What the Dallman model proposes about food as stress relief — and what it does not establish.
What the evidence says
- Calorie-dense foods activate reward pathways[1]
- …and the resulting fat stores are PROPOSED to send signals that dampen HPA axis reactivity — this is the model’s feedback half, and it is rat work[1]
- The authors PROPOSE this reduces the stress response; they did not demonstrate it in people. Their sentence: “We propose that people eat comfort food in an attempt to reduce the activity in the chronic stress-response network with its attendant anxiety”[1]
- “Self-medication” with food being biologically effective short-term is the model’s claim, not a measured human finding. In humans chronic stress runs BOTH ways — increased intake with weight gain, OR decreased intake with weight loss[1]
- Eating to regulate emotions is one of Macht’s five classes, and the review holds that ”emotions may regulate eating, and eating may regulate emotions”[2]
- Long-term cost: weight gain, metabolic dysfunction[1]
- One human study enrolled 59 healthy premenopausal women aged 20–50, but compared only the extreme quartiles of perceived stress — 17 high against 16 low. The grouping was by perceived stress score, not by caregiver status: the sample was recruited to span a wide stress range, containing caregivers of chronically ill children (n = 40) and of healthy children (n = 19). High-stress women reported more emotional eating (3.16 vs 2.18 on a 5-point scale), had greater central adiposity (sagittal diameter 20.92 vs 18.24 cm; BMI 25.97 vs 23.89), and showed significantly lower cortisol output in response to a lab stressor despite reporting greater distress to that same stressor[3]
- In the high-stress women, abdominal fat correlated negatively with diurnal cortisol (r = −.44, p = .05) and with greater dexamethasone suppression (r = −.55, p = .02), among what the authors call ”signs of a heightened sensitivity to cortisol” — the pattern the Dallman model predicts, observed in people. The authors report that these relationships ”did not emerge in the low stress group”, which is 16 women: a null in a small group, not a demonstration that the pattern is confined to high stress[3]
- And it is cross-sectional, which the authors state as a limitation. It cannot say whether stress eating precedes the hormonal changes or follows them, and the sample is small. So this is the model’s shape found in humans, not the model’s loop demonstrated in humans — the feedback direction remains proposed[3]