[evol262]

There are essentially zero papers which show anything even remotely resembling this. In basically every controlled environment, CICO more or less works, albeit not with perfect efficiency. It's not until prolonged starvation periods or massive overfeeding periods that any sort of genetic/epigenetic limiters come in (overfeeding tends to be prisoner studies, the MN starvation experiment is the most commonly cited/most data from long-term starvation).

If dieting works in controlled environments, the obvious conclusion is that dieting does not work in uncontrolled environment precisely because of the poor food choices, and because we, as humans, tend to discount the amount of calories in small things through out the day, nor do we rigorously measure.

All a "NEAT decrease" is saying is "people are more lethargic when dieting". It would be unsurprising if this were due to crash dieting, but this cannot "negate" a diet. Reduce calories further until the scale starts to move. Or be more active, despite lethargy.