AI-driven personalized metabolic models of postprandial glucose to mixed meals

AI-driven personalized metabolic models of postprandial glucose to mixed meals | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article AI-driven personalized metabolic models of postprandial glucose to mixed meals Ricardo Gutierrez-Osuna, Anurag Das, Ghady Nasrallah, Bobak Mortazavi This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7511427/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 11 You are reading this latest preprint version Abstract Background/Objectives Diet monitoring for chronic disease management is often hindered by burdensome self-reporting. Automated techniques are possible using continuous glucose monitors (CGMs), which capture post-prandial glucose responses (PPGRs), but inter-individual variability limits predictive utility of models. We hypothesized that machine learning models that jointly represent meals and individual health parameters would improve prediction of PPGRs from meal composition (direct) and inference of meal composition from PPGRs (inverse). Subjects/Methods : Data came from the CGMacros dataset. Forty-five individuals with varied HbA1c, demographics, labs, and gut microbiome profiles were included and their PPGRs captured by CGMs. We developed JointCGMacros, a deep neural network framework that encodes PPGRs and participant health features. Using triplet loss, we aligned similar meal-health pairs while separating dissimilar ones in the embedding space. From this, we estimated the two-hour incremental area under the curve (iAUC) of PPGRs (direct) and carb-to-calorie ratio of meals (inverse). Models were trained and validated with 10-fold cross-validation, with performance measured using Pearson correlation and normalized root mean square error (NRMSE). Interventions/methods used and duration of administration CGMacros provided participants 10 breakfast shakes with varied macronutrient compositions, 10 Chipotle lunches with varied compositions, and free-choice dinners. Participants fasted before breakfasts and avoided eating for three hours after breakfast and lunch. Results: On the direct task, we achieved an NRMSE=0.22 and Pearson correlation r=0.59, outperforming baseline predictions (NRMSE=0.23, r =0.52), though not significantly at p<0.01. On the inverse task, we achieved an NRMSE=0.28 and r = 0.47, a statistically significant improvements over baseline predictions (NRMSE=0.62,p<0.01; r=0.17,p<0.001). Conclusions While PPGRs alone are not predictive of meal composition without accounting for individual health parameters, JointCGMacros integrates these factors, enabling more accurate automated dietary monitoring and estimation of meal macronutrient effects. Health sciences/Health care/Nutrition Health sciences/Diseases/Endocrine system and metabolic diseases/Diabetes/Type 2 diabetes Figures Figure 1 Figure 2 INTRODUCTION Monitoring and controlling diet are critical to preventing and managing chronic diseases, including obesity, diabetes and cardiovascular disease. However, manual recording of food intake (e.g., food diaries, 24-hour recall) is often time-consuming and error-prone [ 1 ]. Compounding the problem are the very large inter-individual differences in the response to the same foods [ 2 ], which puts into question the utility of universal dietary recommendations. Thus, there is a need for new techniques that can (1) reduce the burden of monitoring food intake and (2) allow individuals to personalize their diets to achieve optimum health. Several sensing technologies have been proposed to assist in automatic diet monitoring . Physical sensors, in particular, have been the subject of multiple studies [ 3 ]. For example, wearable inertial measurements (e.g., from accelerometers and gyroscopes in smartwatches or fitness trackers) can be used to detect eating gestures [ 4 ], and wearable acoustic sensors can be used to identify chewing and swallowing sounds [ 5 ]. Physical sensors have also been placed in “smart utensils” to detect eating moments and eating speed [ 6 ]. Chemical sensors have also been developed to analyze biomarkers associated with metabolism, including breath measurements of ketones (Ketonix AB, Varberg, Sweden) and CO2 (Lumen, MetaFlow Ltd., Tel Aviv, Israel), and ketones in interstitial fluid [ 7 , 8 ]. However, these technologies only capture limited information, such as the timing and type of foods, but not their macronutrient composition, which is critical in disease management. Studies over the past decade have explored the use of continuous glucose monitors (CGMs) to develop personalized nutrition programs that account for inter-individual differences. CGMs are in-dwelling wearable devices that measure glucose concentration in interstitial fluid. Though initially developed for Type 1 Diabetes, CGMs are increasingly being used to treat individuals with Type 2 Diabetes [ 9 ]. Several companies (e.g., Levels, NutriSense) also offer CGM-based services for the wellness market. The origins of CGM-based personalized nutrition programs can be traced back to the seminal study of Zeevi, Korem [ 10 ]. In this study, the authors collected CGM data and meal logs from an 800-person cohort and developed a machine-learning (ML) model that integrated clinical features, dietary habits, physical activity, and gut microbiome profiles to predict glycemic responses to meals. Their approach was validated in a randomized controlled trial, where personalized dietary recommendations significantly reduced postprandial glucose excursions. These studies have been replicated in different cohorts [ 11 , 12 ]. CGM-based personalized nutrition approaches such as that of Zeevi, Korem [ 10 ] can be thought of as developing a direct metabolic model that predicts post-prandial glucose responses (PPGRs) from meal macronutrients. Over the past decade, our work has focused on the inverse metabolic model: predicting meal macronutrients from PPGRs, as an approach for automatic diet monitoring [ 13 ]. The primary contributor to elevated PPGRs is the amount (and type) of carbohydrates, but PPGRs are also influenced by other, non-glycemic macronutrients in the meal –see Fig. 1 a-b. For example, adding protein and fat to a carbohydrate-rich meal can reduce the initial peak and slow down recovery to baseline [ 14 , 15 ]. Unfortunately, the inverse problem is far more challenging, as many different meals can lead to similar PPGRs. Compounding the problem, metabolic health parameters (e.g., HbA1c, insulin sensitivity) and demographics (e.g., sex, age) also play a major role in PPGRs to identical meals. As a result, PPGRs from two individuals cannot be compared without controlling for these health/demographic factors. To address this challenge, we present an ML model that addresses both the direct (i.e., personalized nutrition) and the inverse (i.e., automatic diet monitoring) problems. Our model learns a numerical representation (referred to as an embedding in ML research) that aligns PPGRs with the macronutrient composition of a meal, conditioned on metabolic and demographic parameters. The result is an interpretable and personalized embedding of a meal that accounts for inter-individual differences. MATERIALS/SUBJECTS AND METHODS Model design Our proposed model (JointCGMacros) uses neural networks to learn the underlying relationship between PPGRs, meal macronutrients and the individual’s health parameters. Illustrated in Fig. 1 b, the model consists of three distinct components: (1) a neural network that converts PPGRs into a single numeric value, i.e., a one-dimensional (1D) embedding, (2) a weighted parametric equation that converts meals macronutrients into a separate 1D embedding, and (3) a second neural network that converts each subject’s health parameters into the set of weights used by the parametric equation. In what follows, we will refer to the first neural network as the “PPGR encoder”, and the second neural network as the “health encoder”. We train the coefficients of the two neural networks so that their respective outputs match using an error function known as the triplet loss [ 16 ] (see Supplementary materials). To use the triplet loss, we give the model two pairs of examples. The first pair contains PPGR and macronutrients from the same meal, whereas the second pair contains PPGR and macronutrients from different meals. The triplet loss forces the neural network outputs from the first pair to be closer to each other, and those from the second pair to be as far from each other as possible –see Fig. 1 d. In doing so, the model learns how meal macronutrients and health parameters contribute to the observed PPGR, as measured by a CGM. Formally, the PPGR encoder takes a 13-dimensional vector containing the 3-hour PPGR to a meal (as measured by an Abbot Freestyle Libre Pro CGM every 15 min), which we denote by \(\:g\left(t\right)\) , and generates a 1D embedding \(\:{z}_{P}\) . In turn, the health encoder takes a 34-dimensional vector of health parameters \(\:H\) (see Table 1 c) and generates a personalized set of weights that are combined with the macronutrient amounts of a meal \(\:M\) to produce a 1D embedding \(\:{z}_{M}\) as: $$\:{z}_{M}=\frac{C-{w}_{P}P-{w}_{F}F-{w}_{B}B}{1+{w}_{C}^{{\prime\:}}C+{w}_{P}^{{\prime\:}}P+{w}_{F}^{{\prime\:}}F+{w}_{B}^{{\prime\:}}B}$$ 1 where \(\:C\) , \(\:P\) , \(\:F\) , and \(\:B\) are the amount of net carbs, protein, fat and fiber, respectively, in calories 3 . Coefficients \(\:{w}_{P}\) , \(\:{w}_{F}\) and \(\:{w}_{B}\) in the numerator denote how much protein, fat, and fiber subtract from the effect of carbs, whereas coefficients \(\:{w}_{C}^{{\prime\:}}\) , \(\:{w}_{P}^{{\prime\:}}\) , \(\:{w}_{F}^{{\prime\:}}\) and \(\:{w}_{B}^{{\prime\:}}\) in the denominator produce an adjusted caloric content. As such, the embedding \(\:{z}_{M}\) can be interpreted as a generalization of the carb-caloric-ratio (CCR), the proportion of calories in a meal that come from carbs, which is a measure of the glycemic load of a meal: $$\:CCR=\frac{C}{C+P+F+B}$$ 2 Note that for \(\:{w}_{C}^{{\prime\:}}=\:{w}_{P}^{{\prime\:}}=\:{w}_{F}^{{\prime\:}}={w}_{B}^{{\prime\:}}=0\) in Eq. ( 1 ), the denominator becomes \(\:1\) , in which case \(\:{z}_{M}\) only considers the subtractive effect of non-glycemic macronutrients (protein, fat and fiber). In turn, for \(\:{w}_{C}^{{\prime\:}}={w}_{P}^{{\prime\:}}={w}_{F}^{{\prime\:}}={w}_{B}^{{\prime\:}}=1\) and \(\:{w}_{P}={w}_{F}={w}_{B}=0\) , \(\:{z}_{M}\) reduces to the CCR in Eq. ( 2 ). This parametric embedding has two main advantages. First, it allows the model to learn the optimum combination of macronutrients that is related to PPGRs (either linearly or non-linearly) for each subject. Second, because each of the resulting weights corresponds to a single macronutrient, the model is interpretable, as we will see in the Results section. Both \(\:{z}_{M}\) and \(\:{z}_{P}\) are 1D embeddings to prioritize interpretability and reduce the risk of overfitting the training dataset. The inverse metabolic model: predicting macro embeddings from PPGRs Once trained, the model can be used to predict the macro embedding ( \(\:{z}_{M}\) ) of a meal from the corresponding PPGR, as illustrated in Fig. 1 e. First, the PPGR of the meal is passed to the PPGR encoder to obtain the corresponding embedding \(\:{z}_{P}\) . Then, the individual's health parameters \(\:H\) are passed to the macro encoder and combined with the meal’s macronutrients as a vector \(\:M\) , to produce the embedding \(\:{z}_{M}\) . Since \(\:M\) is unknown for a test PPGR, we generate multiple embeddings \(\:({z}_{M,1},\:{z}_{M,2},\:...\:{z}_{M,n})\) from a list of potential meals \(\:({M}_{1},\:{M}_{2},\:...\:{M}_{n})\) in a dataset. The meal \(\:{M}_{i}\) whose macronutrient embedding \(\:{z}_{M,i}\) is closest to \(\:{z}_{P}\) is treated as the prediction. We compare the inverse mapping in JointCGMacros against a baseline model that predicts the CCR in Eq. ( 2 ) directly from its corresponding PPGR. The baseline model is a feedforward network with the same number of parameters as the PPGR encoder in JointCGMacros; see Supplementary materials. Because the baseline model does not use health parameters H, this comparison allows us to determine if the personalized macro embedding \(\:{z}_{M}\) can reduce the effect of inter-individual differences. We train both models using 10-fold cross-validation. For each training fold, we randomly select 80% of the data for training and 20% for validation. We use validation data for early stopping (10 epochs) to prevent overfitting. We report two performance measures on the test fold: the normalized root mean squared error (NRMSE) and the Pearson correlation between ground truth and prediction. Because the NRMSE is the percentage of error relative to ground truth, it allows us to compare errors between target variables of different scales. The direct metabolic model: predicting PPGRs from macro embeddings Once trained, the joint model can also be used to predict the full PPGR of a meal from the corresponding set of an individual’s health parameters \(\:H\) and meal macronutrients \(\:M\) , as illustrated in Fig. 1 e. First, parameters \(\:\left(H,M\right)\) are passed to the macro encoder to obtain the corresponding embedding \(\:{z}_{M}\) . Then, we find the PPGR embedding \(\:{z}_{P}\) in the training set that is closest to \(\:{z}_{M}\) . Finally, we recover PPGR from \(\:{z}_{P}\) by inverting the PPGR encoder via linear regression using \(\:\left({\sqrt{{z}_{M}},\:z}_{M},\:{z}_{M}^{2},{z}_{M}^{3}\right)\) as predictors, one linear regression model for each time step in the PPGR, i.e., \(\:g\left(t\right)\) , for \(\:t=15,\:30,\dots\:\) minutes; see Fig. 1 b. We compare the direct mapping in JointCGMacros against a baseline model that predicts each PPGR timestep from parameters \(\:\left(H,M\right)\) . The baseline model is a feedforward network with the same number of parameters as the PPGR encoder in JointCGMacros; see Supplementary materials. As before, we train both models using 10-fold cross-validation and use validation data for early stopping (10 epochs) to prevent overfitting. We report the NRMSE and Pearson correlation between ground truth and prediction for the two-hour incremental area-under-the-curve (iAUC2hr), as is common to examine results from the Oral Glucose Tolerance Test (OGTT). Note that, in this case, the odds are stacked against the JointCGMacros because it predicts PPGRs from a single variable ( \(\:{z}_{M}\) ), whereas the baseline model uses 38 predictors (i.e., the 34 health parameters in Table 1 c plus the 4 macronutrients). Thus, this comparison allows us to determine if the embedding \(\:{z}_{M}\) is a good summary statistic of \(\:\left(H,M\right)\) . Data records We trained the proposed model on the CGMacros dataset [ 17 ] publicly-available on PhysioNet, which contains PPGRs, meal macronutrients and health/demographics parameters for 45 subjects (ages 18–69 and body mass index (BMI) 21–46 kgm 2 ); clinical trial #NCT04991142. Table 1 b summarizes the demographic information of study participants. As part of the screening process, we measured the subjects’ BMI, glycated hemoglobin (HbA1c), fasting glucose, fasting insulin, triglyceride, and cholesterol levels, as well as their demographic information (age, gender, and race). Participants wore an Abbott FreeStyle Libre Pro CGM (15-min sampling period) and a Dexcom G6 Pro CGM (5-min sampling) in the upper arm and abdomen, respectively. Both CGMs were blinded to prevent glucose readings from influencing participants. Participants were also provided with a Fitbit smartwatch (Fitbit Sense) to log exercise, and were trained to use the MyFitnessPal mobile app to log their meals and take pictures of their foods using the WhatsApp mobile app. Each subject recorded their meals for 10 days, including breakfast, lunch and dinner. Breakfasts consisted of protein shakes with varying amounts of carbohydrates, protein, fat, and fiber -see Table 1 a. Lunches were ordered from a local, fast-casual restaurant chain (Chipotle Mexican Grill). The breakfast and lunch meals were designed to cover a range of macronutrient contents. For dinners, participants ate foods of their own choice. To minimize interferences in glucose responses from prior meals, participants were instructed to eat lunch at least three hours after breakfast, with only water or coffee (without sugar) in between, and dinner at least three hours after lunch. They also took photographs of the meals before and after eating, from which we extracted the meal timestamps and the proportion of the meal they consumed. At the start of the study, we also collected stool samples using a Viome microbiome kit (Viome Life Sciences, Inc.) to obtain the gut microbiome profile for each subject. We report results from the breakfast meals using PPGRs from the Abbott FreeStyle CGM. We used 34 variables available on the CGMacros dataset for each subject; see Table 1 c. RESULTS Predictive accuracy on test data First, we compare the performance of JointCGMacros model against the baseline inverse model that predicts CCR directly from PPGRs (i.e., without personalization). Results are summarized in Fig. 2 a. JointCGMacros strongly outperforms (lower NRMSE, \(\:p<0.01\) ; higher correlation, \(\:p<0.001\) ) the baseline model, indicating that JointCGMacros accounts for inter-individual differences in postprandial glucose responses. Note that, for the baseline model, the correlation coefficient can become negative and the NRMSE can exceed an error of 100%, indicating that PPGRs are not predictive of the macronutrient composition of meals unless individual health/demographics parameters are considered . We also compare the performance of JointCGMacros model against a baseline direct model that predicts PPGRs directly from parameters \(\:\left(H,M\right)\) . In this case, we compute the iAUC-2hr from the predicted PPGRs and compare against the ground truth’s iAUC-2hr. Results are summarized in Fig. 2 b. JointCGMacros achieves similar NRMSE and correlation coefficients ( \(\:p>0.05\) ) despite using a single predictor ( \(\:{z}_{M})\) , whereas the baseline model uses 38 predictors. This suggests that the complex relationship between health parameters, gut microbiome and food macronutrients is intrinsically low dimensional. We also examined pairwise predictions from the macro and PPGR encoders. Results are illustrated in Fig. 2 c as scatterplot of the corresponding embeddings ( \(\:{z}_{M},{z}_{P}\) ), color-coded by the ground-truth CCR in Eq. ( 2 ). Embeddings towards the bottom left (dark blue) have the lowest CCR, which represent meals with high amounts of carbs, protein, fat, and fiber (e.g., meal HHHH in Table 1 ). Since these meals have the highest amounts of non-glycemic macronutrients, they tend to result in low postprandial glucose. In contrast, meals with the highest CCR (dark red) appear towards the top right. These meals have a high amount of carbs and low amounts of other macronutrients (HLLL in Table 1 ), which leads to higher postprandial glucose. Model interpretation A major limitation of deep-learning models is their lack of interpretability, which can be problematic in critical applications, such as health. To address this limitation, we also conduct a series of analyses to shed light on the inner workings of our proposed model. In a first step, we quantify the importance of the \(\:34\) health parameters in JointCGMacros using Bidirectional Search (BDS), a procedure that alternates between forward selection and backward elimination of predictors. Starting from the empty set, the forward step adds the best single predictor (i.e., lowest NRMSE) to the model; in turn, starting from the full set, the backward step eliminates the worst single predictor (i.e., highest NRMSE). As a result, both searches converge in the middle ( \(\:17\) predictors). Results are shown in Fig. 2 d. The two most important features are HOMA-IR, a clinical measure of insulin resistance and beta-cell function, and fasting glucose, a key parameter for the diagnosis of T2D. The final feature set includes a mix of demographics (age, gender), metabolic health (lipid profiles, HbA1c), and several pathways from gut microbiome analytics. This optimal set achieves lower NRMSE than the full feature set (rightmost boxplot, in gray), a difference that is statistically significant (0.247 vs. 0.291, \(\:p=0.03\) ). We also examined the relationship between the two 1D embeddings (i.e., \(\:{z}_{P}\) from PPGRs; \(\:{z}_{M}\) from health parameters and meal macros) learned by the model. Figure 2 e shows the average PPGR at increments of 5% in the quantile distribution of the PPGR embedding \(\:{z}_{P}\) . PPGRs in the lowest quantiles of \(\:{z}_{P}\) (i.e., 5–20%) show a biphasic characteristic, with a rise in the first 30–60 minutes, a subsequent drop, and another increase 90–120 minutes later. This type of PPGR is considered an indicator of good metabolic health [ 18 ]. In contrast, PPGRs in the highest quantiles of \(\:{z}_{P}\) (i.e., 70–95%) show a monophasic characteristic, with a single peak followed by a continuous drop, which is indicative of less optimal metabolic health. Likewise, we analyzed the distribution of health parameters for samples in the first (25%) and fourth (75%) quartiles of the distribution of \(\:{z}_{M}\) . Results for a selected few health parameters and meal macronutrients are shown in Fig. 2 f. On average, subjects with low HbA1c, low fasting glucose, low HOMA-IR and low BMI tend to produce lower embeddings \(\:{z}_{M}\) , regardless of meal macronutrients. Likewise, meals high in carbs, low in protein, or low in fiber tend to produce higher embeddings \(\:{z}_{M}\) , results that are consistent with the literature [ 19 ]. To corroborate these observations, we examined the weights in Eq. ( 1 ) learned by the model. In the numerator, fiber ( \(\:{w}_{B}=3.95\) ) has the strongest effect, followed by protein ( \(\:{w}_{P}=1.22\) ) and fat ( \(\:{w}_{F}=0.34\) ), suggesting healthy alternatives to control elevated post-prandial glucose response (i.e., add fiber and protein, but not fat). In the denominator, net carbs ( \(\:{w}_{C}^{{\prime\:}}\) ) have the strongest contribution ( \(\:{w}_{C}^{{\prime\:}}=0.33\) ), whereas the effect of the three non-glycemic macronutrients is negligible ( \(\:{w}_{P}^{{\prime\:}}=0.03,\:{w}_{F}^{{\prime\:}}=0.01,\:{w}_{B}^{{\prime\:}}=0.01\) ). Ignoring the smallest coefficients, these results indicate that, when accounting for health parameters , post-prandial glucose responses are proportional to the expression \(\:(C-3.95B-1.22P)/(1+0.33C)\) . DISCUSSION Monitoring diet in an automated fashion is made possible by ML methods that account for both a person’s response to a meal (as captured by a CGM), and their individual demographics and characteristics. By jointly learning an individual’s health characteristics and observing his/her responses to individual meals, we have developed a model (JointCGMacros) that makes possible automated modeling of diet by both generating a PPGR from a known meal (direct) and estimate a meal’s macronutrient composition from a captured response as measured by a CGM (inverse). Baseline models that estimate the direct or indirect task without health parameters can generate errors over 100% NRMSE and JointCGMacros significantly outperforms the baselines. Perhaps more importantly, when compared with a baseline that uses higher dimensionality, JointCGMacros, which compresses data into a single embedding for PPGRs and a single embedding for health parameters, uses a low dimensional structure to create meaningful metabolic representations. Prior research has shown that the relationship between post-prandial glucose and meal macronutrients is largely driven by inter-individual differences. Our results show that this three-way relationship can be unveiled by learning a joint embedding of PPGRs and macronutrient composition conditioned on a few demographics, metabolic health, and gut microbiota parameters. We find that the key contributors to elevated postprandial glucose are net carbs (as expected) with a linear correction for fiber and protein. It is the expression \(\:(C-3.95B-1.22P)/(1+0.33C)\) , and not the conventional carb-caloric-ratio, that can be recovered from PPGRs, but only when accounting for an individual's health variables. Beyond predictive performance, we also identified factors related to the meals and health parameters that drove model performance. These features included HOMA-IR and fasting glucose information as the most influential health predictors, aligning with established risk markers for Type 2 Diabetes, while embedding low-dimensional representation of gut microbiome pathways. Moreover, the embeddings revealed biologically plausible distinctions between monophasic and biphasic glucose curves, echoing prior associations with metabolic health. At the macronutrient level, fiber and protein emerged as protective factors, while excess carbohydrate intake strongly drove elevated responses. Together, these insights suggest that JointCGMacros not only improves prediction but also uncovers mechanistic relationships that may guide personalized nutrition strategies. Limitations The study did have some limitations that can be addressed with future study. First, we used the parametric expression in Eq. ( 1 ) to improve the interpretability of our model, but further improvements in macronutrient prediction from PPGRs may be achieved by relaxing this constraint, e.g., concatenating macronutrient and health vectors \(\:[M,H]\) and feeding them to a deep-learning model, but this would likely require access to a larger dataset than CGMacros, the only of its nature that is publicly available. Second, CGMacros contains individuals who are not taking insulin. Therefore, monitoring for individuals requiring insulin will require data and modeling of additional health characteristics and insulin dosage to model the effect on the PPGR. Declarations Acknowledgements This work was supported by National Science Foundation Award No. 2014475. Author contributions AD was responsible for data curation and data analytics. AD and GN were responsible for implementation and evaluation of the proposed machine learning model. BM was Principal Investigator of the NSF award; he was responsible for project oversight and co-lead data analytics and meal design. RGO was co-Investigator of the NSF award; he co-led data analytics for the NSF award and was responsible for manuscript preparation. Data availability statement The dataset used for this study is publicly available at PhysioNet (https://physionet.org/content/cgmacros) Competing Interests: The authors declare no competing interests for this work. This work was supported by National Science Foundation Award No. 2014475. References Mortazavi, B.J. and R. Gutierrez-Osuna, A review of digital innovations for diet monitoring and precision nutrition. J Diabetes Sci Technol, 2023. 17 (1): p. 217--223. Zeevi, D., et al., Personalized Nutrition by Prediction of Glycemic Responses. Cell, 2015. 163 (5): p. 1079-1094. Bell, B.M., et al., Automatic, wearable-based, in-field eating detection approaches for public health research: a scoping review. npj Digital Medicine, 2020. 3 (1): p. 38. Zhang, B., et al., Enabling Eating Detection in a Free-living Environment: Integrative Engineering and Machine Learning Study. J Med Internet Res, 2022. 24 (3): p. e27934. Kumar, Y., et al., Automated detection and recognition system for chewable food items using advanced deep learning models. Scientific Reports, 2024. 14 (1): p. 6589. Huang, Q., Z. Yang, and Q. Zhang. Smart-u: Smart utensils know what you eat . in IEEE Conference on Computer Communications (INFOCOM) . 2018. IEEE. Zhang, J.Y., et al., Continuous Ketone Monitoring: A New Paradigm for Physiologic Monitoring. J Diabetes Sci Technol, 2021: p. 19322968211009860. Alva, S., et al., Feasibility of Continuous Ketone Monitoring in Subcutaneous Tissue using a Ketone Sensor. J Diabetes Sci Technol, 2021: p. 19322968211008185. Jancev, M., et al., Continuous glucose monitoring in adults with type 2 diabetes: a systematic review and meta-analysis. Diabetologia, 2024. 67 (5): p. 798-810. Zeevi, D., et al., Personalized Nutrition by Prediction of Glycemic Responses. Cell, 2015. 163 (5): p. 1079--1094. Mendes-Soares, H., et al., Assessment of a Personalized Approach to Predicting Postprandial Glycemic Responses to Food Among Individuals Without Diabetes. JAMA Netw Open, 2019. 2 (2): p. e188102. Mendes-Soares, H., et al., Model of personalized postprandial glycemic response to food developed for an Israeli cohort predicts responses in Midwestern American individuals. Am J Clin Nutr, 2019. 110 (1): p. 63--75. Zhang, L., et al. Joint Embedding of Food Photographs and Blood Glucose for Improved Calorie Estimation . in 2023 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI) . 2023. Moghaddam, E., J.A. Vogt, and T.M.S. Wolever, The effects of fat and protein on glycemic responses in nondiabetic humans vary with waist circumference, fasting plasma insulin, and dietary fiber intake. J Nutr, 2006. 136 (10): p. 2506--2511. Rytz, A. and D. Adeline, Predicting glycemic index and glycemic load from macronutrients to accelerate development of foods and beverages with lower glucose responses. Nutrients, 2019. 11 (5): p. 1172. Weinberger, K.Q. and L.K. Saul, Distance metric learning for large margin nearest neighbor classification. Journal of machine learning research, 2009. 10 (2). Gutierrez-Osuna, R., et al. CGMacros: a scientific dataset for personalized nutrition and diet monitoring . 2025; Available from: https://physionet.org/content/cgmacros. Robbins, D.C., et al., Biphasic patterns of peripheral insulin and glucose levels after lunch in normal subjects. Diabetes Care, 1987. 10 (3): p. 293-9. Ketema, E.B. and K.T. Kibret, Correlation of fasting and postprandial plasma glucose with HbA1c in assessing glycemic control; systematic review and meta-analysis. Arch Public Health, 2015. 73 : p. 43. Ioffe, S. and C. Szegedy, Batch normalization: Accelerating deep network training by reducing internal covariate shift , in International Conference on Machine Learning (ICML) . 2015. p. 448--456. Li, Y. and Y. Yuan, Convergence analysis of two-layer neural networks with relu activation , in Advances in Neural Information Processing Systems (NIPS) . 2017. Footnotes The dataset used for this study contains six unique meal combinations. This results in five distinct triplets per CGM, each sharing the same anchor and positive, but differing in the negative Tables Additional Declarations There is NO conflict of interest to disclose Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: revise 27 Apr, 2026 Review # 3 received at journal 23 Mar, 2026 Reviewer # 3 agreed at journal 19 Feb, 2026 Review # 2 received at journal 13 Oct, 2025 Reviewer # 2 agreed at journal 29 Sep, 2025 Reviewer # 1 agreed at journal 23 Sep, 2025 Reviewers invited by journal 22 Sep, 2025 Submission checks completed at journal 03 Sep, 2025 First submitted to journal 02 Sep, 2025 Unknown event 02 Sep, 2025 Editor assigned by journal 01 Sep, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7511427","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":519019477,"identity":"99a77156-b337-47dc-8c6c-4f174f4b2014","order_by":0,"name":"Ricardo Gutierrez-Osuna","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAu0lEQVRIiWNgGAWjYLCCBAYGOSiTmSgNjA1ALcY8pGkBEok9RGsx719j/uBBxb30/eyn0yQYKqwTGwhpkbnxxrAh4Uxxbg9P7jYJhjPphLVISJwxbEhsS8jtYQBqYWw7TKyWfwnpPPxvgVr+EaOFvweopSEhgUcCZEsDUbawFc5IOJZg2HPj7WaLhGPpxkTYcnjDxx81CfLs/bkbb3yosZYlqIVBIgGJk4BDESrgP0CUslEwCkbBKBjJAABKqT3yUHvh1QAAAABJRU5ErkJggg==","orcid":"","institution":"Texas A\u0026M University","correspondingAuthor":true,"prefix":"","firstName":"Ricardo","middleName":"","lastName":"Gutierrez-Osuna","suffix":""},{"id":519019478,"identity":"df798ca3-7b02-4179-a863-797b0a41da2e","order_by":1,"name":"Anurag Das","email":"","orcid":"","institution":"Texas A\u0026M University","correspondingAuthor":false,"prefix":"","firstName":"Anurag","middleName":"","lastName":"Das","suffix":""},{"id":519019479,"identity":"2a4fe908-89a1-4fe8-9650-9a44166b88bb","order_by":2,"name":"Ghady Nasrallah","email":"","orcid":"","institution":"Texas A\u0026M University","correspondingAuthor":false,"prefix":"","firstName":"Ghady","middleName":"","lastName":"Nasrallah","suffix":""},{"id":519019480,"identity":"0fd718bf-39c9-4a24-a7f6-875b4e3f485d","order_by":3,"name":"Bobak Mortazavi","email":"","orcid":"","institution":"Texas A\u0026M University","correspondingAuthor":false,"prefix":"","firstName":"Bobak","middleName":"","lastName":"Mortazavi","suffix":""}],"badges":[],"createdAt":"2025-09-01 20:35:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7511427/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7511427/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":92848306,"identity":"ec258324-48d2-4420-82e1-012933053654","added_by":"auto","created_at":"2025-10-06 10:07:05","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1279631,"visible":true,"origin":"","legend":"","description":"","filename":"ijoFINAL.docx","url":"https://assets-eu.researchsquare.com/files/rs-7511427/v1/bf1e6f9feacdd20d734eb0b7.docx"},{"id":92848297,"identity":"a93075c2-5f34-439a-a189-9ace198d479c","added_by":"auto","created_at":"2025-10-06 10:07:05","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6654,"visible":true,"origin":"","legend":"","description":"","filename":"2025IJO01445.json","url":"https://assets-eu.researchsquare.com/files/rs-7511427/v1/7bb61a950d6c71fbb378ce39.json"},{"id":92849070,"identity":"8d737c89-4737-4e60-bb8b-872a550bbdd3","added_by":"auto","created_at":"2025-10-06 10:15:05","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":86469,"visible":true,"origin":"","legend":"","description":"","filename":"2025IJO014450enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7511427/v1/4935a95efab2549256061c79.xml"},{"id":92849071,"identity":"ef77dce7-363b-410a-91fd-9557c51953b6","added_by":"auto","created_at":"2025-10-06 10:15:05","extension":"emf","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3071148,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.emf","url":"https://assets-eu.researchsquare.com/files/rs-7511427/v1/c245d501b397bb1df2c0e1df.emf"},{"id":92848298,"identity":"28d1846c-5306-4238-80ec-3b0201750edf","added_by":"auto","created_at":"2025-10-06 10:07:05","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":91509,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7511427/v1/bee48a06827619841df60c2f.png"},{"id":92848307,"identity":"2811c23f-adb8-4203-80d6-579cba92d795","added_by":"auto","created_at":"2025-10-06 10:07:05","extension":"jpeg","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":516889,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7511427/v1/05a69ade097b310af84154d9.jpeg"},{"id":92848301,"identity":"4c747fbe-2f32-4854-8d25-fe6183843b64","added_by":"auto","created_at":"2025-10-06 10:07:05","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":482879,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7511427/v1/7ee3c4e34c3c160d1739afe0.png"},{"id":92848302,"identity":"da5b6809-ac20-45fc-8141-29c662a02477","added_by":"auto","created_at":"2025-10-06 10:07:05","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":25993,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7511427/v1/93879ce645abf1ed94a6a932.png"},{"id":92848303,"identity":"489315a8-d585-468c-8eec-83137c32c69d","added_by":"auto","created_at":"2025-10-06 10:07:05","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":270202,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7511427/v1/0fe2b90442de4c80e68b1597.png"},{"id":92848310,"identity":"a9d7b919-7a4e-4a30-8420-7f3b6a73e582","added_by":"auto","created_at":"2025-10-06 10:07:05","extension":"xml","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":82000,"visible":true,"origin":"","legend":"","description":"","filename":"2025IJO014450structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7511427/v1/a3bf6b966de7c7edbec6a4b0.xml"},{"id":92848309,"identity":"166a2d08-1b26-4277-b896-88352ddd1b49","added_by":"auto","created_at":"2025-10-06 10:07:05","extension":"html","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":103807,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7511427/v1/f94206d8d7a293cdd0f9f72c.html"},{"id":92849069,"identity":"422140ae-0f26-4870-b1f7-878873e60666","added_by":"auto","created_at":"2025-10-06 10:15:05","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1036596,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u003c/p\u003e","description":"","filename":"fig1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7511427/v1/2879470c810bb06a7caad01f.jpg"},{"id":92848300,"identity":"6969b8a4-112e-42f6-a8ec-35e396db8d60","added_by":"auto","created_at":"2025-10-06 10:07:05","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1161491,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u003c/p\u003e","description":"","filename":"fig2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7511427/v1/b75061c784f27367dc169356.jpg"},{"id":92849275,"identity":"7a5635df-b1ef-4786-874b-f068a7f106d8","added_by":"auto","created_at":"2025-10-06 10:23:11","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2966030,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7511427/v1/52bf9dc3-c35a-4975-844b-0ec26745461a.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e conflict of interest to disclose","formattedTitle":"AI-driven personalized metabolic models of postprandial glucose to mixed meals","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eMonitoring and controlling diet are critical to preventing and managing chronic diseases, including obesity, diabetes and cardiovascular disease. However, manual recording of food intake (e.g., food diaries, 24-hour recall) is often time-consuming and error-prone [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Compounding the problem are the very large inter-individual differences in the response to the same foods [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e], which puts into question the utility of universal dietary recommendations. Thus, there is a need for new techniques that can (1) reduce the burden of monitoring food intake and (2) allow individuals to personalize their diets to achieve optimum health.\u003c/p\u003e\u003cp\u003eSeveral sensing technologies have been proposed to assist in \u003cem\u003eautomatic diet monitoring\u003c/em\u003e. Physical sensors, in particular, have been the subject of multiple studies [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. For example, wearable inertial measurements (e.g., from accelerometers and gyroscopes in smartwatches or fitness trackers) can be used to detect eating gestures [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e], and wearable acoustic sensors can be used to identify chewing and swallowing sounds [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Physical sensors have also been placed in \u0026ldquo;smart utensils\u0026rdquo; to detect eating moments and eating speed [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Chemical sensors have also been developed to analyze biomarkers associated with metabolism, including breath measurements of ketones (Ketonix AB, Varberg, Sweden) and CO2 (Lumen, MetaFlow Ltd., Tel Aviv, Israel), and ketones in interstitial fluid [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. However, these technologies only capture limited information, such as the timing and type of foods, but not their macronutrient composition, which is critical in disease management.\u003c/p\u003e\u003cp\u003eStudies over the past decade have explored the use of continuous glucose monitors (CGMs) to develop \u003cem\u003epersonalized nutrition\u003c/em\u003e programs that account for inter-individual differences. CGMs are in-dwelling wearable devices that measure glucose concentration in interstitial fluid. Though initially developed for Type 1 Diabetes, CGMs are increasingly being used to treat individuals with Type 2 Diabetes [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Several companies (e.g., Levels, NutriSense) also offer CGM-based services for the wellness market. The origins of CGM-based personalized nutrition programs can be traced back to the seminal study of Zeevi, Korem [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. In this study, the authors collected CGM data and meal logs from an 800-person cohort and developed a machine-learning (ML) model that integrated clinical features, dietary habits, physical activity, and gut microbiome profiles to predict glycemic responses to meals. Their approach was validated in a randomized controlled trial, where personalized dietary recommendations significantly reduced postprandial glucose excursions. These studies have been replicated in different cohorts [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eCGM-based personalized nutrition approaches such as that of Zeevi, Korem [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e] can be thought of as developing a \u003cem\u003edirect\u003c/em\u003e metabolic model that predicts post-prandial glucose responses (PPGRs) from meal macronutrients. Over the past decade, our work has focused on the \u003cem\u003einverse\u003c/em\u003e metabolic model: predicting meal macronutrients from PPGRs, as an approach for automatic diet monitoring [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. The primary contributor to elevated PPGRs is the amount (and type) of carbohydrates, but PPGRs are also influenced by other, non-glycemic macronutrients in the meal \u0026ndash;see Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea-b. For example, adding protein and fat to a carbohydrate-rich meal can reduce the initial peak and slow down recovery to baseline [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Unfortunately, the \u003cem\u003einverse\u003c/em\u003e problem is far more challenging, as many different meals can lead to similar PPGRs. Compounding the problem, metabolic health parameters (e.g., HbA1c, insulin sensitivity) and demographics (e.g., sex, age) also play a major role in PPGRs to identical meals. As a result, PPGRs from two individuals cannot be compared without controlling for these health/demographic factors.\u003c/p\u003e\u003cp\u003eTo address this challenge, we present an ML model that addresses both the direct (i.e., personalized nutrition) and the inverse (i.e., automatic diet monitoring) problems. Our model learns a numerical representation (referred to as an embedding in ML research) that aligns PPGRs with the macronutrient composition of a meal, \u003cem\u003econditioned\u003c/em\u003e on metabolic and demographic parameters. The result is an interpretable and personalized embedding of a meal that accounts for inter-individual differences.\u003c/p\u003e"},{"header":"MATERIALS/SUBJECTS AND METHODS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eModel design\u003c/h2\u003e\u003cp\u003eOur proposed model (JointCGMacros) uses neural networks to learn the underlying relationship between PPGRs, meal macronutrients and the individual\u0026rsquo;s health parameters. Illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb, the model consists of three distinct components: (1) a neural network that converts PPGRs into a single numeric value, i.e., a one-dimensional (1D) embedding, (2) a weighted parametric equation that converts meals macronutrients into a separate 1D embedding, and (3) a second neural network that converts each subject\u0026rsquo;s health parameters into the set of weights used by the parametric equation. In what follows, we will refer to the first neural network as the \u0026ldquo;PPGR encoder\u0026rdquo;, and the second neural network as the \u0026ldquo;health encoder\u0026rdquo;. We train the coefficients of the two neural networks so that their respective outputs match using an error function known as the triplet loss [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e] (see Supplementary materials). To use the triplet loss, we give the model two pairs of examples. The first pair contains PPGR and macronutrients from the same meal, whereas the second pair contains PPGR and macronutrients from different meals. The triplet loss forces the neural network outputs from the first pair to be closer to each other, and those from the second pair to be as far from each other as possible \u0026ndash;see Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ed. In doing so, the model learns how meal macronutrients and health parameters contribute to the observed PPGR, as measured by a CGM.\u003c/p\u003e\u003cp\u003eFormally, the PPGR encoder takes a 13-dimensional vector containing the 3-hour PPGR to a meal (as measured by an Abbot Freestyle Libre Pro CGM every 15 min), which we denote by \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:g\\left(t\\right)\\)\u003c/span\u003e\u003c/span\u003e, and generates a 1D embedding \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{P}\\)\u003c/span\u003e\u003c/span\u003e. In turn, the health encoder takes a 34-dimensional vector of health parameters \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:H\\)\u003c/span\u003e\u003c/span\u003e (see Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec) and generates a personalized set of weights that are combined with the macronutrient amounts of a meal \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:M\\)\u003c/span\u003e\u003c/span\u003e to produce a 1D embedding \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e as:\u003c/p\u003e\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\n$$\\:{z}_{M}=\\frac{C-{w}_{P}P-{w}_{F}F-{w}_{B}B}{1+{w}_{C}^{{\\prime\\:}}C+{w}_{P}^{{\\prime\\:}}P+{w}_{F}^{{\\prime\\:}}F+{w}_{B}^{{\\prime\\:}}B}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:C\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:P\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:F\\)\u003c/span\u003e\u003c/span\u003e, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:B\\)\u003c/span\u003e\u003c/span\u003e are the amount of net carbs, protein, fat and fiber, respectively, in calories\u003csup\u003e3\u003c/sup\u003e. Coefficients \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{P}\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{F}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{B}\\)\u003c/span\u003e\u003c/span\u003e in the numerator denote how much protein, fat, and fiber subtract from the effect of carbs, whereas coefficients \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{C}^{{\\prime\\:}}\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{P}^{{\\prime\\:}}\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{F}^{{\\prime\\:}}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{B}^{{\\prime\\:}}\\)\u003c/span\u003e\u003c/span\u003e in the denominator produce an adjusted caloric content. As such, the embedding \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e can be interpreted as a generalization of the carb-caloric-ratio (CCR), the proportion of calories in a meal that come from carbs, which is a measure of the glycemic load of a meal:\u003cdiv id=\"Equ2\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ2\" name=\"EquationSource\"\u003e\n$$\\:CCR=\\frac{C}{C+P+F+B}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e2\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eNote that for \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{C}^{{\\prime\\:}}=\\:{w}_{P}^{{\\prime\\:}}=\\:{w}_{F}^{{\\prime\\:}}={w}_{B}^{{\\prime\\:}}=0\\)\u003c/span\u003e\u003c/span\u003e in Eq.\u0026nbsp;(\u003cspan refid=\"Equ1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), the denominator becomes \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:1\\)\u003c/span\u003e\u003c/span\u003e, in which case \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e only considers the subtractive effect of non-glycemic macronutrients (protein, fat and fiber). In turn, for \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{C}^{{\\prime\\:}}={w}_{P}^{{\\prime\\:}}={w}_{F}^{{\\prime\\:}}={w}_{B}^{{\\prime\\:}}=1\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{P}={w}_{F}={w}_{B}=0\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e reduces to the CCR in Eq.\u0026nbsp;(\u003cspan refid=\"Equ2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). This parametric embedding has two main advantages. First, it allows the model to learn the optimum combination of macronutrients that is related to PPGRs (either linearly or non-linearly) for each subject. Second, because each of the resulting weights corresponds to a single macronutrient, the model is interpretable, as we will see in the Results section. Both \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{P}\\)\u003c/span\u003e\u003c/span\u003e are 1D embeddings to prioritize interpretability and reduce the risk of overfitting the training dataset.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eThe inverse metabolic model: predicting macro embeddings from PPGRs\u003c/h3\u003e\n\u003cp\u003eOnce trained, the model can be used to predict the macro embedding (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e) of a meal from the corresponding PPGR, as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ee. First, the PPGR of the meal is passed to the PPGR encoder to obtain the corresponding embedding \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{P}\\)\u003c/span\u003e\u003c/span\u003e. Then, the individual's health parameters \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:H\\)\u003c/span\u003e\u003c/span\u003e are passed to the macro encoder and combined with the meal\u0026rsquo;s macronutrients as a vector \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:M\\)\u003c/span\u003e\u003c/span\u003e, to produce the embedding \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e. Since \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:M\\)\u003c/span\u003e\u003c/span\u003e is unknown for a test PPGR, we generate multiple embeddings \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:({z}_{M,1},\\:{z}_{M,2},\\:...\\:{z}_{M,n})\\)\u003c/span\u003e\u003c/span\u003e from a list of potential meals \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:({M}_{1},\\:{M}_{2},\\:...\\:{M}_{n})\\)\u003c/span\u003e\u003c/span\u003e in a dataset. The meal \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{M}_{i}\\)\u003c/span\u003e\u003c/span\u003e whose macronutrient embedding \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M,i}\\)\u003c/span\u003e\u003c/span\u003e is closest to \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{P}\\)\u003c/span\u003e\u003c/span\u003e is treated as the prediction.\u003c/p\u003e\u003cp\u003eWe compare the inverse mapping in JointCGMacros against a baseline model that predicts the CCR in Eq.\u0026nbsp;(\u003cspan refid=\"Equ2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) directly from its corresponding PPGR. The baseline model is a feedforward network with the same number of parameters as the PPGR encoder in JointCGMacros; see Supplementary materials. \u003cem\u003eBecause the baseline model does not use health parameters H, this comparison allows us to determine if the personalized macro embedding\u003c/em\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e\u003cem\u003ecan reduce the effect of inter-individual differences.\u003c/em\u003e We train both models using 10-fold cross-validation. For each training fold, we randomly select 80% of the data for training and 20% for validation. We use validation data for early stopping (10 epochs) to prevent overfitting. We report two performance measures on the test fold: the normalized root mean squared error (NRMSE) and the Pearson correlation between ground truth and prediction. Because the NRMSE is the percentage of error relative to ground truth, it allows us to compare errors between target variables of different scales.\u003c/p\u003e\n\u003ch3\u003eThe direct metabolic model: predicting PPGRs from macro embeddings\u003c/h3\u003e\n\u003cp\u003eOnce trained, the joint model can also be used to predict the full PPGR of a meal from the corresponding set of an individual\u0026rsquo;s health parameters \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:H\\)\u003c/span\u003e\u003c/span\u003e and meal macronutrients \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:M\\)\u003c/span\u003e\u003c/span\u003e, as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ee. First, parameters \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\left(H,M\\right)\\)\u003c/span\u003e\u003c/span\u003e are passed to the macro encoder to obtain the corresponding embedding \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e. Then, we find the PPGR embedding \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{P}\\)\u003c/span\u003e\u003c/span\u003e in the training set that is closest to \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e. Finally, we recover PPGR from \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{P}\\)\u003c/span\u003e\u003c/span\u003e by inverting the PPGR encoder via linear regression using \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\left({\\sqrt{{z}_{M}},\\:z}_{M},\\:{z}_{M}^{2},{z}_{M}^{3}\\right)\\)\u003c/span\u003e\u003c/span\u003e as predictors, one linear regression model for each time step in the PPGR, i.e., \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:g\\left(t\\right)\\)\u003c/span\u003e\u003c/span\u003e, for \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:t=15,\\:30,\\dots\\:\\)\u003c/span\u003e\u003c/span\u003eminutes; see Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb.\u003c/p\u003e\u003cp\u003eWe compare the direct mapping in JointCGMacros against a baseline model that predicts each PPGR timestep from parameters \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\left(H,M\\right)\\)\u003c/span\u003e\u003c/span\u003e. The baseline model is a feedforward network with the same number of parameters as the PPGR encoder in JointCGMacros; see Supplementary materials. As before, we train both models using 10-fold cross-validation and use validation data for early stopping (10 epochs) to prevent overfitting. We report the NRMSE and Pearson correlation between ground truth and prediction for the two-hour incremental area-under-the-curve (iAUC2hr), as is common to examine results from the Oral Glucose Tolerance Test (OGTT). Note that, in this case, the odds are stacked against the JointCGMacros because it predicts PPGRs from a single variable (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e), whereas the baseline model uses 38 predictors (i.e., the 34 health parameters in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec plus the 4 macronutrients). Thus, this comparison allows us to determine if the embedding \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e is a good summary statistic of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\left(H,M\\right)\\)\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e\n\u003ch3\u003eData records\u003c/h3\u003e\n\u003cp\u003eWe trained the proposed model on the CGMacros dataset [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e] publicly-available on PhysioNet, which contains PPGRs, meal macronutrients and health/demographics parameters for 45 subjects (ages 18\u0026ndash;69 and body mass index (BMI) 21\u0026ndash;46 kgm\u003csup\u003e2\u003c/sup\u003e); clinical trial #NCT04991142. Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb summarizes the demographic information of study participants. As part of the screening process, we measured the subjects\u0026rsquo; BMI, glycated hemoglobin (HbA1c), fasting glucose, fasting insulin, triglyceride, and cholesterol levels, as well as their demographic information (age, gender, and race). Participants wore an Abbott FreeStyle Libre Pro CGM (15-min sampling period) and a Dexcom G6 Pro CGM (5-min sampling) in the upper arm and abdomen, respectively. Both CGMs were blinded to prevent glucose readings from influencing participants. Participants were also provided with a Fitbit smartwatch (Fitbit Sense) to log exercise, and were trained to use the MyFitnessPal mobile app to log their meals and take pictures of their foods using the WhatsApp mobile app.\u003c/p\u003e\u003cp\u003eEach subject recorded their meals for 10 days, including breakfast, lunch and dinner. Breakfasts consisted of protein shakes with varying amounts of carbohydrates, protein, fat, and fiber -see Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea. Lunches were ordered from a local, fast-casual restaurant chain (Chipotle Mexican Grill). The breakfast and lunch meals were designed to cover a range of macronutrient contents. For dinners, participants ate foods of their own choice. To minimize interferences in glucose responses from prior meals, participants were instructed to eat lunch at least three hours after breakfast, with only water or coffee (without sugar) in between, and dinner at least three hours after lunch. They also took photographs of the meals before and after eating, from which we extracted the meal timestamps and the proportion of the meal they consumed. At the start of the study, we also collected stool samples using a Viome microbiome kit (Viome Life Sciences, Inc.) to obtain the gut microbiome profile for each subject. We report results from the breakfast meals using PPGRs from the Abbott FreeStyle CGM. We used 34 variables available on the CGMacros dataset for each subject; see Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003ePredictive accuracy on test data\u003c/h2\u003e\u003cp\u003eFirst, we compare the performance of JointCGMacros model against the baseline inverse model that predicts CCR directly from PPGRs (i.e., without personalization). Results are summarized in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea. JointCGMacros strongly outperforms (lower NRMSE, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:p\u0026lt;0.01\\)\u003c/span\u003e\u003c/span\u003e; higher correlation, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:p\u0026lt;0.001\\)\u003c/span\u003e\u003c/span\u003e) the baseline model, indicating that JointCGMacros accounts for inter-individual differences in postprandial glucose responses. Note that, for the baseline model, the correlation coefficient can become negative and the NRMSE can exceed an error of 100%, indicating that PPGRs are not predictive of the macronutrient composition of meals \u003cem\u003eunless individual health/demographics parameters are considered\u003c/em\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eWe also compare the performance of JointCGMacros model against a baseline direct model that predicts PPGRs directly from parameters \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\left(H,M\\right)\\)\u003c/span\u003e\u003c/span\u003e. In this case, we compute the iAUC-2hr from the predicted PPGRs and compare against the ground truth\u0026rsquo;s iAUC-2hr. Results are summarized in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb. JointCGMacros achieves similar NRMSE and correlation coefficients (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:p\u0026gt;0.05\\)\u003c/span\u003e\u003c/span\u003e) despite using a single predictor (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M})\\)\u003c/span\u003e\u003c/span\u003e, whereas the baseline model uses 38 predictors. This suggests that the complex relationship between health parameters, gut microbiome and food macronutrients is intrinsically low dimensional.\u003c/p\u003e\u003cp\u003eWe also examined pairwise predictions from the macro and PPGR encoders. Results are illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec as scatterplot of the corresponding embeddings (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M},{z}_{P}\\)\u003c/span\u003e\u003c/span\u003e), color-coded by the ground-truth CCR in Eq.\u0026nbsp;(\u003cspan refid=\"Equ2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Embeddings towards the bottom left (dark blue) have the lowest CCR, which represent meals with high amounts of carbs, protein, fat, and fiber (e.g., meal HHHH in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Since these meals have the highest amounts of non-glycemic macronutrients, they tend to result in low postprandial glucose. In contrast, meals with the highest CCR (dark red) appear towards the top right. These meals have a high amount of carbs and low amounts of other macronutrients (HLLL in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), which leads to higher postprandial glucose.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eModel interpretation\u003c/h3\u003e\n\u003cp\u003eA major limitation of deep-learning models is their lack of interpretability, which can be problematic in critical applications, such as health. To address this limitation, we also conduct a series of analyses to shed light on the inner workings of our proposed model.\u003c/p\u003e\u003cp\u003eIn a first step, we quantify the importance of the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:34\\)\u003c/span\u003e\u003c/span\u003e health parameters in JointCGMacros using Bidirectional Search (BDS), a procedure that alternates between forward selection and backward elimination of predictors. Starting from the empty set, the forward step adds the best single predictor (i.e., lowest NRMSE) to the model; in turn, starting from the full set, the backward step eliminates the worst single predictor (i.e., highest NRMSE). As a result, both searches converge in the middle (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:17\\)\u003c/span\u003e\u003c/span\u003e predictors). Results are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed. The two most important features are HOMA-IR, a clinical measure of insulin resistance and beta-cell function, and fasting glucose, a key parameter for the diagnosis of T2D. The final feature set includes a mix of demographics (age, gender), metabolic health (lipid profiles, HbA1c), and several pathways from gut microbiome analytics. This optimal set achieves lower NRMSE than the full feature set (rightmost boxplot, in gray), a difference that is statistically significant (0.247 vs. 0.291, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:p=0.03\\)\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eWe also examined the relationship between the two 1D embeddings (i.e., \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{P}\\)\u003c/span\u003e\u003c/span\u003e from PPGRs; \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e from health parameters and meal macros) learned by the model. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ee shows the average PPGR at increments of 5% in the quantile distribution of the PPGR embedding \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{P}\\)\u003c/span\u003e\u003c/span\u003e. PPGRs in the lowest quantiles of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{P}\\)\u003c/span\u003e\u003c/span\u003e (i.e., 5\u0026ndash;20%) show a biphasic characteristic, with a rise in the first 30\u0026ndash;60 minutes, a subsequent drop, and another increase 90\u0026ndash;120 minutes later. This type of PPGR is considered an indicator of good metabolic health [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. In contrast, PPGRs in the highest quantiles of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{P}\\)\u003c/span\u003e\u003c/span\u003e (i.e., 70\u0026ndash;95%) show a monophasic characteristic, with a single peak followed by a continuous drop, which is indicative of less optimal metabolic health. Likewise, we analyzed the distribution of health parameters for samples in the first (25%) and fourth (75%) quartiles of the distribution of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e. Results for a selected few health parameters and meal macronutrients are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ef. On average, subjects with low HbA1c, low fasting glucose, low HOMA-IR and low BMI tend to produce lower embeddings \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e, regardless of meal macronutrients. Likewise, meals high in carbs, low in protein, or low in fiber tend to produce higher embeddings \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{M}\\)\u003c/span\u003e\u003c/span\u003e, results that are consistent with the literature [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eTo corroborate these observations, we examined the weights in Eq.\u0026nbsp;(\u003cspan refid=\"Equ1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) learned by the model. In the numerator, fiber (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{B}=3.95\\)\u003c/span\u003e\u003c/span\u003e) has the strongest effect, followed by protein (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{P}=1.22\\)\u003c/span\u003e\u003c/span\u003e) and fat (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{F}=0.34\\)\u003c/span\u003e\u003c/span\u003e), suggesting healthy alternatives to control elevated post-prandial glucose response (i.e., add fiber and protein, but not fat). In the denominator, net carbs (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{C}^{{\\prime\\:}}\\)\u003c/span\u003e\u003c/span\u003e) have the strongest contribution (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{C}^{{\\prime\\:}}=0.33\\)\u003c/span\u003e\u003c/span\u003e), whereas the effect of the three non-glycemic macronutrients is negligible (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{w}_{P}^{{\\prime\\:}}=0.03,\\:{w}_{F}^{{\\prime\\:}}=0.01,\\:{w}_{B}^{{\\prime\\:}}=0.01\\)\u003c/span\u003e\u003c/span\u003e). Ignoring the smallest coefficients, these results indicate that, \u003cem\u003ewhen accounting for health parameters\u003c/em\u003e, post-prandial glucose responses are proportional to the expression \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:(C-3.95B-1.22P)/(1+0.33C)\\)\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eMonitoring diet in an automated fashion is made possible by ML methods that account for both a person\u0026rsquo;s response to a meal (as captured by a CGM), and their individual demographics and characteristics. By jointly learning an individual\u0026rsquo;s health characteristics and observing his/her responses to individual meals, we have developed a model (JointCGMacros) that makes possible automated modeling of diet by both generating a PPGR from a known meal (direct) and estimate a meal\u0026rsquo;s macronutrient composition from a captured response as measured by a CGM (inverse). Baseline models that estimate the direct or indirect task without health parameters can generate errors over 100% NRMSE and JointCGMacros significantly outperforms the baselines. Perhaps more importantly, when compared with a baseline that uses higher dimensionality, JointCGMacros, which compresses data into a single embedding for PPGRs and a single embedding for health parameters, uses a low dimensional structure to create meaningful metabolic representations.\u003c/p\u003e\u003cp\u003ePrior research has shown that the relationship between post-prandial glucose and meal macronutrients is largely driven by inter-individual differences. Our results show that this three-way relationship can be unveiled by learning a joint embedding of PPGRs and macronutrient composition conditioned on a few demographics, metabolic health, and gut microbiota parameters. We find that the key contributors to elevated postprandial glucose are net carbs (as expected) with a linear correction for fiber and protein. It is the expression \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:(C-3.95B-1.22P)/(1+0.33C)\\)\u003c/span\u003e\u003c/span\u003e, and not the conventional carb-caloric-ratio, that can be recovered from PPGRs, but only when accounting for an individual's health variables.\u003c/p\u003e\u003cp\u003eBeyond predictive performance, we also identified factors related to the meals and health parameters that drove model performance. These features included HOMA-IR and fasting glucose information as the most influential health predictors, aligning with established risk markers for Type 2 Diabetes, while embedding low-dimensional representation of gut microbiome pathways. Moreover, the embeddings revealed biologically plausible distinctions between monophasic and biphasic glucose curves, echoing prior associations with metabolic health. At the macronutrient level, fiber and protein emerged as protective factors, while excess carbohydrate intake strongly drove elevated responses. Together, these insights suggest that JointCGMacros not only improves prediction but also uncovers mechanistic relationships that may guide personalized nutrition strategies.\u003c/p\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eLimitations\u003c/h2\u003e\u003cp\u003eThe study did have some limitations that can be addressed with future study. First, we used the parametric expression in Eq.\u0026nbsp;(\u003cspan refid=\"Equ1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) to improve the interpretability of our model, but further improvements in macronutrient prediction from PPGRs may be achieved by relaxing this constraint, e.g., concatenating macronutrient and health vectors \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:[M,H]\\)\u003c/span\u003e\u003c/span\u003e and feeding them to a deep-learning model, but this would likely require access to a larger dataset than CGMacros, the only of its nature that is publicly available. Second, CGMacros contains individuals who are not taking insulin. Therefore, monitoring for individuals requiring insulin will require data and modeling of additional health characteristics and insulin dosage to model the effect on the PPGR.\u003c/p\u003e\u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003eAcknowledgements\u003c/p\u003e\n\u003cp\u003eThis work was supported by National Science Foundation Award No. 2014475.\u003c/p\u003e\n\u003cp\u003eAuthor contributions\u003c/p\u003e\n\u003cp\u003eAD was responsible for data curation and data analytics. AD and GN were responsible for implementation and evaluation of the proposed machine learning model. \u0026nbsp;BM was Principal Investigator of the NSF award; he was responsible for project oversight and co-lead data analytics and meal design. RGO was co-Investigator of the NSF award; he co-led data analytics for the NSF award and was responsible for manuscript preparation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eData availability statement\u003c/p\u003e\n\u003cp\u003eThe dataset used for this study is publicly available at PhysioNet (https://physionet.org/content/cgmacros)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interests:\u0026nbsp;\u003c/strong\u003eThe authors declare no competing interests for this work. This work was supported by National Science Foundation Award No. 2014475.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eMortazavi, B.J. and R. Gutierrez-Osuna, \u003cem\u003eA review of digital innovations for diet monitoring and precision nutrition.\u003c/em\u003e J Diabetes Sci Technol, 2023. \u003cstrong\u003e17\u003c/strong\u003e(1): p. 217--223.\u003c/li\u003e\n\u003cli\u003eZeevi, D., et al., \u003cem\u003ePersonalized Nutrition by Prediction of Glycemic Responses.\u003c/em\u003e Cell, 2015. \u003cstrong\u003e163\u003c/strong\u003e(5): p. 1079-1094.\u003c/li\u003e\n\u003cli\u003eBell, B.M., et al., \u003cem\u003eAutomatic, wearable-based, in-field eating detection approaches for public health research: a scoping review.\u003c/em\u003e npj Digital Medicine, 2020. \u003cstrong\u003e3\u003c/strong\u003e(1): p. 38.\u003c/li\u003e\n\u003cli\u003eZhang, B., et al., \u003cem\u003eEnabling Eating Detection in a Free-living Environment: Integrative Engineering and Machine Learning Study.\u003c/em\u003e J Med Internet Res, 2022. \u003cstrong\u003e24\u003c/strong\u003e(3): p. e27934.\u003c/li\u003e\n\u003cli\u003eKumar, Y., et al., \u003cem\u003eAutomated detection and recognition system for chewable food items using advanced deep learning models.\u003c/em\u003e Scientific Reports, 2024. \u003cstrong\u003e14\u003c/strong\u003e(1): p. 6589.\u003c/li\u003e\n\u003cli\u003eHuang, Q., Z. Yang, and Q. Zhang. \u003cem\u003eSmart-u: Smart utensils know what you eat\u003c/em\u003e. in \u003cem\u003eIEEE Conference on Computer Communications (INFOCOM)\u003c/em\u003e. 2018. IEEE.\u003c/li\u003e\n\u003cli\u003eZhang, J.Y., et al., \u003cem\u003eContinuous Ketone Monitoring: A New Paradigm for Physiologic Monitoring.\u003c/em\u003e J Diabetes Sci Technol, 2021: p. 19322968211009860.\u003c/li\u003e\n\u003cli\u003eAlva, S., et al., \u003cem\u003eFeasibility of Continuous Ketone Monitoring in Subcutaneous Tissue using a Ketone Sensor.\u003c/em\u003e J Diabetes Sci Technol, 2021: p. 19322968211008185.\u003c/li\u003e\n\u003cli\u003eJancev, M., et al., \u003cem\u003eContinuous glucose monitoring in adults with type 2 diabetes: a systematic review and meta-analysis.\u003c/em\u003e Diabetologia, 2024. \u003cstrong\u003e67\u003c/strong\u003e(5): p. 798-810.\u003c/li\u003e\n\u003cli\u003eZeevi, D., et al., \u003cem\u003ePersonalized Nutrition by Prediction of Glycemic Responses.\u003c/em\u003e Cell, 2015. \u003cstrong\u003e163\u003c/strong\u003e(5): p. 1079--1094.\u003c/li\u003e\n\u003cli\u003eMendes-Soares, H., et al., \u003cem\u003eAssessment of a Personalized Approach to Predicting Postprandial Glycemic Responses to Food Among Individuals Without Diabetes.\u003c/em\u003e JAMA Netw Open, 2019. \u003cstrong\u003e2\u003c/strong\u003e(2): p. e188102.\u003c/li\u003e\n\u003cli\u003eMendes-Soares, H., et al., \u003cem\u003eModel of personalized postprandial glycemic response to food developed for an Israeli cohort predicts responses in Midwestern American individuals.\u003c/em\u003e Am J Clin Nutr, 2019. \u003cstrong\u003e110\u003c/strong\u003e(1): p. 63--75.\u003c/li\u003e\n\u003cli\u003eZhang, L., et al. \u003cem\u003eJoint Embedding of Food Photographs and Blood Glucose for Improved Calorie Estimation\u003c/em\u003e. in \u003cem\u003e2023 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI)\u003c/em\u003e. 2023.\u003c/li\u003e\n\u003cli\u003eMoghaddam, E., J.A. Vogt, and T.M.S. Wolever, \u003cem\u003eThe effects of fat and protein on glycemic responses in nondiabetic humans vary with waist circumference, fasting plasma insulin, and dietary fiber intake.\u003c/em\u003e J Nutr, 2006. \u003cstrong\u003e136\u003c/strong\u003e(10): p. 2506--2511.\u003c/li\u003e\n\u003cli\u003eRytz, A. and D. Adeline, \u003cem\u003ePredicting glycemic index and glycemic load from macronutrients to accelerate development of foods and beverages with lower glucose responses.\u003c/em\u003e Nutrients, 2019. \u003cstrong\u003e11\u003c/strong\u003e(5): p. 1172.\u003c/li\u003e\n\u003cli\u003eWeinberger, K.Q. and L.K. Saul, \u003cem\u003eDistance metric learning for large margin nearest neighbor classification.\u003c/em\u003e Journal of machine learning research, 2009. \u003cstrong\u003e10\u003c/strong\u003e(2).\u003c/li\u003e\n\u003cli\u003eGutierrez-Osuna, R., et al. \u003cem\u003eCGMacros: a scientific dataset for personalized nutrition and diet monitoring\u003c/em\u003e. 2025; Available from: https://physionet.org/content/cgmacros.\u003c/li\u003e\n\u003cli\u003eRobbins, D.C., et al., \u003cem\u003eBiphasic patterns of peripheral insulin and glucose levels after lunch in normal subjects.\u003c/em\u003e Diabetes Care, 1987. \u003cstrong\u003e10\u003c/strong\u003e(3): p. 293-9.\u003c/li\u003e\n\u003cli\u003eKetema, E.B. and K.T. Kibret, \u003cem\u003eCorrelation of fasting and postprandial plasma glucose with HbA1c in assessing glycemic control; systematic review and meta-analysis.\u003c/em\u003e Arch Public Health, 2015. \u003cstrong\u003e73\u003c/strong\u003e: p. 43.\u003c/li\u003e\n\u003cli\u003eIoffe, S. and C. Szegedy, \u003cem\u003eBatch normalization: Accelerating deep network training by reducing internal covariate shift\u003c/em\u003e, in \u003cem\u003eInternational Conference on Machine Learning (ICML)\u003c/em\u003e. 2015. p. 448--456.\u003c/li\u003e\n\u003cli\u003eLi, Y. and Y. Yuan, \u003cem\u003eConvergence analysis of two-layer neural networks with relu activation\u003c/em\u003e, in \u003cem\u003eAdvances in Neural Information Processing Systems (NIPS)\u003c/em\u003e. 2017.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Footnotes","content":"\u003col start=\"3\"\u003e\u003cli\u003e\u003cspan\u003e The dataset used for this study contains six unique meal combinations. This results in five distinct triplets per CGM, each sharing the same anchor and positive, but differing in the negative\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cimg src=\"data:image/png;base64,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\" style=\"width: 635px; height: 531.306px;\"\u003e\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"international-journal-of-obesity","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"ijo","sideBox":"Learn more about [International Journal of Obesity](http://www.nature.com/ijo/)","snPcode":"41366","submissionUrl":"https://mts-ijo.nature.com/cgi-bin/main.plex","title":"International Journal of Obesity","twitterHandle":"@intjobesity","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7511427/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7511427/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground/Objectives\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDiet monitoring for chronic disease management is often hindered by burdensome self-reporting. Automated techniques are possible using continuous glucose monitors (CGMs), which capture post-prandial glucose responses (PPGRs), but inter-individual variability limits predictive utility of models. We hypothesized that machine learning models that jointly represent meals and individual health parameters would improve prediction of PPGRs from meal composition (direct) and inference of meal composition from PPGRs (inverse).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSubjects/Methods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e: Data came from the CGMacros dataset. Forty-five individuals with varied HbA1c, demographics, labs, and gut microbiome profiles were included and their PPGRs captured by CGMs. We developed JointCGMacros, a deep neural network framework that encodes PPGRs and participant health features. Using triplet loss, we aligned similar meal-health pairs while separating dissimilar ones in the embedding space. From this, we estimated the two-hour incremental area under the curve (iAUC) of PPGRs (direct) and carb-to-calorie ratio of meals (inverse). Models were trained and validated with 10-fold cross-validation, with performance measured using Pearson correlation and normalized root mean square error (NRMSE).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInterventions/methods used and duration of administration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCGMacros provided participants 10 breakfast shakes with varied macronutrient compositions, 10 Chipotle lunches with varied compositions, and free-choice dinners. Participants fasted before breakfasts and avoided eating for three hours after breakfast and lunch.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOn the direct task, we achieved an NRMSE=0.22 and Pearson correlation r=0.59, outperforming baseline predictions (NRMSE=0.23, r =0.52), though not significantly at p\u0026lt;0.01. On the inverse task, we achieved an NRMSE=0.28 and r = 0.47, a statistically significant improvements over baseline predictions (NRMSE=0.62,p\u0026lt;0.01; r=0.17,p\u0026lt;0.001).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWhile PPGRs alone are not predictive of meal composition without accounting for individual health parameters, JointCGMacros integrates these factors, enabling more accurate automated dietary monitoring and estimation of meal macronutrient effects.\u003c/p\u003e","manuscriptTitle":"AI-driven personalized metabolic models of postprandial glucose to mixed meals","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-06 10:07:00","doi":"10.21203/rs.3.rs-7511427/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"revise","date":"2026-04-27T16:07:20+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-03-23T12:07:14+00:00","index":3,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2026-02-19T13:26:08+00:00","index":3,"fulltext":"This content is not available."},{"type":"editorInvitedReview","content":"This content is not available.","date":"2025-10-13T05:59:56+00:00","index":2,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2025-09-29T15:10:02+00:00","index":2,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2025-09-23T12:49:59+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewersInvited","content":"","date":"2025-09-23T00:32:57+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-03T14:13:10+00:00","index":"","fulltext":""},{"type":"submitted","content":"International Journal of Obesity","date":"2025-09-02T17:35:27+00:00","index":"","fulltext":""},{"type":"checksFailed","content":"","date":"2025-09-02T10:28:47+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-01T20:30:36+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"international-journal-of-obesity","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"ijo","sideBox":"Learn more about [International Journal of Obesity](http://www.nature.com/ijo/)","snPcode":"41366","submissionUrl":"https://mts-ijo.nature.com/cgi-bin/main.plex","title":"International Journal of Obesity","twitterHandle":"@intjobesity","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"92209154-2c18-46d7-8725-96f761922f0a","owner":[],"postedDate":"October 6th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[{"id":55153780,"name":"Health sciences/Health care/Nutrition"},{"id":55153781,"name":"Health sciences/Diseases/Endocrine system and metabolic diseases/Diabetes/Type 2 diabetes"}],"tags":[],"updatedAt":"2026-04-27T16:15:18+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-06 10:07:00","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7511427","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7511427","identity":"rs-7511427","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}