"Weight gain isn't purely a measure of the calories taken in; it's also related to how those calories are perceived.
"If we can validate this in humans, perhaps we can make a drug that doesn't interfere with smell but still blocks that metabolic circuitry. That would be amazing."
What tennis grunts mean
Perhaps the most unglamorous part of top-level tennis is the grunts from players as they send balls hurling down the court.
But could they have something to do with the outcome of a match? A UK study shows grunts produced by players during tennis matches they lost were higher in pitch than during the matches they won.
What's more, psychologists at the University of Sussex found players displayed differences in their grunt pitch long before the scoreboard made it clear whether they would win or lose.
While the pitch or frequency of grunts increased as matches progressed, the study found the likely match outcome for a player may become apparent from the outset.
"This suggests that this shift in pitch is not due to short-term changes in scoreboard dominance, but, instead, may reflect longer term physiological or psychological factors that may manifest even before the match," says doctoral researcher and the university's tennis team captain, Jordan Raine.
When competitive tennis players were played short clips of other players' grunts, with no access to any other information, they were able to identify which of two grunt sequences produced by the same player came from a match the player lost.
Can a self-driving vehicle be moral?
A groundbreaking study has found that human morality can be modelled, meaning machine-based moral decisions are, in principle, possible.
German researchers used immersive virtual reality to study human behaviour in simulated road traffic scenarios.
Participants were asked to drive a car in a typical suburban neighbourhood on a foggy day and were subjected to "unexpected, unavoidable dilemma" situation involving inanimate objects, animals and humans, forcing them to decide which was to be spared.
Drawing on statistical methods, the researchers were able to show that moral decisions in these scenarios could be explained - and modelled - by a single value-of-life for every human, animal or inanimate object.
Until now, it had been assumed that such moral decisions couldn't be described algorithmically.
The German Government has just defined 20 ethical principles around self-driving vehicles based on the critical assumption that human moral behaviour couldn't be modelled.
"But we found quite the opposite," said study author Leon Sutfeld, of the Institute of Cognitive Science at the University of Osnabruck.