Please note, this is recovered content from the former website of the New York Obesity Research Center website.


Direct and Indirect Influencers of Meal Size: The First 10 Years

Date: March 8th, 2007
Speaker’s Name and Affiliation: Gerard P. Smith, Cornell-Weill Medical School
Title: “Direct and Indirect Controls of Meal Size: The First 10 Years”
Physiological, psychological, ecological, conditioned, social, and pathological states, stimuli, and molecular mechanisms increase or decrease the size of meals. There was no theory for analyzing the relationships and relative potency of these various states, stimuli, and mechanisms until1996 when Smith proposed “The Direct and Indirect Controls of Meal Size” (Smith, 1996). This theory was unambiguous, neurologically coherent, and more comprehensive than any previous theory, such as the glucostatic theory of Mayer or the gastric contraction theory of Cannon. In this seminar, Smith described the main components of his theory and reviewed the experimental evidence for it that has accrued in the ten years since it was proposed.

A control of meal size is composed of an adequate stimulus, specific receptor, neural and molecular mediators, and brain processing that involves a comparator function and a motor command to the final common motor pathway of central pattern generators in the hindbrain for oromotor movements of eating, such as licking, lapping, or chewing.

Direct controls are stimulated during a meal by ingested food stimuli or their digestive products acting on preabsorptive enteral receptors in the mucosal surfaces of the mouth, stomach, and small intestine. A comparison of meal size of the same diet during sham feeding and real feeding demonstrated that orosensory direct controls provide primarily positive feedback to the brain that acts to maintain eating and postingestive direct controls provide primarily negative feedback to the brain that acts to terminate eating.

When the processed central potency of enteroreceptive positive feedback is judged larger than the central potency of enteroreceptive negative feedback by the central comparator function of the brain, eating continues. When the processed central potency of enteroreceptive negative feedback is equal to or larger than the processed central potency of enteroceptive positive feedback, eating stops and the meal ends. The disconnected caudal brainstem in the chronic decerebrate rat has sufficient neural complexity to process, compare, and integrate the peripheral positive and negative feedbacks produced by enteroreceptive stimulation into a meal.

Indirect controls are distinguished from direct controls by two criteria: 1) they do not stimulate enteroreceptors directly; 2) they do not act in the chronic decerebrate rat. The second criterion means that the size of a meal in the chronic decerebrate rat is entirely determined by direct controls and that indirect controls are dependent on neural connections between the forebrain and caudal brainstem.

Note that an animal or human must be eating in order for the effect of an indirect control on meal size to be measured. This means that direct controls stimulated by the ingested diet will also be acting. Thus, if an indirect control, such as food deprivation for 24 hours or the ovarian rhythm of estrogen, increases or decreases meal size, it does so by changing the central potency of positive or negative feedback produced by enteroreceptive stimulation. The change in central potency produced by the indirect control can be quantitated by curve-shift analysis (Smith, 1996).

Identifying the specific effect of any indirect control of meal size by measuring its effect on the central processing of peripheral positive or negative feedback not only advances the analysis of the controls of meal size beyond the descriptive statements of increased or decreased meal size, it also suggests the cellular and synaptic mechanisms involved because the mechanisms of the two feedbacks are different.

Research during the past ten years has supported the theory. Leptin and insulin, the peripheral hormonal mediators of the negative feedback control of meal size produced by adipose mass, decrease meal size by increasing the central potency of cholecystokinin (CCK), a mediator of the peripheral negative feedback control of the small intestine stimulated by digestive products of ingested food, particularly fatty acids. Estrogen, the hormonal mediator of the ovarian control that decreases meal size, also decreases meal size by increasing the central potency of CCK.

A third example is Neuropeptide Y, a central peptide mediator of increased meal size produced by the indirect controls of food deprivation or diabetes mellitus, increases meal size by increasing the central potency of peripheral enteroreceptive positive feedback.

In addition to this work in the adult rat, recent review of the ontogeny of independent ingestion in the rat and mouse (Smith, 2006) supported the theory’s classification of direct and indirect controls and their interaction and measurement. The identification of direct and indirect controls operating in the preweaning rodent during its first independently ingested meal is strong evidence that the direct-indirect classification is not simply a new name for an old descriptive distinction, such as short-term and long-term, but accurately maps the biological organization of the controls of meal size revealed by the natural experiment of development.

Smith concluded that the past 10 years of research produced encouraging confirmations of the theory and no significant challenges to it.


Q. Data that were collected from real and sham feeding studies done by Campbell & Davis and Gibbs, Smith, and Young in 1973 are not inconsistent with findings originally proposed in the glucostatic theory, are they?

A. The glucostatic hypothesis did not provide a verified explanation for the size of a meal under any condition.

Q. Do you have a specific definition for a “direct” control of meal intake?

A. The definition has three parts. First, a direct control is stimulated by ingested food or digestive products during a meal that act directly on specific receptors in the preabsorptive surfaces of the gut from the mouth to the end of the small intestine. Second, a direct control is mediated by orosensory, gastric, and small intestinal sensory afferents that go directly to the hindbrain. The afferent nerves are stimulated by ingested stimuli or by peptides released by the ingested stimuli. Third, the caudal brainstem disconnected from the forebrain has sufficient neural complexity to integrate the positive and negative feedbacks of the direct controls of meal size when chronic decerebrate rats eat liquids infused intraorally.

Comment: That seems like a circular definition.

A. All definitions are logically circular.

Q. What about intermeal interval, couldn’t you also say that your definition of a direct control also applies to that?

A. Direct controls are controls of meal size that function during the ingestion of a meal and have the three criteria I just mentioned. These criteria do not apply to intermeal interval. The controls of intermeal interval have received little attention although Woods recently suggested that conditioning was a major mechanism for them and Gibbs has suggested that gut peptides prolong the intermeal interval.

Q. Why is adiposity an indirect control. Can’t you measure it directly?

A. You can measure adiposity directly as well as the two major adiposity signals, leptin and insulin. None of these measurements satisfy the criteria of direct controls that I listed above. Thus,adiposity is an indirect control and leptin and insulin mediate its effect on the central neural network that controls eating. The inhibitory effect of leptin and insulin on meal size can only be measured during eating which stimulates direct controls. When leptin and insulin decrease meal size, it has been demonstrated that they do it by increasing the central potency of the negative feedback effect of CCK which mediates the direct inhibitory control activated by food stimuli in the proximal small intestine.

Q. What is a direct measure of a direct control. If you have sucrose concentrations, would that be an example of a direct measure?

A. Yes, sucrose concentration is an adequate stimulus for orosensory positive feedback. Geary and I demonstrated this in 1985.

Q. Is it correct to say that CCK works on indirect controls, but leptin does not?

A. No, that is not correct. Indirect controls related to adiposity and estrogen rhythm decrease meal size by increasing the central potency of the negative feedback effect of CCK which is a mechanism of the direct inhibitory control from the proximal small intestine. Leptin from adipose tissue, on the other hand, is a mechanism of the indirect inhibitory control of meal size exerted by the mass of adipose tissue.

Q. In humans, we put some value on reports of hunger and fullness, and how people are feeling during the process of eating. If leptin affects feelings of hunger and fullness, can’t we then say it has an effect on meal size?

A. Subjective reports of hunger and fullness do not reliably predict meal size. When leptin decreases meal size, it does so by increasing the central potency of the direct inhibitory control of meal size produced by food stimuli in the stomach and small intestine. The quantitative relationships among plasma leptin concentration, decrease of meal size, and self reports of hunger and fullness are unknown.

Q. What do you mean by an increased positive feedback?

A. By increased orosensory positive feedback, I mean that the concentration of adequate stimuli in the mouth, such as sucrose concentration, increased and this increased the stimulation of afferent nerves from the mouth to the nucleus tractus solitarius in the hindbrain. Another way in which orosensory positive feedback can be increased is by the release of an orexic peptide or amine in the central network that processes the orosensory input so that its central potency is increased. An example of this is our recent report that NPY in the third ventricle increased meal size of a sucrose solution by increasing positive orosensory feedback without changing postingestive negative feedback (Torregrossa et al., 2006).

Q. Are you suggesting that estrogen acts on food intake through CCK and GRP?

A. Yes. Over the past decade, Geary and his colleagues established this for CCK. There is much less information about GRP and glucagon.

Q. Does the theory tell us which pathway would be affected by one signal or the other?

A. By requiring you to demonstrate whether a signal that changed meal size did it by changing positive or negative feedback, the theory leads you to a different group of candidate mechanisms, because most of the central and peripheral mechanisms of positive feedback are different from the mechanisms of negative feedback.

Q. In terms of the negative feedback that occurs when animals are sham feeding, wouldn’t that be an internal feedback that modulates these signals?

A. Davis and I have shown that this negative feedback is conditioned and the unconditioned postingestive stimulus is small intestinal hypertonicity.

Q. Direct controls are acting through indirect controls. There are several sensory interfaces where dopamine affects food intake, but this is very far removed from the hindbrain. How does your theory take into account that some animals may be eating specifically to increase levels of dopamine in the nucleus accumbens?

A. No, indirect controls affect meal size by modulating the central potency of orosensory positive feedback or postingestive negative feedback of direct controls. The possibility that under some conditions, rats eat to increase levels of dopamine in the nucleus accumbens is intriguing, but not proven. If this were shown to be a physiological fact, the theory would encompass it within orosensory positive feedback because Hajnal and Norgren and their colleagues have shown that dopamine levels in the accumbens shell are a function of orosensory positive feedback stimulated by sucrose or corn oil.

Q. What about cephalic phase learning. For example, the animal learns to have an increased cephalic phase response that will affect the amount of negative feedback on a meal. Is this a case where a typical indirect control would be acting as a direct control?

A. This is a case where conditioning a visceral response decreased the adequate intestinal stimulus for direct inhibitory control of meal size. No indirect control of meal size is involved.

Suggested Readings:

  1. G.P. Smith. The direct and indirect controls of meal size. Neurosci. Biobehav. Rev. 20: 41-46, 1996.
  2. G. P. Smith. The controls of eating: brain meanings of food stimuli. Prog. Brain. Res. 122: 173-186, 2000.
  3. G.P. Smith. The controls of eating: A shift from nutritional homeostasis to behavioral neuroscience. Nutrition 16: 814-820, 2000.
  4. G.P. Smith. Controls of Food Intake. In: M.E. Shils, M. Shike, A.C. Ross, B. Cabllero, R.L. Weinsier, and R.J. Cousins, (Eds.), Modern Nutrition in Health and Disease, 10th Edition. Baltimore: Lippincott Williams & Wilkins, 2005, 707-719.
  5. G.P. Smith and G. Dockray. Introduction to the reviews on appetite.Philos Trans R Soc Lond B Biol Sci. 361(1471):1089-93, 2006.
  6. G.P. Smith. Ontogeny of ingestive behavior. Dev Psychobiol. Jul;48(5):345-359, 2006.
  7. A.-M. Torregrossa, J.D. Davis, and G.P. Smith. Orosensory stimulation is sufficient and postingestive negative feedback is not necessary for neuropeptide Y to increase sucrose intake. Physiol. Behav. 87: 773-780, 2006.