Eating is an essential act of our everyday life, and it involves complicated cognitive appraisal and gustatory evaluation. Multiple meta-analyses have attempted to pool data across studies to identify brain regions responsible for primary tasting (1–3); for processing visual, taste, and odor food stimuli (4), food commercials (5), affect, intensity, and quality of food stimuli (6), tastants delivered with and without swallowing (7); for differential processing between hungry and satiated state (8), male and female (9), obese and normal weight (10); and regions targeted by anti-appetite medications (11). (For a more complete list of meta-analyses on taste and food, please refer to (12)). Meanwhile, different kinds of food label, such as food brand logo, nutritional label, or label about food nature should modulate the cerebral processing of food information. An obvious analogy is graphic cigarette warning label. Some smokers and potential smokers may refrain from taking the cigarettes upon seeing such warning label printed on a pack of cigarettes. Compared to control warnings, it was reported that such graphic warning labels could significantly improve subjects’ motivation to quit smoking, and elicited stronger activation in numerous brain regions (13). The information provided by a nutritional label may similarly influence consumer behavior. For example, consumers were likely to pay 11% more for a box of cookies with a nutritional label compared to its counterpart without a label (14). There could be more modulating factors such as consumers’ nutritional knowledge and trust toward the nutrition and health claims (15, 16). On the other hand, food brand could also affect how much a consumer liked the product and his/her subsequent purchasing behavior (17). Understanding the neurobiology underlying such food choice and consumption behavior is obviously beneficial for marketing, nutritional advice, and pharmaceuticals. For instance, the cerebral regions highly related to gustatory processing, such as the insula and orbitofrontal cortex, were activated in response to visual cues of shower gel and dish liquid with beverage-imitating packages (18). This might explain why some people would unintentionally drink shampoo or other poisonous home products. In fact, anti-appetite medications often targeted the gustatory processing at the insula (11). Besides, it was found that discrete regions of the prefrontal cortex, responsible for decision making and self-control, were activated in response to seeing popular brand logos of foods such as French fries versus corresponding organic brand logos (19). It is important to know more about the neural correlates of various types of food labels, so that the designs of the food labels can be tested and optimized by neuroimaging evaluations.

Here, three types of food labels were examined: food brand logo, nutritional label, or label about food nature. Food brand logos might trigger a consumer’s memory about the past consumption experience and involve social facets such as brand loyalty. In 1985, the Coca Cola company introduced the so-called “New Coke” into the market, which was rated more superiorly than the original product in blinded taste tests. However, the word “new” printed on the can was largely rejected by consumers as they preferred the original image of the can and thus the original “package” of Coca Cola including the drink filled into the can (20). Similarly, it was found that human brain had differential gustatory processing of Coca Cola relative to Pepsi in multiple regions such as dorsolateral prefrontal cortex and hippocampus, implying differential drink preference (21). Meanwhile, nutritional label provides nutritional information of the food product that may influence liking, perception of healthiness, and ultimately food choice. There are three main groups of nutritional labels, traffic light, guideline daily amounts (GDA), and health logos/ratings (22). Consumers were more willing to pay for foods tagged with healthy nutritional labels than foods with unhealthy labels (23). In particular, seeing foods tagged with red (unhealthy) traffic light nutritional label would activate the inferior frontal gyrus and dorsolateral prefrontal cortex, regions implicated in self-control in food choice (23). Finally, the remaining type of label is about food nature. Food nature has diverse context. One fundamental aspect is food versus non-food. For instance, isovaleric acid is a chemical with a strong odor. It is present in foods (e.g., cheese and apple juice) and also produced by human body, attributing to sweaty smell and foot odor (24). The smell of isovaleric acid was rated much more unpleasant when it was labeled as body odor compared to labeled as cheddar cheese, with differential activations at the anterior cingulate, medial orbitofrontal cortex, and the amygdala (25). Apart from food versus non-food, the social/environment meanings of the food can also be expressed by the labels, such as Halal versus non-Halal for Muslims (26) was differentiated in the ventromedial prefrontal cortex, organic versus conventional (27) was differentiated in the ventral striatum, and fair trade versus conventional (28) was differentiated in the ventral striatum, cingulate cortex, and superior frontal gyrus.

Therefore, this meta-analytic study pooled data across functional magnetic resonance imaging (fMRI) studies. The activation likelihood estimation (ALE) meta-analytic approach was selected, which calculated the consistency of significantly activated brain coordinates reported from published studies by means of probability distribution modeling (29, 30). Based on the existing literature, it was hypothesized that the neural correlates of food brand logo, nutritional label, and label about food nature would involve the gustatory network, such as the insula and the prefrontal cortex.

The details of the 19 fMRI studies on food labels are listed in Table 1. Each study recruited 9–50 subjects, with two studies recruiting men only and one study women only. Regarding subject handedness, 10 studies recruited right-handed subjects only, whereas 4 studies recruited mostly right-handed subjects and 5 studies did not report details. Regarding subject age, most of the studies recruited young and middle-aged adults, whereas 3 studies focused on children and adolescents below the age of 18, and no study involved the elderly above the age of 65. For body mass index (BMI) of the subjects, 8 studies provided the mean data that ranged from 17.7 to 23.6. Meanwhile, 7 studies reported the number of hours the subjects fasted prior to the fMRI scan (1–6 h) and 1 study reported that the subjects did not fast. For the dimensions investigated, 5 studies investigated food brand, 6 studies investigated food nature, 5 studies investigated food nutrition, and 3 studies investigated food brand/nature.

For ALE meta-analysis, the overall analysis on all 19 studies, the analysis on food brand, and the analysis on food nature revealed no significant brain regions. Results remained insignificant even when a more liberal threshold of P < 0.05 (false-discovery rate corrected) with cluster size > 200 mm³ was applied.

The ALE meta-analysis on food nutrition revealed that the precuneus on the right hemisphere was significantly activated (Figure 3). The peak voxel of the activated cluster was at [6, −68, 40], with ALE value of 1.75 × 10^–2 and cluster size of 520 mm³.

The additional label-based meta-analysis showed that the activated brain regions for food brand processing seemed to be more diverse/inconsistent compared to food nature and food nutrition (Tables 2–4). Figure 4 shows the unthresholded ALE brain map for food brand. It illustrated that multiple brain regions were reported from the included studies, though none of them reached statistical significance.

This is the first meta-analysis on fMRI studies regarding brain responses to food labels. Surprisingly, results did not yield significant activations in the hypothesized brain regions of the insula and the prefrontal cortex. Results have shown that the cerebral processing of food labels on nutrition significantly activated the right precuneus. The processing of food labels on food brand and food nature did not yield significant results.

The insula is known to be the primary taste cortex that processes neural signals relayed from taste receptors in the oral cavity (51). The meta-analytic results here did not find significant activation of it, which might be partly explained by the fact that the analyzed studies did not involve actual tasting or stimulation of the taste receptors. Instead, the stimuli involved visual and cognitive aspects of food processing. Meanwhile, the prefrontal cortex is responsible for top-down affective modulation (51) and should be activated. The insignificant results might be attributed to heterogeneous designs of the studies.

The precuneus was reported to consume much more glucose than other parts of the cerebral cortex and thus sensitive to and easily affected by malnutrition (52). Among healthy subjects, precuneus was found to activate more in response to personalized nutritional messages compared to untailored nutritional messages (53). Meanwhile, its cortical thickness correlated with nutritional state and cognitive functions in patients with anorexia nervosa (54). Hence, the results of current meta-analysis on food nutrition labels were consistent to the existing literature that the precuneus seems to play an important role in nutritional evaluation.

At first glance, the insignificant result from food brand was unexpected. However, a prior fMRI study that presented visual cues of car brands showed that the cerebral processing of brands indeed involved diverse brain regions, e.g., familiar brands preferentially activated superior frontal gyrus, hippocampus and posterior cingulate; sports and luxury brands activated medial prefrontal cortex and precuneus; whereas brands deemed value products activated superior frontal gyrus and anterior cingulate (55). Therefore, the heterogeneity of subject background and the exact instructions given to them during fMRI scan might have attributed to the inconsistency across the analyzed studies. As a matter of fact, one of the analyzed studies on food brand did investigated Coca Cola and Pepsi separately, and compared brand-cued drink delivery to light-cued (colored circle) delivery (21). Significant results were only found for Coca Cola but not Pepsi. Another study similarly found that Coca Cola elicited greater activation than Pepsi in the right amygdala, whereas the opposite comparison resulted in no significant activation (32). This illustrated that the cerebral processing of food brand(s) could be complicated and not easily elicited in a consistent manner.

The heterogeneous situation was more obvious for food nature studies. Three studies investigated food versus non-food labels, followed by two investigated organic versus conventional food. Other comparisons were each investigated in one study, such as fair-trade label versus no label, food name label versus no label, eating alone label versus eating together label, and a correct match between flavor name and color on potato chip packaging versus a wrong match. It could be reasonably argued that different brain regions could be utilized for the cognitive processing of such distinct group of labels about food nature, rendering the results insignificant. There were even other labels on food nature being investigated. For instance, it was reported that there was a significant difference in the ventromedial prefrontal cortex activity level between viewing meat images labeled with Halal logo and those labeled with non-Halal logo, though this study was not included in the present analysis due to its lack of description on whether whole-brain analysis was performed (and also no brain coordinates provided) (26).

This meta-analysis has several limitations. First, the participants involved in the analyzed studies were mainly young and middle-aged adults with some children and adolescents. The results therefore might not be well applied to the elderly population, especially that aging was shown to associate with decreasing activation in the precuneus in response to food stimuli in a previous study (56). In addition, the analyses on the food brand and food nature subsets did not reveal significant results, so that potential follow-up analyses to compare between food brand, food nature, and food nutrition could not be conducted. Moreover, the majority of the analyzed studies exclusively recruited right-handed subjects or did not report subject handedness, similar to the scenario revealed by a prior survey on nutritional neuroscience literature (57). It was established that handedness would affect the lateralization of cerebral processing of tasting (58, 59). However, the methodology to gauge and determine subject handedness lacked a consensus and remained disputable (60). Regardless, due to the lack of results from exclusively left-handed subjects, it was not possible to test and determine if handedness would affect cerebral processing of food labels. Readers should be aware of these limitations during results interpretation. Finally, ALE meta-analysis can only compute the consistency of brain coordinates activated across studies but it cannot compute the effect size (30). Subsequently, future studies should include left-handed subjects in their sample. Moreover, the elderly people should be recruited so that the results could be more representative of the whole population. Researchers should share the statistical brain maps resulted from future original studies or report their effect size, so that the effect size could also be meta-analyzed.

In conclusion, the cerebral processing of food labels on nutrition significantly activated the right precuneus, a brain region related to nutritional evaluation. The peak voxel of the activated cluster was at [6, −68, 40], with ALE value of 1.75 × 10^–2 and cluster size of 520mm³. The processing of food labels on food brand and food nature did not yield significant results. More future studies are needed to better understand the cognitive processing of different kinds of food labels in our brain. They should include left-handed subjects as well as the elderly people, so that results could be more generalizable.

The original contributions presented in this study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

The author confirms being the sole contributor of this work and has approved it for publication.