"Obese women show signs of food learning impairment," is the headline on the BBC News website.
It reports on a behavioural study involving 67 individuals of normal weight and 68 obese individuals.
Each participant was shown a series of either blue or purple squares, and then asked to predict whether the square would yield a reward. Depending on the phase of the experiment, this would be a picture of either food or money, followed by actual food or money at the end of the experiment.
The pattern of rewards was not random – one colour square was more weighted to provide a reward than another. Crucially, halfway through the experiment the reward pattern was reversed.
Researchers were interested in seeing how long it took for participants to recognise and adapt to the shift.
They found obese women were less able to recognise and adapt to the shift compared with other groups (non-obese women and men of any weight) when the rewards were food based. However, their performance was the same as other groups when the rewards involved money.
According to the researchers, the implication is that an image of food somehow skews the rational predictive and decision making part of the brain in obese women (but, strangely, not obese men).
Overall, this study will add to research in the field of behavioural learning related to food, but on its own it provides a very limited explanation or new therapeutic angles for tackling the obesity epidemic.
Where did the story come from?
The study was carried out by researchers from Yale University and School of Medicine, and Icahn School of Medicine at Mount Sinai, New York. No sources of financial support are reported.
The study was published in the peer-reviewed scientific journal Current Biology.
BBC News' reporting of the research is broadly accurate. However, the limitations of this piece of experimental research were not acknowledged – deficits in learning were seen in a very specific test scenario in a small group of people.
The LA Times provides a more informative summary of how the experiment was carried out and the alleged implications of the results.
What kind of research was this?
This was an experimental study that aimed to see if there is a difference in learning when responding to food cues, comparing obese people with those of a normal weight.
One of the main drivers of the global obesity epidemic is known to be the consumption of foods high in fat and sugar. The "rewarding" properties of these types of foods are thought to be what causes us to keep on eating them.
However, it is thought these reward circuits in the brain may differ between people, causing some to have a propensity for overeating and obesity. This is what this study aimed to investigate.
What did the research involve?
The study included 67 individuals of a normal weight (35 of these were women) and 68 obese individuals (of which 34 were women) who were recruited from the community.
They took part in a behavioural learning test assessing their reward association. The participants had to try to work out the relationship between two different coloured squares (blue and purple), and images of either food or monetary rewards.
Half the participants took part in a money task where the reward was either $5 or $10, and half took part in a food task where the reward was either 10 or 15 peanut M&Ms, or 6 or 12 pretzels (depending on the person's preference).
In the first part of the test, a picture of a reward appeared after the colour A one-third of the time, and never after colour B (total: 14 presentations of each colour, intermixed with 7 presentations where colour A was associated with the reward).
The colours were reversed in the second part of the test, so a picture of a reward appeared after the colour B one-third of the time, and never after colour A was paired with the reward in one-third of the trials, and the colour B was never paired with the rewards (total: 18 presentations of each colour, intermixed with 9 presentations, where colour B was associated with the reward).
When the participants were shown a colour, they had to indicate on a scale of one to nine the degree to which they expected to get a reward.
The researchers asked all participants to fast for four hours before taking part in the trials to try to increase the salience (importance) of the food rewards.
The participants were told at the end of the tasks that they would received the accumulated amount of all the money or food they saw during the experiment. This resulted in $100 in the money condition and a bag of peanut M&Ms or pretzels in the food condition.
The researchers examined the differences between obese people and people with normal body mass indexes (BMIs), and looked to see if there were differences between men and women.
What were the basic results?
The researchers found a significant association between test performance and BMI. The magnitude of the difference between obese and normal-weight participants was further influenced by the test modality (food or money) and whether the participant was male or female.
When looking at all participants who took part in the food tests, they found that compared to normal weight people, obese people had a food-specific learning deficit. However, when they split the group by gender, they found the association was significant only in obese women, but not in obese men.
Obese women were less able to tell which of the two colours was associated with the food on the first part of the test, or then able to switch this association on the second part of the test.
Meanwhile, on the monetary test, there was no significant learning deficit between obese and normal-weight men or women.
How did the researchers interpret the results?
The researchers say their analyses demonstrated a "robust negative association between BMI and learning performance in the food domain in female participants" – that is, as BMI goes up, learning performance goes down when food is part of the equation. The same impairment was not observed in obese men.
They say: "These findings suggest that obesity may be linked to impaired reward-based associative learning and that this impairment may be specific to the food domain."
This experimental study included fairly small groups of obese and normal-weight men and women.
It found that, overall, obese women demonstrated a learning deficit when food was used as a reward compared with normal-weight women.
On the food tasks, obese women were generally less able to discriminate between which of the two colours was associated with the food, and then to respond when the association was switched.
The difference was not significant between obese men and normal-weight men. There was also no difference among participants when money was used as a reward.
While this may possibly demonstrate some difference in food-related learning and reward associations between obese and normal-weight people – and specifically obese and normal-weight women – the applications of this single small piece of research seem to be quite limited.
The study only included a small number of people in the US: 67 normal-weight individuals and 68 obese individuals. These people where then divided between the two monetary and food tasks.
This meant all of the results related to the "food-related learning deficit in obese women" were obtained from tests in only 18 obese and 18 normal-weight women.
This is a very small group, and it is possible the results could be down to chance. Other groups of people, including those of different countries and cultures, could have given different results.
This was also only a single very specific test, seeing if people could spot which of two colours was associated with the food reward of either some M&Ms or pretzels. Interpreting the meaning from this single test is very difficult. It tells us very little about how people become obese.
For example, someone being unable to link which particular colour is associated with a food item doesn't tell us about the various drivers that have led to that individual becoming obese.
Even if we do take the findings of the study at face value, it leaves several important questions unanswered.
For example, does "food-related learning deficit" lead to obesity, or does being obese make it more likely you will go on to develop a food-related learning deficit?
And why were these deficits only seen in obese women and not obese men?
One possible answer to the second question is this could be because of the very small samples of men and women tested. The findings of no difference in men but a difference in women could be purely down to chance, and there may not be a difference between the genders at all.