The insurance hypothesis proposed by Nettle et al. is primarily presented as a psychological hypothesis. Yet, energy intake, hunger, and obesity are not just psychological, but are also physiological processes. Including a consideration of these physiological processes – and, in particular, the cues affecting them – can help expand the insurance hypothesis to relate food insecurity to obesity, even when food is secure, and to incorporate other causes of obesity, including stress, disease, dietary micronutrient composition, and sedentism.

Of these, stress bears the closest relationship to conscious perceptions of food insecurity. Activation of the hypothalamic-pituitary axis and the release of glucocorticoids can induce both increased energy expenditure and increased energy intake, particularly when stress is chronic (Tataranni et al. Reference Tataranni, Larson, Snitker, Young, Flatt and Ravussin1996; Torres & Nowson Reference Torres and Nowson2007). Through much of human history, glucocorticoids would have served as reliable cues of both an increase in energy need and an increase in energy unpredictability. Even psychosocial stress would have been related to changes in social food-sharing networks and thus resource buffering and food security (Gurven et al. Reference Gurven, Jaeggi, von Rueden, Hooper and Kaplan2015). Among subsistence populations, food anxiety, psychological well-being, and actual caloric productivity are closely related (Stieglitz et al. Reference Stieglitz, Jaeggi, Blackwell, Trumble, Gurven, Kaplan, Weinstein and Lane2014). However, for humans in post-industrial countries, there may often be little linkage between HPA activation and food unpredictability; stressors lead to physiological increases in energy uptake that are not accompanied by changes in expenditure and unpredictability, thus causing the accumulation of excess weight.

A second cue worth considering is immune activation. Immune function is energetically costly, thus frequent immune activation might be expected to serve as a cue signaling greater needs for preparatory energy storage, particularly when coupled with resource unpredictability. In contexts in which disease is actually more prevalent, these excess calories might be used (Blackwell et al. Reference Blackwell, Trumble, Maldonado Suarez, Stieglitz, Beheim, Snodgrass, Kaplan and Gurven2016; Gurven et al. Reference Gurven, Trumble, Stieglitz, Yetish, Cummings, Blackwell, Beheim, Kaplan and Pontzer2016). However, if disease cues are not actually linked to disease, then excess energy might again be accumulated. This might be the case if disease is present at early ages but reduced at older ages, similar to other formulations of the thrifty phenotype hypothesis.

Immune activation may also affect intake by increasing the need for particular micronutrients (Cotter et al. Reference Cotter, Simpson, Raubenheimer and Wilson2011). Organisms do their best to optimize the micronutrient composition of their diets, but when micronutrients are limited, excess consumption can be necessary to meet micronutrient demands (Simpson et al. Reference Simpson, Sibly, Lee, Behmer and Raubenheimer2004). Internal processes likely monitor micronutrients, allowing food content to cue micronutrient unpredictability, even in the absence of caloric unpredictability. Such cues may also affect perceived unpredictability. For example, the nutritional content of foods affects both energy intake and the experiences of hunger (Fuhrman et al. Reference Fuhrman, Sarter, Glaser and Acocella2010). Changes in the food supply with industrialization mean that foods are increasingly put together from constituent components, rather than prepared in ways that reflect natural associations between micronutrients present in animal and plant material (Cordain et al. Reference Cordain, Eaton, Sebastian, Mann, Lindeberg, Watkins, O'Keefe and Brand-Miller2005; Pollan Reference Pollan2006). Thus, physiological mechanisms that evolved to balance micronutrients might have a hard time motivating correct intake when the content of foods does not resemble the content of foods in the past.

Finally, physical activity may also serve as a physiological cue, not only to expected energy expenditures, but also to the expected costs of carrying excess weight. Carrying excess weight is only functionally costly when an organism must move to escape predators or acquire food. Through most of human history, mobility would have been critical for survival. Yet, the degree and type of mobility required would have varied between individuals. Thus, we might expect individuals to monitor how much physical activity they engage in and use this as a cue to the future need for activity and thus the expected costs of carrying weight. If such a mechanism exists, sedentism might lead to excess weight gain above and beyond changes in activity energy expenditure, by also affecting things like the rate of adipose deposition. For example, high-density lipoproteins (HDLs) may be part of a mechanism for regulating fat deposition in relation to physical activity, as HDLs are increased by physical activity (Warburton et al. Reference Warburton, Nicol and Bredin2006), and they also reduce the deposition of body fat by affecting lipolysis (Wei et al. Reference Wei, Averill, McMillen, Dastvan, Mitra, Subramanian, Tang, Chait and Leboeuf2014).

In the mathematical model described by Nettle et al., individuals have perfect knowledge about food security in their environments. Yet, in real life, knowledge is imperfect and might be particularly imperfect for stochastic food fluctuations. Thus, humans must regulate their intake based on inexact cues to multiple variables, including expected energy needs, expected unpredictability of food, and the expected costs of carrying extra weight. A consideration of inexact cues and the systems that interpret them is important because even though a general theory like the insurance hypothesis can describe why we see associations between food insecurity and obesity, it says little about why we see humans exceeding optimal levels of fatness.

Thinking about cues can also help us make clear predictions leading to novel interventions against obesity. Many of the cues discussed here may misfire in modern environments, leading to feelings of food insecurity in the absence of actual caloric insecurity, or conversely, affecting behavior without changing conscious assessments of food insecurity. For example, stress might be associated with obesity only in contexts where it affects perceived food security, but not actual fluctuations in food availability. Similarly, poverty and food insecurity lead people to choose high-calorie, low-micronutrient foods, but we might also predict that the micronutrient density of a person's diet could influence perceived food insecurity, creating positive feedback and exacerbating weight gain.

A complete explanation for the obesity epidemic should explain not only why food insecurity is associated with increased weight, but also why it is associated with excess weight – that is, why these mechanisms overshoot what might be expected to be adaptive levels. In short, thinking about the cues associated with expected energy needs, expected unpredictability of food, and expected costs of carrying extra weight can help link the insurance hypothesis to other theories of obesity.