HSRP-310
Human amygdala responses to visual threat begin at approximately 74 ms, preceding visual cortex and selective to low spatial frequency; replicated at 88 ms for backward-masked stimuli outside conscious report. Behavioural threat judgement is possible from backward-masked faces at 39 ms. SUPERSEDES the widely circulated figure of 12 ms, which derives from auditory fear conditioning in RATS (LeDoux 1996; Quirk, Repa & LeDoux 1995) and is not a human value, nor a visual one. COUNTER-EVIDENCE AND SCOPE: the subcortical-route model is contested. Pessoa and Adolphs propose a "many roads" account and note that a visual equivalent of the rodent auditory pathway in humans is assumed rather than demonstrated. The human auditory low road itself remains newly mapped. Both human studies used epilepsy patients with implanted electrodes, who are not neurologically typical; this is the only way to obtain these latencies in humans and the constraint should be stated. All measurements used fearful FACES, not conceptual or ideological threat — the framework's extension to belief-threat is inference, not measurement. The claim that survives without the route: affective appraisal precedes deliberative evaluation. That is not in dispute.
Proof page — generated from the HSRP research notes register. Where a block has no data, it says so.
1. Header
2. The claim, in full
Human amygdala responses to visual threat begin at approximately 74 ms, preceding visual cortex and selective to low spatial frequency; replicated at 88 ms for backward-masked stimuli outside conscious report. Behavioural threat judgement is possible from backward-masked faces at 39 ms. SUPERSEDES the widely circulated figure of 12 ms, which derives from auditory fear conditioning in RATS (LeDoux 1996; Quirk, Repa & LeDoux 1995) and is not a human value, nor a visual one. COUNTER-EVIDENCE AND SCOPE: the subcortical-route model is contested. Pessoa and Adolphs propose a "many roads" account and note that a visual equivalent of the rodent auditory pathway in humans is assumed rather than demonstrated. The human auditory low road itself remains newly mapped. Both human studies used epilepsy patients with implanted electrodes, who are not neurologically typical; this is the only way to obtain these latencies in humans and the constraint should be stated. All measurements used fearful FACES, not conceptual or ideological threat — the framework's extension to belief-threat is inference, not measurement. The claim that survives without the route: affective appraisal precedes deliberative evaluation. That is not in dispute.
3. What it rests on
-
Mendez-Bertolo, Moratti, Toledano, Lopez-Sosa, Martinez-Alvarez, Mah, Vuilleumier, Gil-Nagel & Strange, "A fast pathway for fear in human amygdala," Nature Neuroscience 19:1041-1049 (2016).
Sources: SRC-429 -
Intracranial depth electrodes in eleven epilepsy patients. Amygdala responses began 74 ms after stimulus onset to fearful faces — but not to neutral or happy faces — at considerably shorter latency than fear responses recorded in visual cortex.
Sources: SRC-429 -
Two features confirm the route rather than merely the timing. The fast responses appeared only for low spatial frequency components, exactly as a magnocellular subcortical pathway predicts. And they were not evoked by arousing scenes generally: the route is tuned to socially relevant threat, not to danger in the abstract.
Sources: SRC-429 -
Time-frequency analysis in the second study below located a low-gamma signature from as early as 45 ms (J. Neuroscience 43:1405, 2023).
Sources: SRC-430 -
Second study, independent: an amygdala potential beginning 88 ms post-stimulus, preferentially evoked by fearful faces rendered invisible by backward masking, again low-spatial-frequency selective (J. Neuroscience 43:1405, 2023). The amygdala discriminated threat in stimuli the subject could not consciously perceive.
Sources: SRC-430 -
Rodent antecedent, correctly labelled: Quirk, Repa & LeDoux, Neuron 15:1029-1039 (1995). 74 single units across 24 freely behaving rats; conditioning effects concentrated in response components arriving under 15 ms, too early to have passed through cortex given the same laboratory's stimulation data showing cortical activation of the amygdala could not occur under 20 ms. These are rodent auditory values and are not human figures.
Sources: SRC-428
No confidence intervals are recorded for any figure in this block.
4. Sources
-
SRC-428 · PRIORITY · Fear conditioning enhances short-latency auditory responses of lateral amygdala neurons — Quirk, Repa & LeDoux, Neuron 15(5):1029–1039 (1995)
Pinpoint: not recorded · Access: Access status not recorded
-
SRC-429 · PRIORITY · A fast pathway for fear in human amygdala — Mendez-Bertolo et al., Nature Neuroscience 19(8):1041–1049 (2016)
Pinpoint: not recorded · Access: Access status not recorded
-
SRC-430 · PRIORITY · Rapid Processing of Invisible Fearful Faces in the Human Amygdala — J. Neurosci. 43(8):1405 (2023)
Pinpoint: not recorded · Access: Access status not recordedIncomplete record: author list, full page range not yet recorded.
-
SRC-431 · PRIORITY · Emotion processing and the amygdala: from a 'low road' to 'many roads' of evaluating biological significance — Pessoa & Adolphs, Nat. Rev. Neurosci. 11(11):773–783 (2010)
Pinpoint: not recorded · Access: Access status not recorded
-
SRC-550 · PRIORITY · The Emotional Brain: The Mysterious Underpinnings of Emotional Life — Joseph E. LeDoux (Simon & Schuster) (1996)
Pinpoint: not recorded · Access: Access status not recordedThe popularising source of the "low road / high road" framing and of the millisecond figures that circulate in secondary accounts. Cited here as the route by which rodent auditory latencies entered discussion of HUMAN VISUAL threat processing, where they do not apply. The underlying experiment is SRC-428 (Quirk, Repa & LeDoux 1995), which is rodent auditory fear conditioning. Not a source for any human figure.
5. Counter-evidence
-
Pessoa & Adolphs, Nature Reviews Neuroscience 11:773-783 (2010) argued from physiological data that the subcortical route does not process threats faster than the cortical one, proposing a "multiple waves" model. de Gelder, van Honk & Tamietto replied in the same journal (2011) defending the subcortical account. The 2016 intracranial data postdates and addresses the challenge directly.
-
Latency estimates remain heterogeneous across methods. MEG studies commonly report sub-100 ms amygdala responses; several single-neuron and intracranial studies in humans and monkeys report latencies beyond 100 ms.
Source: none linked in the register -
Sample constraint: eleven patients, all with epilepsy, all with electrodes placed for clinical rather than research reasons. Electrode placement is not experimenter-controlled and the population is not neurologically typical.
Sources: SRC-429
6. Why this grade
Direct intracranial recording in humans, not inference from rodent work and not reconstruction from scalp EEG. Two independent studies, two latencies, two stimulus conditions — one of them unconscious presentation.
The spatial-frequency selectivity is what raises this above a timing observation. A fast response alone would be consistent with several routes. Low-spatial-frequency selectivity is a positive prediction of the magnocellular subcortical pathway specifically, and it was confirmed.
The social-threat tuning matters for the framework's use of this finding: identity-protective cognition is triggered by challenges to standing within a group, not by danger in general. The route is tuned to exactly that class of stimulus.
Held at ESTABLISHED for the human precedence finding. The specific anatomical route remains CONTESTED and is graded separately.
The register defines ESTABLISHED as: Supported directly by primary texts or published findings.
7. Correction history
No correction events are recorded. The entry's status is AMENDED.
8. Related observations
- HSRP-020 · ESTABLISHED · PV mechanism: thalamo-amygdala precedence; Interpreter backward justificationRelation: not recorded
9. Where this appears on the site
Content pages that cite this observation inline:
- Entheogens“A second study replicated this at 88 milliseconds for backward-masked faces the subject could not consciously perceive (J. Neurosci. 43:1405, 2023).”
- Historic Warnings“The model was seriously challenged by Pessoa and Adolphs (Nature Reviews Neuroscience 11:773–783, 2010), who proposed a "multiple waves" account instead; the intracranial studies above answered that challenge directly.”
Pages citing a source this observation rests on: /collapse-of-nations, /entheogens, /historic-warnings, /neuroscience, /psycho-vulgarism.
10. Open questions
11. Machine block
The register record this page was generated from, with the derived fields added:
{
"grade": "ESTABLISHED",
"status": "AMENDED",
"category": "NEU",
"title": "Human amygdala responses to visual threat begin at approximately 74 ms, preceding visual cortex and selective to low spatial frequency; replicated at 88 ms for backward-masked stimuli outside conscious report. Behavioural threat judgement is possible from backward-masked faces at 39 ms. SUPERSEDES the widely circulated figure of 12 ms, which derives from auditory fear conditioning in RATS (LeDoux 1996; Quirk, Repa & LeDoux 1995) and is not a human value, nor a visual one. COUNTER-EVIDENCE AND SCOPE: the subcortical-route model is contested. Pessoa and Adolphs propose a \"many roads\" account and note that a visual equivalent of the rodent auditory pathway in humans is assumed rather than demonstrated. The human auditory low road itself remains newly mapped. Both human studies used epilepsy patients with implanted electrodes, who are not neurologically typical; this is the only way to obtain these latencies in humans and the constraint should be stated. All measurements used fearful FACES, not conceptual or ideological threat — the framework's extension to belief-threat is inference, not measurement. The claim that survives without the route: affective appraisal precedes deliberative evaluation. That is not in dispute.",
"entry_form": "full",
"note": null,
"n": 310,
"ref": "HSRP-310",
"cited_by": [
"SRC-428",
"SRC-429",
"SRC-430",
"SRC-431",
"SRC-550"
],
"category_name": "Neuroscience",
"supplement": "004",
"evidence": [
{
"statement": "Mendez-Bertolo, Moratti, Toledano, Lopez-Sosa, Martinez-Alvarez, Mah, Vuilleumier, Gil-Nagel & Strange, \"A fast pathway for fear in human amygdala,\" Nature Neuroscience 19:1041-1049 (2016).",
"sources": [
"SRC-429"
]
},
{
"statement": "Intracranial depth electrodes in eleven epilepsy patients. Amygdala responses began 74 ms after stimulus onset to fearful faces — but not to neutral or happy faces — at considerably shorter latency than fear responses recorded in visual cortex.",
"sources": [
"SRC-429"
]
},
{
"statement": "Two features confirm the route rather than merely the timing. The fast responses appeared only for low spatial frequency components, exactly as a magnocellular subcortical pathway predicts. And they were not evoked by arousing scenes generally: the route is tuned to socially relevant threat, not to danger in the abstract.",
"sources": [
"SRC-429"
]
},
{
"statement": "Time-frequency analysis in the second study below located a low-gamma signature from as early as 45 ms (J. Neuroscience 43:1405, 2023).",
"sources": [
"SRC-430"
]
},
{
"statement": "Second study, independent: an amygdala potential beginning 88 ms post-stimulus, preferentially evoked by fearful faces rendered invisible by backward masking, again low-spatial-frequency selective (J. Neuroscience 43:1405, 2023). The amygdala discriminated threat in stimuli the subject could not consciously perceive.",
"sources": [
"SRC-430"
]
},
{
"statement": "Rodent antecedent, correctly labelled: Quirk, Repa & LeDoux, Neuron 15:1029-1039 (1995). 74 single units across 24 freely behaving rats; conditioning effects concentrated in response components arriving under 15 ms, too early to have passed through cortex given the same laboratory's stimulation data showing cortical activation of the amygdala could not occur under 20 ms. These are rodent auditory values and are not human figures.",
"sources": [
"SRC-428"
]
}
],
"counter_evidence": [
{
"statement": "Pessoa & Adolphs, Nature Reviews Neuroscience 11:773-783 (2010) argued from physiological data that the subcortical route does not process threats faster than the cortical one, proposing a \"multiple waves\" model. de Gelder, van Honk & Tamietto replied in the same journal (2011) defending the subcortical account. The 2016 intracranial data postdates and addresses the challenge directly.",
"sources": [
"SRC-431",
"SRC-429"
]
},
{
"statement": "Latency estimates remain heterogeneous across methods. MEG studies commonly report sub-100 ms amygdala responses; several single-neuron and intracranial studies in humans and monkeys report latencies beyond 100 ms.",
"sources": []
},
{
"statement": "Sample constraint: eleven patients, all with epilepsy, all with electrodes placed for clinical rather than research reasons. Electrode placement is not experimenter-controlled and the population is not neurologically typical.",
"sources": [
"SRC-429"
]
}
],
"grade_rationale": [
"Direct intracranial recording in humans, not inference from rodent work and not reconstruction from scalp EEG. Two independent studies, two latencies, two stimulus conditions — one of them unconscious presentation.",
"The spatial-frequency selectivity is what raises this above a timing observation. A fast response alone would be consistent with several routes. Low-spatial-frequency selectivity is a positive prediction of the magnocellular subcortical pathway specifically, and it was confirmed.",
"The social-threat tuning matters for the framework's use of this finding: identity-protective cognition is triggered by challenges to standing within a group, not by danger in general. The route is tuned to exactly that class of stimulus.",
"Held at ESTABLISHED for the human precedence finding. The specific anatomical route remains CONTESTED and is graded separately."
],
"open_questions": {
"strengthen": [
"recording during identity-threatening stimuli rather than face-threat; a non-clinical population, which current methods do not permit; measurement of the interval between amygdala onset and reported subjective certainty."
],
"falsify": [
"demonstration that the 74 ms response is a volume-conducted artefact of nearby cortical activity; failure to replicate the spatial-frequency selectivity in an independent sample."
]
},
"related": [
"HSRP-020"
],
"appearances": [
"/entheogens",
"/historic-warnings"
],
"missing_fields": [
"entered",
"revised",
"intervals",
"pinpoint for SRC-428",
"pinpoint for SRC-429",
"pinpoint for SRC-430",
"pinpoint for SRC-431",
"pinpoint for SRC-550",
"source for an authored item",
"history",
"relation to HSRP-020"
]
}Generated from hsrp-research-notes.json (register generated 2026-09-19). Nothing on this page is written by hand. 11 fields not yet recorded.