Background somatic states
How the somatic-marker-hypothesis handles mood: not as a separate system, but as the noise floor against which every triggered somatic signal is read. From Bechara & Damasio (2005), §4.1–4.2.
The claim
Somatic markers are not evaluated in a vacuum. Pre-existing states — from market news, from a streak of wins or losses, from being in pain or hungry — modulate both the triggering and the feeling of subsequent states. “Prior emotional events influence future economic choices.”
The circuitry is therefore a complete circle, in the authors’ phrase: primary and secondary inducers trigger somatic states → feedback from those states modulates activity in the structures critical for induction (amygdala, VM cortex, insula/SII/SI) → that modulated activity changes the next round of induction.
Two mechanisms
At the brain — before the next state is triggered. Pre-existing states influence neurotransmitter release (dopamine is the example), which lowers the threshold of neuronal firing in insula/SII/SI, amygdala, and VM cortex. So after a streak of losses the thought of another loss becomes more painful and triggers a stronger negative state; after a streak of gains the thought of another gain becomes more pleasurable.
At the soma — after a state has been triggered. Pre-existing states shape the feedback signals the new state can generate. The worked example: if a negative state means high heart rate and a positive state low heart rate, then in a negative state keeping the heart rate high “is physiologically easier to achieve,” while switching high→low is harder. Negative states are reinforced; positive ones impeded.
Both converge on the same result — changing the magnitude of the somatic feedback that biases feeling and decision. Hence: “negative states breed pessimism and positive states breed optimism,” and once in one it is harder to switch to the other.
The signal-to-noise model (Fig. 10)
The framework’s formalization, such as it is. The triggered somatic state is signal; the background state is noise; what reaches the brain depends on their relationship.
| background | signal-to-noise | result | market analogue |
|---|---|---|---|
| neutral or weak | high (“big splash”) | strong, clean somatic feedback; decisions most sensitive to long-term consequences | stable, uneventful market |
| strong, incongruous | low (“no splash”) | triggered signal cancelled by background noise; the state is ineffective | crash: thoughts of future gain stop biasing |
| strong, congruous | signal in phase with noise (“strong splash”) | somatic feedback exaggerated | crash: thoughts of further loss dominate |
So in a crashing market, thoughts signalling another loss gain control over behaviour while thoughts signalling future gain become ineffective; in a growing market, the reverse. Note the asymmetry this predicts and the wiki should hold onto: a strong background does not simply amplify or simply mute — it acts as a filter tuned to its own valence.
The appetite/bias distinction
The subtlest and most valuable thing in the section, and easy to read past.
After a streak of losses, a greater appetite for gain develops. This looks like it should contradict the model — if a negative background suppresses positive somatic states, why does wanting-to-win go up? Bechara & Damasio’s answer is that appetite and bias are different quantities, and appetite is generated by the aversive state itself.
The analogy: a hungry person has a greater appetite for food, but that appetite comes from the hunger. Thoughts of food exacerbate the hunger state rather than alleviating it. During hunger, decisions are desensitized to the biasing influence of positive somatic states — “promises of delicious food and great restaurants do not bias a hungry person towards waiting.” The behaviour is energized by the aversive state, not by the positive states of the food options.
Transposed: in a crashing market, the investor’s increased appetite for gain reflects a drive to escape the state of loss, not a bias toward good options. So the investor is biased toward stopping the ongoing loss — selling — rather than toward choosing promising stocks.
This is a genuine, non-obvious, and in-principle testable prediction, and it dissolves the apparent paradox without hand-waving. It is also where the model does work that prospect theory does not.
The prediction that was never tested here
The framework’s one advance beyond re-describing prospect theory. Since sure outcomes are processed by posterior VM cortices and therefore trigger disproportionately stronger somatic states than probable ones, the model predicts that framing effects are modulated by background state rather than fixed:
- Risk seeking in the face of sure loss is enhanced when the background state is negative. After several losses, aversion to another sure loss increases disproportionately, so seeking risky alternatives increases.
- Risk aversion in the face of sure gain is enhanced when the background state is positive. After several gains, desire for another sure gain increases, so risk-seeking decreases.
The paper offers no test. Its supporting evidence is indirect: Loewenstein et al.’s (2001) hot/cold states — people in pain over-buy pain medicine, hungry people over-buy food — read as congruous background exaggerating a congruent triggered state. That fits, but it is other people’s data recruited after the fact, not a discriminating test.
Where this touches the rest of the wiki
It presupposes autonomic-specificity. The signal-to-noise model only works if the brain can tell positive from negative somatic signals — otherwise “congruous” and “incongruous” are undefined. Bechara & Damasio assert the discrimination (§4.1.1: positive and negative states “induce distinct physiological patterns… changes in heart rate, skin conductance, respiration”) but cite Cacioppo et al. (2000) for it rather than any Iowa data, and their own measure is skin conductance alone, which cannot distinguish patterns. So the model’s load-bearing assumption is imported from the very literature autonomic-specificity-of-emotion records as unsettled. This is worth flagging as a dependency, not a refutation: valence discrimination is a much weaker requirement than emotion-specific patterning, and is the part of the specificity literature that has held up best.
It is a mood model without a mood construct. Background somatic states do the work core affect does in Barrett’s framework — a non-emotion-specific bodily-valence state that colours what follows. The two are never connected in either literature. The instructive difference: for Barrett core affect is primitive, the raw material out of which emotion is constructed; for Bechara & Damasio the background state is residue, left over from previous inducers. Same position in the architecture, opposite provenance.
It is allostatic in spirit and never says so. The circle — states change thresholds change states — is regulation-through-change rather than around a set point. See allostasis. The connection is the wiki’s, not the paper’s.