Tempting fast food and sugary treats are among the culprits in the obesity epidemic currently sweeping much of the world. Researchers are intensively trying to understand what makes people choose some types of food over others. Experiments to determine this often depend on self-reports. Now researchers have developed a method that enlists eye tracking and facial expressions to help them identify which types of food turn us on – even before we become aware of it.
Searching for food is no longer a daily battle for most people in affluent countries. Many people struggle not to consume too much – especially energy-dense and hyperpalatable food. Our senses are still programmed to seek these types of food, and this has contributed to increasing obesity. Studies of the mental and physiological responses to food are therefore important for understanding and preventing obesity and the associated comorbid conditions.
“What people choose to eat depends on many factors, including how hungry they feel, what they know about food and their desire for sensory pleasure. So far, measuring the behavioural and psychological aspects has been challenging since it has been based primarily on self-reports, but now we have developed a method that uses eye tracking, electrodermal activity and facial expressions to objectively assess a person’s immediate reactions to food cues. In particular, visual attention towards foods differs substantially according to the fat content and taste,” explains Hanne Pedersen, industrial PhD student from Steno Diabetes Center Copenhagen.
Much deeper insight
Many people think that eating at an ad libitum buffet is difficult to manage, and the researchers therefore used such a buffet in their experiment to monitor the reactions of the participants and how long they looked at images of foods. The researchers used biometric methods such as eye tracking to measure where and for how long the participants looked at various places on a computer screen. Electrodermal sensors and facial recognition cameras were also used for measuring reactions.
“We used machine learning algorithms to code the facial expressions in videos into emotional expressions. These were used to examine how participants’ facial reactions differed to foods that varied in fat content and taste. We also examined how these biometric signatures for foods were related to which foods these individuals ended up choosing from the ad libitum buffet,” says Hanne Pedersen.
The new method is based on an existing and highly recognised method: the Leeds Food Preference Questionnaire, which was developed to examine food reward by collecting ratings, choices and reaction times in response to images of food from various food categories.
“However, the results leave many unanswered questions about why we react the way we do and what psychological and physiological processes are activated when we react to various types of food. We therefore decided, in collaboration with the leading researcher behind the Leeds Food Preference Questionnaire, Graham Finlayson, and iMotions A/S in Denmark, to add biometric measurements to the existing methods. That way we can now obtain much deeper insight into the immediate and unconscious reactions to foods,” explains Hanne Pedersen.
Improved ability to measure health interventions
The researchers initially tested the new method on 100 normal-weight adults to establish a basis for comparison for future studies. However, this has already given the researchers exciting new insight into the interaction between subjective assessment, objective biometric reactions and the actual food intake from the ad libitum buffet. All this behaviour is governed by complex interactions between psychological and physiological processes that function both consciously and unconsciously.
“Our eyes can interpret information about the fat and energy content of food. The study showed that the participants’ visual attention to certain types of foods was associated with how much they ate of them and how well they actually liked them. Finally, we are all endowed with knowledge of health and especially a set of cultural norms about what we should eat. The interaction between all these impulses makes us choose one food over another in an ad libitum buffet,” says Hanne Pedersen.
This knowledge can be absolutely crucial in helping people to choose healthier solutions – not only in ad libitum buffets but generally in a society that bombards us with advertisements and offers that encourage us to make choices that more easily cause weight gain and health problems.
“Now we have a method we can use to start studying how choices and reactions differ in different target groups – for example between people with different weight levels. This may enable us to understand the interaction between psychological and physiological mechanisms and how some people are more susceptible to an environment that promotes obesity than other people. Now that we have tools to measure more parameters, we can also measure how these parameters change collectively or separately when we try to get people to make lifestyle changes, so we know what works best for each individual,” concludes Hanne Pedersen.