4 research outputs found
Classifying Food Items During an Eating Occasion: A Machine Learning Approach with Slope Dynamics for Windowed Kinetic Data
No full textBackground: Wearable devices equipped with a range of sensors have emerged as promising tools for monitoring and improving individuals’ health and lifestyle. Objectives: Contribute to the investigation and development of effective and reliable methods for dietary monitoring based on raw kinetic data generated by wearable devices. Methods: This study uses resources from the NOTION study. A total of 20 healthy subjects (9 women and 11 men, aged 20–31 years) were equipped with two commercial smartwatches during four eating occasions under semi-naturalistic conditions. All meals were video-recorded, and acceleration data were extracted and analyzed. Food recognition on these features was performed using random forest (RF) models with 5-fold cross-validation. The performance of the classifiers was expressed in out-of-bag sensitivity and specificity. Results: Acceleration along the x-axis and power show the highest and lowest rates of median variable importance, respectively. Increasing the window size from 1 to 5 s leads to a gain in performance for almost all food items. The RF classifier reaches the highest performance in identifying meatballs (89.4% sensitivity and 81.6% specificity) and the lowest in identifying sandwiches (74.6% sensitivity and 72.5% specificity). Conclusions: Monitoring food items using simple wristband-mounted wearable devices is feasible and accurate for some foods while unsatisfactory for others. Machine learning tools are necessary to deal with the complexity of signals gathered by the devices, and research is ongoing to improve accuracy further and work on large-scale and real-time implementation and testing
Analyzing the Caloric Variability of Bites in a Semi-Naturalistic Dietary Setting
No full textBackground: Obesity is a major public health issue in developed countries, primarily managed through dietary interventions and physical activity. Food portion sizes influence the estimation of energy intake, particularly through bites, of which characteristics remain insufficiently defined. This study investigates the variability in bite energy content. Methods: This observational study was conducted over 14 months. Thirteen types of packaged food were provided to 30 Italian healthy volunteers (mean age 26.8 ± 8.5 years) in a semi-naturalistic dietary feeding setting. Participants’ anthropometric measurements were recorded. A total of 1850 bites were weighed and 420 bites were assessed for volume and energy content. Results: Bite volume and mass explained bite energy content at different rates. The most influential anthropometric feature was waist circumference. Gender modified the association between waist circumference and bite characteristics; males showed increased bite volume, mass, and energy content as waist circumference increased, whereas females showed little or no association. Age was inversely associated with bite volume and mass, with younger participants having larger bites. Gender significantly influenced average bite size, with females showing lower values than males. The use of a fork was associated with higher bite volume, mass, and energy compared to a spoon. Food eaten with bare hands had lower mass but higher energy content compared to food eaten with a spoon. The variability in bite energy was considerably greater per bite than per gram, reflecting the combined influence of food texture, bite size, and cutlery used. Conclusions: Bite energy variability, influenced by intrinsic factors (gender, age, waist circumference) and extrinsic factors (cutlery, food texture), significantly impacts portion size effect. Future bite counters should consider these elements for accurate dietary assessment
Lower intakes of protein, carbohydrate, and energy are associated with increased global DNA methylation in 2‐ to 3‐year‐old urban slum children in Bangladesh
No full textStunting in children is a global public health concern. We investigated how global DNA methylation relates to food intakes, dietary diversity, and development of stunting among 324 children aged 24-36 months in a slum community in Dhaka, Bangladesh. Stunted children (height-for-age z score ˂-2; n = 162) and their age- and sex-matched nonstunted counterparts (height-for-age z score ˃-1; n = 162) were selected by active community surveillance. We studied global DNA methylation, measured as 5-mC% content in whole blood. Dietary intake, anthropometric measurement, and sociodemographic information were obtained. In the multiple linear regression model, increased global DNA methylation level in children was significantly associated with consumption of lower amount of energy, coef: .034 (95% CI [.014, .053]); P = .001, protein, coef: .038 (95% CI [.019, .057]); P = .000, carbohydrate, coef: .027 (95% CI [.008, .047]); P = .006, zinc, coef: .020 (95% CI [.001, .039]); P = .043, total dietary intakes, coef: .020 (95% CI [.001, .039]); P = .043, and intake from plant sources, coef: .028 (95% CI [.009, .047]); P = .005, after adjusting for other covariates. Moreover, higher fruits and vegetables consumption was significantly associated with lower 5-mC% level, coef: -.022 (95% CI [-.041, -.002]); P = .028. Our findings suggest a significant association between low dietary intakes and increased global DNA methylation. We also found increased global DNA methylation in stunted children. To establish the relationship among the macronutrient intakes, global DNA methylation, and stunting, future prospective studies are warranted in resource-poor settings.No Full Tex
