questions the rodent work is built on.
- The name is a theoretical commitment. ‘Fear conditioning’ imports the feeling word the survival circuit framework is trying to eliminate; LeDoux’s relabeling as defense conditioning is deliberate.
Connections
The paradigm behind most of what is known about survival circuit anatomy, and the empirical base of LeDoux’s career. See LeDoux (2012).
The naming argument
Conventionally fear conditioning. LeDoux (2012) renames it Pavlovian defense conditioning on purpose: the procedure conditions defensive responding, and whether the animal feels afraid is a separate question the paradigm cannot address. Keeping “fear” in the name is precisely the move — using introspective feeling words as guideposts to animal phenomena — the survival circuit framework exists to block. The rename is not cosmetic; it is the thesis applied to the method. See Basic emotions and natural kinds, Can we know whether other animals have conscious feelings?.
Circuit
Sensory inputs reach the lateral amygdala (LA); LA connects to the central amygdala (CEA) directly and indirectly via basal (BA), accessory basal (ABA), and intercalated (ITC) nuclei; medial CEA projects to ventrolateral PAG, whose outputs to motor control regions produce freezing. Unconditioned threats route differently — olfactory ones via vomeronasal → MEA → VMH → PMH → dorsal PAG. Full detail on Amygdala.
What it measures beyond freezing
The same CS does three jobs (LeDoux’s Table 1), which is why the paradigm underwrites the emotion/motivation/reinforcement unification rather than just threat learning:
- Trigger: elicits innate defensive responses.
- Incentive: suppresses food-motivated bar-pressing (Estes & Skinner 1941), facilitates aversively motivated behaviour (Hammond 1970).
- Reinforcer: rats learn new instrumental responses rewarded by elimination of an aversive CS (Cain & LeDoux 2007).
Conditioning is not one thing — an unexpected corroboration
From Damasio (1996), where D. Bishop asks in the discussion whether the Iowa Gambling Task deficit is a failure to inhibit an earlier association rather than a failure to form associations, and whether such patients would be impaired on a standard classical conditioning paradigm.
Damasio’s answer: most patients who fail the gambling task acquire classical conditioning normally — which he takes to imply dissociable aspects of conditioning relative to the nature of the task.
Worth recording here for who is saying it. LeDoux argues from nucleus-level anatomy that “conditioning” fractionates — different circuits by trigger modality, by innate vs. learned status, by CS-US associability (taste with nausea but not with shock; Garcia et al. 1968). Damasio arrives at the same fractionation from the opposite end: a lesion group that fails one task and passes another. Two researchers who agree on almost nothing else in this wiki — see Amygdala, where Bechara & Damasio treat “the amygdala” as exactly the explanatory unit LeDoux denies exists — converge on conditioning being task-relative rather than unitary.
Hold it loosely. Damasio’s claim is stated conversationally in a discussion transcript, with no citation, no n, and the hedge “most patients.” It is an assertion about unpublished clinical observation, not a result. But it is a testable one, and if it holds it constrains what the gambling deficit can be: not a general associative-learning failure, and therefore not explicable by anything this paradigm measures.
The claim now has a procedure and an n
Upgraded with the Bechara, Damasio & Damasio (2000) ingest, which reports the experiment behind the 1996 remark.
Ten VM patients and ten matched controls. Conditioned stimuli were four monochrome slides of different colours; the unconditioned stimulus was a startlingly aversive 100 dB sound; SCR was the dependent measure. Three phases — habituation, conditioning (one colour paired with the US, presented at random among the others), extinction. The VM patients conditioned to the loud noise (Tranel et al. 1996; Bechara et al. 1999a).
Three things this changes:
- The conclusion the wiki drew from the conversational version now rests on something firmer. “The VM cortex is not essential for emotional conditioning,” in the authors’ words — so the absent anticipatory SCRs in the Iowa Gambling Task cannot be explained by a failure to couple an exteroceptive stimulus to a punishment’s somatic state. The fractionation point stands: conditioning and the gambling deficit come apart.
- The convergence with LeDoux noted above is on firmer ground too — arrived at from nucleus-level anatomy on one side and from a lesion group passing one task and failing another on the other.
- But the discount does not disappear, it changes shape. This is still a null result at n = 10, reported in a review with no statistics and no power analysis (see Bechara et al. (2000)), and its primary citations are a Society for Neuroscience abstract and a paper not in
raw/. “Ten patients conditioned, no numbers given” is better than “most patients, no citation” and is not the same as a demonstration.
One anatomical caveat the 2000 paper does not connect but this wiki can. Ventromedial prefrontal cortex (VM) records Tranel et al. (1996) reporting that VM patients whose lesions extend posteriorly into ACC and basal forebrain fail conditioning as well as gambling. The conditioning-intact result above is therefore about anterior VM damage — and the same posterior margin predicts which patients also fail working-memory tasks and show perseveration. So “VM patients condition normally” is a claim about a subgroup, and which subgroup is legible from the lesion.
Relation to this wiki’s clinical material
The direct descendant in the interoception literature is Interoceptive exposure — extinction logic applied to feared bodily sensations rather than external cues. The bridge is LeDoux’s own note that the PAG detects internal physiological signals triggering defense (Schimitel et al. 2012, suffocation alarm): if interoceptive states can serve as unconditioned triggers, then panic-relevant conditioning is a survival circuit phenomenon in his sense. Expect this to become load-bearing when the Week 8 panic/interoception sources (Ehlers; van der Does & Ehlers; Zoellner & Craske) are ingested.