Obesity: can environmental food odours make you lose control? Investigation of implicit priming effects on reactivity and inhibitory control towards foods

The food environment can interact with cognitive processing and influence eating behaviour. Our objective was to characterize the impact of implicit olfactory priming on inhibitory control towards food, in groups with different weight status. Ninety-two adults completed the Food Inhibition Task: they had to detect target stimuli and ignore distractor stimuli while primed with non-attentively perceived odours. We measured reactivity and inhibitory control towards food pictures. In all participants, food pictures were detected more quickly and induced more disinhibition than neutral pictures. Only individuals with obesity were slower to detect foods when primed with a high energy-dense food odour than in control conditions. Common mechanisms were observed for the top-down processing of foods, regardless of weight status, but we observed specific priming effects related to weight status on bottom-up processes. Our results contribute to current knowledge regarding the relationship between cognitive load and food reactivity in an obesogenic environment.


INTRODUCTION
Studies have shown that individuals with obesity tend to have poorer inhibition capacities when it comes to food 37 (1,2). In our food-abundant environment, this tendency inevitably leads to overeating, i.e. eating more than one's 38 physiological needs. This type of impaired inhibition can naturally lead to weight gain and even to obesity.

39
Environmental factors and bottom-up cognitive processing of foods. 40 The combination of excess calorie intake and a lack of caloric expenditure results in weight excess, overweight, 41 and often obesity. This phenomenon is related to our environment: for most people in modern day society, food 42 is abundant and easily accessible. Moreover, daily exercise is now a choice rather than an obligation. Scientists 43 have therefore introduced the idea of the "obesogenic" environment, inferring that the influence of the in a specific type of stimulus processing, which helps individuals to adapt to changing situations by enabling 83 voluntary behaviours and inhibiting possible perturbations.

84
The hypothesis of a deficit in inhibitory control among individuals with obesity has been widely explored by 85 researchers in an effort to explain why weight loss remains difficult, and to find innovative opportunities to 86 reduce obesity (15). Such a deficit could lead to a decrease in the ability to pursue goal-directed behaviour, such 87 as maintaining a healthy lifestyle. In this line of study, some authors showed that individuals with obesity have 88 lower inhibitory control, (2,12,16,17) while other studies found no differences related to weight status (18, 19). 89 No consensus has been found so far, potentially due to the diversity of methodologies (20). Additionally, other 90 variables (such as frequent comorbidities in obesity, or specific eating styles) are susceptible to modulate 91 inhibitory control capacities beyond weight status (19,(21)(22)(23) found that non-attentively-primed olfactory HED food cues led individuals with obesity to direct their attention 100 more toward foods (9). These observations led us to question whether olfactory priming could facilitate a deficit 101 of inhibitory control toward foods. Previously, we demonstrated the differentiated effects of non-attentively 102 perceived food cues on attentional biases (implicit measure of a bottom-up process) depending on weight status; 103 here we aimed to measure the same effects on inhibitory control toward foods. To our knowledge, our study is 104 the first to explore the relationship between a context of facilitation and inhibitory control toward foods (high 105 and low energy-dense foods vs. neutral non-food stimuli) in male and female adults of various weight statuses 106 (normal-weight, overweight, obese) and with no eating disorder.

107
The first aim of this study was to characterize inhibitory control toward food pictures in individuals with normal-108 weight, overweight and obesity. Our second aim was to study how olfactory priming affected top-down 109 processes in individuals with various weight statuses, by measuring their inhibitory control capacities when non-110 attentively exposed to olfactory food cues compared to non-exposed. Our main hypothesis was that, compared with neutral stimuli (objects), individuals facing food stimuli would have decreased inhibitory control, especially 112 when the food stimuli were HED. We expected that this deficit would be increased in individuals with higher 113 weight status, especially when non-attentively primed with olfactory food cues.

182
The Go/no-Go subtest explores response inhibition through a simple task in which the participant must detect 183 target stimuli "X" and withhold a response when presented with distractor stimuli "+". The flexibility subtest 184 assesses shifting abilities in mental flexibility. In this subtest, two stimuli appear, one on the left and one on the 185 right side of the screen. One of the stimuli is round while the other is an angular shape.

Weight status
Normal-weight (NW) Though we hypothesized that individuals with higher weight status would show less inhibitory control toward 385 foods than lean individuals, it was not the case in our experiment. In fact, we found common patterns of 386 inhibitory control toward food stimuli in individuals across the weight status spectrum.

387
In our experiment, participants made more commission errors when they were facing HED food stimuli. No 388 difference was found in regard to weight status, which is congruent with part of the literature (63,64). This 389 observation strongly suggests that the lack of inhibition toward foods is a common process for all individuals and 390 it is also consistent with the idea that the rewarding quality of HED foods makes them more appealing (65-67), 391 leading to an increased approach bias. The saliency of HED foods combined with the associated approach bias 392 makes the detection of HED food stimuli a prepotent response for the individual. A prepotent response is 393 cognitively more difficult to inhibit than other response options, which need to be inhibited in order to exhibit 394 goal-congruent behaviour. This effect appears to be even stronger when cognitive load is high because 395 individuals make significantly more commission errors toward HED food stimuli in this condition.

396
We found different patterns of inhibitory control toward HED and LED foods, indicating that the top-down 397 processing of those stimuli is differentiated. In lower cognitive load conditions, individuals made fewer 398 commission errors when facing LED food stimuli than when facing HED food or object stimuli. We can thus 399 presume that fruits (LED foods) are processed faster than other stimuli. This assumption is supported by the 400 work of Leleu et al., 2016 (68), who showed that fruit pictures elicited earlier event-related responses in the brain 401 than other food types (vegetables, HED foods) during a food discrimination task.

402
Food stimuli are salient, which induces an approach bias that interferes with the initiation of goal-directed 403 behaviour on a cognitive level, leading to cognitive and behavioural deficits in inhibitory control. This process 404 occurs in individuals regardless of weight status, and its intensity seems to vary in function of food 405 characteristics (i.e. category and/or energy density). Moreover, the deficit in inhibitory control induced by food 406 stimuli is modulated by the cognitive load in working memory, which means that the more mental effort the 407 individual has to make while performing a task, the fewer resources are available to inhibit prepotent responses.

408
This phenomenon leads to more disinhibition, meaning that individuals may be more likely to eat more HED 409 foods when their cognitive load is heavier.

Priming effects: why does implicit priming only impact bottom-up processes?
411 In our study, we tested whether implicit priming with olfactory food cues would impact inhibitory control, a information. Another type of less subtle but still implicitly perceived cues should be tested in order to observe 430 the effects we were expecting in this study.

432
Concerning global reaction times, we found some priming effects for individuals with overweight and obesity.

433
More specifically, individuals with obesity and with overweight were slower to detect all kinds of stimuli when primed with a pound cake odour and a pear odour, respectively, regardless of the go/no-go instructions. In our 435 study, the odour signalling HED or LED foods could have slowed the bottom-up processing of foods by adding 436 another element to take into account in the detection of stimuli. This indicates that olfactory food cues were 437 implicated in the detection process by slowing RT in individuals of higher weight status. We consequently 438 hypothesize that priming effects only influence the bottom-up processing of food cues.

439
The result of the priming effect seen here is congruent with the results of our previous study on attentional 440 biases. In this earlier study we found that implicit priming of olfactory food cues had differentiated effects: 441 individuals with obesity were more vulnerable to a non-attentively perceived pound cake odour (9 these aspects leads us to suppose that individuals with obesity might be subject to higher cognitive loads during 495 daily decision-making, which could alter their inhibitory control and consequently, produce goal-unrelated 496 behaviours. In our study, individuals were experimentally confronted to the same amount of cognitive load, 497 which did made it impossible to discriminate individual levels of inhibitory control toward foods according to 498 weight status. We now suggest that variations in everyday cognitive load might explain some of the relationships 499 between behaviourally reflected lack of inhibitory control facing foods and weight status that was identified in 500 other studies. In future research, these relationships should be characterized in order to better understand 501 overweight and obesity. 502

503
Several studies focusing on inhibitory control manipulated the cognitive processing of food stimuli by creating a 504 context of facilitation with priming (priming concepts of impulsivity (24) and unrestrained food consumption 505 (25), which led to interesting results. Nevertheless, such priming was explicit and is therefore not reflective of 506 incidental food cues from the environment, which was part of the objective of our study. Different forms of 507 implicit priming could be used in future research in order to assess the effects of implicit food cues on inhibitory 508 control or other top-down processes toward foods in a unimodal or multimodal manner. For instance, the 509 combination of auditory and olfactory priming has already been suggested as a means to influence individual 510 food choices (82). In future research, this type of multimodal priming could be used as an experimental context 511 of facilitation in order to elicit a lack of inhibitory control for food intake.