A predictive model of rats’ calorie intake as a function of diet energy density

Rahmatollah Beheshti, Yada Treesukosol, Takeru Igusa, Timothy H. MoranbioRxiv 2017 · DOI

TL;DR

This paper introduces a novel mathematical model that accurately predicts rats' daily calorie intake based on their body weight, diet energy density, and the duration of the diet. The model's single, animal-specific parameter (ρ) shows a strong correlation with an animal's calorie intake and weight gain, suggesting it may serve as a proxy for "reward sensitivity."

What Problem Does This Paper Address?

The escalating rates of obesity worldwide are partly attributed to easy access to palatable, high-energy foods. Understanding the complex factors that drive food choices and overconsumption is crucial for developing effective prevention and treatment strategies. While research on high-energy diets is growing, a complete picture of how various elements influence food intake remains elusive. This paper aims to bridge this gap by developing a predictive model for calorie intake in rats, specifically focusing on how diet energy density affects consumption. This is particularly relevant as animal models offer a controlled environment to study these complex interactions, minimizing the confounding variables present in human studies.

Methods

The researchers developed a mathematical model to predict daily calorie intake (C) based on four key variables: body weight (W), energy density of the diet (E), elapsed time since the start of the diet (T), and a unique parameter for each animal, denoted as ρ (rho).

The model's foundation lies in the concept of "perceived reward" (Rc), which decreases as an animal consumes more food within a meal. This relationship is initially modeled as a logarithmic function: Rc = -log(βC), where β is a constant multiplier. To incorporate the influence of body weight, energy density, and time, β is further refined into a more complex term involving W, E, and T.

Through a process of fitting this model to experimental data, the researchers arrived at a specific form of the equation:

$C = P \frac{W \log(E)}{T^{1/4}}$

Here, P is a derived parameter that encapsulates the original β' and the functional forms of W, E, and T. The core of the model's predictive power comes from the parameter ρ, which is determined for each individual rat by fitting the model to their baseline chow diet data. This trained model is then used to predict calorie intake on a high-energy diet.

Datasets

The study utilized two independent datasets from rat experiments:

  • Study-1: Data from 16 male Sprague Dawley rats, approximately 275-300g, fed either a standard chow diet (3.1 kcal/g) or a high-energy (HE) diet (4.73 kcal/g) for 42 days. A recovery period on the chow diet followed.
  • Study-2: Data from 20 rats, similar in type and diet, where half received an HE diet (4.3 kcal/g) for 14 days after a habituation period on a chow diet (3.1 kcal/g).

The data from the chow diet periods were used for training the model (determining the ρ parameter for each rat), and the data from the HE diet periods were used for testing its predictive capabilities.

Key Insights and Ideas

  • A Universal Predictor? The model's ability to predict calorie intake across different diets using a single, animal-specific parameter (ρ) is a significant finding. This suggests that ρ might capture fundamental individual differences in feeding behavior.
  • ρ as a Proxy for Reward Sensitivity: The strong correlation between the ρ parameter and both calorie intake and weight gain on a high-energy diet is particularly intriguing. This hints that ρ could be a quantifiable measure of an animal's sensitivity to rewarding food stimuli, a concept central to understanding overconsumption and obesity.
  • Parsimonious Modeling: The model's strength lies in its simplicity, requiring only one free parameter (ρ) to be determined for each animal. This parsimony makes it a desirable and potentially more generalizable tool compared to more complex models.
  • Time and Energy Density Dynamics: The model explicitly incorporates the influence of diet energy density and the duration of exposure to a diet, reflecting known factors that modulate food intake.

Key Results

  • The developed model demonstrated high accuracy in predicting calorie intake patterns in rats on high-energy diets, as evidenced by the root-mean-square deviation (rmsd) values ranging from 6 to 19 and normalized rmsd (nrmsd) values between 0.08 and 0.57.
  • Regression analyses showed that the ρ parameter explained a substantial portion of the variance in average daily calorie intake during the high-energy diet period (R²=0.44 and R²=0.51 in the two studies).
  • Similarly, ρ explained significant variance in body weight gain by the end of the high-energy diet period (R²=0.26 and R²=0.59).
  • The model successfully replicated observed patterns of initial calorie intake increase followed by a decrease after switching to a high-energy diet, consistent with existing literature.

Limitations and Future Work

  • Single Meal Assumption: The current model implicitly assumes that all daily food intake occurs in a single meal. This simplification might limit its applicability to scenarios with more complex meal patterns.
  • Generalizability to Other Species: While the model shows promise in rats, its applicability to other species, including humans, requires further investigation.
  • Future Directions:
    • Extending the model to simulate meal patterns, including meal size and frequency.
  • Testing the model on genetically distinct rat groups (e.g., diet-induced obesity vs. diet-resistant rats) to explore the role of genetic background.
  • Further validating the hypothesis that ρ represents "reward sensitivity" by comparing it to established measures from reinforcement learning studies.
  • Investigating the model's performance with independent training and testing datasets to bolster confidence in its predictive power.

Summary updated

Abstract

Abstract Easy access to high-energy food has been linked to high rates of obesity in the world. Understanding the way that access to palatable (high fat or high calorie) food can lead to overconsumption is essential for both preventing and treating obesity. Although the body of studies focused on the effects of high energy diets is growing, our understanding of how different factors contribute to food choices is not complete. In this study, we present a mathematical model that is able to predict rats’ calorie intake to a high-energy diet based on their ingestive behavior to a standard chow diet. Specifically, we propose an equation that describes the relation between the body weight (W), energy density (E), time elapsed from the start of diet (T), and daily calorie intake (C). We tested our model on two independent data sets. Our results show that the suggested model is able to predict the calorie intake patterns with high accuracy. Additionally, the only free parameter of our proposed equation (ρ), which is unique to each animal, has a strong correlation with their calorie intake and weight gain. Additionally, we discuss the relevance of our derived parameter in the context of measuring reward sensitivity in reinforcement learning based studies.
BibTeX
@misc{beheshti2017predictive,
  author = {Rahmatollah Beheshti and Yada Treesukosol and Takeru Igusa and Timothy H. Moran},
  title = {{A predictive model of rats’ calorie intake as a function of diet energy density}},
  year = {2017},
  doi = {10.1101/184085},
  howpublished = {bioRxiv},
}
Figure