Effects of single plant-based vs. animal-based meals on satiety and mood in real-world smartphone-embedded studies

Adopting plant-based diets high in fiber may reduce global warming and obesity prevalence. Physiological and psychological determinants of plant-based food intake remain unclear. As fiber has been linked with improved gut-brain signaling, we hypothesized that a single plant-based (vegetarian and vegan) compared to an animal-based (animal flesh) meal, would induce higher satiety, higher mood and less stress. In three large-scale smartphone-based studies, adults (nall = 16,379) ranked satiety and mood before and after meal intake. Meal intake induced satiety and higher mood. Plant-based meal choices did not explain differences in post-meal hunger. Individuals choosing a plant-based meal reported slightly higher mood before and smaller mood increases after the meal compared to those choosing animal-based meals. Protein content marginally mediated post-meal satiety, while gender and taste ratings had a strong effect on satiety and mood in general. We could not detect the profound effects of plant-based vs. animal-based meals on satiety and mood.


Subgroup analysis -Dietary habits (study 1)
To account for potential effects of dietary adherence, we repeated the analysis in subgroups for predominantly omnivorous (n = 11,600), predominantly vegetarian (n = 3,456) and predominantly vegan (n = 911) dieters.
The analysis in predominantly omnivorous dieters only presented the same effects as the whole group, i.e. showing slightly lower mood after a plant-based meal compared to animal-based meals, and also significantly higher post-meal hunger ratings for those choosing plant-based meals (Supplementary Table 4). Influences of taste remained significant for post-meal mood, yet the influence of taste ratings on hunger by meal category was no longer significant.
For predominantly vegetarian dieters only, results again resembled the main analysis, with no effect on post-meal hunger, yet lowered mood for those choosing a plant-based meal (note the unequal group size of animal-based meal n = 577, plant-based meal n = 2,880, Supplementary  Table 4). When accounting for taste ratings, post-meal hunger was lower overall for meals rated with five stars, yet hunger was higher when this was a plant-based meal. Mood ratings showed a similar paradoxical effect when adjusted for taste, namely improved mood for liked meals, yet when those were plant-based, mood was lowered.
For predominantly vegan dieters only, hunger was lower and mood was higher after choosing a plant-based meal (note the unequal group size of animal-based meal n = 117, plant-based meal n = 795 out of 911 vegan dieters, Supplementary Table 4). When accounting for taste, highly liked meals reduced hunger, yet there was no difference in hunger ratings between meal categories. Mood adjusted for taste, showed improved mood for highly liked meals, which was also shown for plant-based meals compared to animal-based ones.

Subgroup analysis -Time stamps (study 1)
To restrict entries to a reasonable timeframe around actual food intake, we additionally restricted the analyses as preregistered to entries that had more than 5 min and up to 3 h time lag between pre-and post-meal (hunger) entries (liberal timeframe) or more than 20 min and up to 1.5 h time lag between pre-and post-meal (hunger) entries (conservative timeframe). This led to a profound drop in sample size (liberal n = 3725, conservative n = 1878).
Results remained largely unchanged for the liberal timeframe from 5 min to 3 h difference between pre-and post-meal entries, namely significant main effects of post-meal timepoint with lower hunger and higher mood. For interaction effects of post-meal by meal category, no significant effects were found for hunger or mood, yet nominal differences were similar to the main analysis for hunger (higher post-meal hunger for plant-based meals), but not for mood (higher post-meal mood for plant-based meals). For the more conservative timeframe from 20 min up to 1.5 h postmeal, main effects for lower hunger and higher mood for post-meal timepoint were sustained. Yet, interaction effects of timepoint by meal category showed significant higher post-meal hunger for those choosing a plant-based meal (b = 0.14, t = 2.3, p = 0.024), while the main analysis showed non-significant results in the same direction of the estimate, and no significant effect on post-meal mood, yet also similar nominal differences with lower mood after plant-based meals (p > 0.14). Note that the reduced sample size when curating for timeframe indicate that the remaining entries had quite a high variety of time lag related to meal intake. Compliance with study design specific timeframes was manually curated in sub-studies 2 and 3 and have already been considered in all of the above analyses.

Confounder analysis -Coffee intake (studies 2+3)
Although not different across meal categories, coffee intake was assessed as a potential confounder on satiety and mood ratings in sub-studies two and three (data was in subgroups only and not available for the app study). Coffee intake did not explain a significant variance on hunger or contentment, but higher coffee consumption was related to higher stress levels (250-500ml: b = -0.17, t = -0.3; 500-1000ml: b = 6.2, t = 3.6; note the limited sample size of n = 32)

Emoji-based scales
Which scale is best to represent the given construct ?