Contribution
A review of fifteen years of causal mapping, mostly from Berridge’s laboratory. Its own move is to separate the components of reward by their mechanisms and to prune two textbook pleasure generators, mesolimbic dopamine and the self-stimulation electrode, by reassigning them to ‘wanting’.
Key insight
Pursuit and pleasure are separable functions, and pursuit can be driven to its maximum while pleasure stays flat or falls. On the review’s account, “dissociated ‘wanting’ is merely a counterfeit pleasure or sham reward, which lacks a true ‘liking’ component” (§ Incentive salience as sham reward).
Assumptions
- Objective and subjective levels. ‘Liking’ (in quotes) is “objective hedonic reactions”, as distinct from “the subjective experience of liking” (§ A neuroscience of pleasure). The review assumes the objective reaction is the core process and that conscious pleasure is built on top of it: “Objective affective reactions likely appeared first during evolution, with subjective affective reactions following in some species, via the evolution of more elaborate and hierarchical brain mesocorticolimbic circuitry to translate core ‘liking’ reactions into conscious feelings of pleasure” (§ Evolutionary origins).
- Taste reactivity is a valid measure of hedonic impact because it is homologous across species and has “a uniquely specific relation to the hedonic evaluation of taste ‘liking’” (§ Objective hedonic reactions).
- Gain and loss of function are different evidence. A site that can amplify liking need not be needed for normal liking (§ Mapping brain pleasure generators?).
Key results
As the review reports them, from experiments not read here:
- The three components are “liking (core reactions to hedonic impact), wanting (motivation process of incentive salience), and learning (Pavlovian or instrumental associations and cognitive representations)”. In the reward cycle, “wanting processes tend to dominate the initial appetitive phase, while liking processes dominate the subsequent consummatory phase that may lead to satiety” (§ A neuroscience of pleasure).
- Taste reactivity across species. “Homologous ‘liking’ orofacial expressions can be elicited also in apes and monkeys, and even in rats and mice”. The expressions are “appropriately modulated physiologically by relevant appetite versus satiety states (Cabanac and Lafrance, 1990”. That is alliesthesia, cited to Cabanac (§ Objective hedonic reactions).
- Liking dissociates from intake. “Yet, taste reactivity ‘liking’ expressions are not diminished by such pharmacological dopamine blockade (Peciña et al., 1997), nor even by complete destruction of mesolimbic dopamine projections.” So “dopamine is not actually needed for the hedonic impact of food pleasure, but rather only for their incentive motivation value” (§ Objective hedonic reactions).
- Hotspots. “The size of hedonic hotspots mapped so far is each about 1 cubic millimeter in volume in rats”. In the accumbens hotspot, “mu opioid stimulation via agonist microinjections can at least double the hedonic impact of sucrose” (§ Causal hedonic hotspots). At the same site, “neither dopamine stimulation or glutamate AMPA blockade alter hedonic ‘liking’ for sucrose at all, even though both elevate ‘wanting’ to eat as effectively as opioid stimulation” (§ Interaction between hotspot site and neurochemical stimulation). Loss of function is concentrated in one place: “the posterior VP hotspot appears more crucial than any other known brain site for loss of hedonic function after damage, at least for taste pleasure” (§ Ventral pallidum hotspot).
- Electrodes. Berridge and Valenstein (1991): “LH stimulation failed to enhance ‘liking’ reactions to sweetness, even though it made the rats ‘want’ to eat at least four times more than normal amounts”. Heath’s patient B-19 was never quoted as saying the stimulation felt pleasant, and Schlaepfer et al.’s accumbens patients reported new desires without pleasure (§ ‘Pleasure electrodes’).
- Dopamine agonists in Parkinson’s patients. They drive compulsive gambling, shopping and similar pursuits, “Yet these cases typically do not report intense pleasure” (§ Dopamine elevations produce higher ‘wanting’ without higher ‘liking’?).
- Incentive sensitization. “Incentive-sensitization would produce addiction, by selectively magnifying cue-triggered ‘wanting’ to take the drug again, and so powerfully cause motivation even if the drug became less pleasant” (§ Resolving the cocaine puzzle?).
- Unconscious liking. “objective hedonic ‘liking’ reactions can sometimes occur alone and unconsciously in ordinary people without any subjective pleasure feeling at all”, and “Unconscious ‘liking’ reactions still effectively change goal-directed human behavior, though those changes may remain undetected or be misinterpreted even by the person who has them”. Hence “Subjective ratings are not always more accurate about hedonic impact than objective hedonic reactions, and the latter can be measured independently of the former” (§ Subjective versus objective levels).
- Wanting can be appraised as pleasure. “To the external observer, and perhaps even sometimes to the experiencing person, ‘wanting’ may be appraised as a positive reward involving eager anticipation.” So “subjective reward experience may be partly synthesized from motivation and cognitive appraisal components” (§§ Incentive salience as sham reward; Resolving the cocaine puzzle?).
- The valence of wanting is plastic. “the hedonic valence of dissociated ‘wanting’ can easily flip from positive incentive into a negative valence of anxiety, frustration or fear. Reversing the hedonic valence of the experience would not necessarily disrupt its motivating power.” Tantalus is the example, and “Motivational salience is never neutral, but its valence is not fixed” (§ Incentive salience as sham reward).
- The keyboard. Inhibition of accumbens shell releases appetite at the front and fear at the back. “Retuning can completely reverse the valence generated at a site from desire to dread, or back from dread to desire” (stressful lights and noise against a home-like room), and “Thus corticolimbic regulation adjusts both the intensity and valence of motivations produced by NAc circuitry” (§ Retuning the affective keyboard).
- A common currency, in imaging. Diverse pleasures engage “a ‘common currency’ reward network of interacting brain regions” (§ Mapping pleasure in the brain). The review takes this from correlational imaging, and contrasts it with the much smaller set of causal hotspots.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | ’Liking’, ‘wanting’ and learning have separable neural mechanisms | strong | many causal double dissociations, reviewed |
| C2 | Taste reactivity measures hedonic impact without report, homologously across mammals | moderate | homology; satiety tracking; dissociation from intake |
| C3 | Mesolimbic dopamine generates ‘wanting’, not ‘liking’ | strong (rat), moderate (human) | lesion, blockade, agonist, human rating studies |
| C4 | Most “pleasure electrodes” did not cause pleasure | moderate | re-reading of case reports; one rat experiment |
| C5 | Objective ‘liking’ can occur without felt pleasure | moderate | subliminal-stimulus studies, cited only |
| C6 | Conscious pleasure is a hierarchically later elaboration of core ‘liking’ | weak | evolutionary argument |
Concepts
- ‘liking’ — the objective hedonic reaction to a reward, whether or not it is felt.
- ‘wanting’ (incentive salience) — motivational value attributed to a cue or representation, which makes it attractive and attention-grabbing.
- hedonic hotspot / coldspot — a site where the same neurochemical stimulation amplifies (or suppresses) ‘liking’.
- incentive sensitization — dopaminergic hyper-reactivity that magnifies cue-triggered ‘wanting’, independently of ‘liking’.
- affective keyboard — the front-to-back gradient of accumbens shell generators for desire and dread, retunable by context.
Connections
The record holds Dayan and Berridge (LIT-584, NOTE-463), which defends model-based Pavlovian ‘wanting’ and names conditioned alliesthesia. Ainslie’s précis (LIT-586, NOTE-443) reads Berridge and Robinson’s wanted-not-liked behaviour as its itch-like cycle. Hesp et al. (LIT-1468) call the wanting/liking dissociation the strongest challenge to their own account of valence. Kahneman, Wakker and Sarin (LIT-1448) draw the parallel distinction between decision and experienced utility in economics.
Bearing on the record
Wanting without liking cuts against valence is potential as written. TERM-076 makes hedonic valence a regime of potential, whose intensity “is the potential’s strength”. Sensitized ‘wanting’ is potential by every mark Frijda gives control precedence (THEORY-055). The addict’s drug-ranking wins conflicts against every other concern. It persists, and it resumes after abstinence when cues return. Yet, on the review’s evidence, its ‘liking’ does not grow with it and may fall. CLAIM-440 names this defeater in advance: “intensity that does not track control precedence”, judged by marks independent of the account. Taste reactivity is such a mark. So the intensity clause fails for positive valence, which is where the claim had already hedged it, and the evidence arrives from a different direction than Carver 2003 (LIT-1445). Four ways to rescue the claim were tried, and none is clean:
- The kind clause. The term restricts valence to an ordering of the system’s own states “by their bearing on its maintenance”, and incentive salience attaches to cues and world representations. But a hungry rat’s ‘wanting’ of food is about maintenance, and dopamine manipulations raise it without raising ‘liking’. Invoked here, the clause looks like a patch.
- Sign, not intensity. The review’s own point is that dissociated ‘wanting’ is “never neutral” and can be Tantalus’s torment. On the account, Tantalus is strong potential with the system behind its course, so intense negative valence, which is right. That reading fits frustrated wanting. It does not fit the addict who gets the drug and enjoys it less: potential is strong, progress is ahead, and the account predicts intense positive valence where ‘liking’ shows little.
- ‘Liking’ is not valence. One could say the account is about motivational valence and ‘liking’ is a different quantity, hedonic impact. The review itself treats the valence of motivational salience as distinct from hedonic impact. But then the term’s word “hedonic” is the wrong word. Its object would be what the review calls the valence of motivation, and the claim would no longer be about pleasure.
- Two regimes. Within the record’s own terms, potential of an incentive ordering and potential of a maintenance ordering could be different regimes, only the second hedonic. That needs a way to tell the two orderings apart that does not use ‘liking’, and the record has none.
The candid reading is this. Potential, as defined, measures what Berridge calls ‘wanting’ better than what he calls ‘liking’. The tranche supports separating the two in the record’s vocabulary, not identifying either with “hedonic valence” by fiat.
It vindicates layer (2) and supplies a measure for it. The record’s plan measures potential from conflicts won, separately from a hedonic marker that does not depend on choice. This review is the best argument that the two must be measured separately, since they come apart under known manipulations. Taste reactivity is a non-report, non-choice marker that works across species, because “the latter can be measured independently of the former”. The marker is limited to taste and to consummatory ‘liking’. It has no counterpart for an organization, so the record’s organizational line (CLAIM-440’s organizations) has no layer-(2) measure from this work.
It partly anticipates the felt-as-self-model candidate, and also complicates it. “Unconscious ‘liking’” separates layer (2) from felt pleasure, as the candidate needs. The review’s evolutionary story, in which “hierarchical” circuitry translates core ‘liking’ into conscious feeling, has the same shape as Brouillet and Friston’s felt fluency as “a hierarchically deeper representation” (THEORY-662). But the deeper level it posits takes ‘liking’, not potential, as its object. And the claim that ‘wanting’ “may be appraised as a positive reward” even by the experiencing person implies something awkward. A self-model of one’s own potential could feel like pleasure while ‘liking’ is absent. If felt valence is the system’s model of its potential, the addict’s felt craving could be mistaken for pleasure, and the review says this can happen. That makes the candidate a theory of felt motivation, open to exactly that mistake. It is not obviously a theory of felt pleasure.
On common currency. The review’s “common currency” is an imaging finding of overlapping activation. It is not Cabanac’s behavioural trade-off (LIT-1467), and it is not the record’s ordering (CLAIM-436). It should not be cited as support for either.
Limitations
A review by the principal advocate of the view it reviews. Much of the human evidence is case reports read against the reporters’ own interpretations, as with B-19. Taste reactivity is restricted to taste. Its validity as a measure of hedonic impact rests on homology and on convergence with other measures, and those arguments are not re-made here. “Unconscious ‘liking’” in humans is cited, not shown.
Open questions
- Is there a non-taste, non-report marker of ‘liking’ that could serve as layer (2) outside the consummatory case?
- Does ‘liking’ ever prevail, winning conflicts and persisting, without ‘wanting’? The review gives no case of liking-without-wanting, and opioid hotspot stimulation raises both.